Intro

My weekly routine is finding and reading new research for Stacy’s Brain. As I segment a big, heavy paper into cards with clear sample and methodology data, I often find disagreements between what different authors say about the same topic.

This is probably one of the most interesting parts of my work — because controversy is where the market hasn’t yet found its efficiency zone, and where additional examination is worth it.

Our “patient report” for today is Sprout Social’s report: 2,305 consumers and 1,200 marketing leaders across the US, UK, and Australia — with a clean headline: consumers and marketers are looking in different directions.

Good, useful report overall. It’s also a survey of what people say.

So we did what we always do: we pulled every finding into our library, graded it, and checked it against what other sources — including those that measure real posts and real campaigns — say about the same thing.

TL;DR

1. Consumers say they’ll spend more time on Facebook. Marketers are putting money into Instagram and TikTok. In the US, 46% of social users plan more Facebook time, while marketers rank Facebook third (59%) behind Instagram (72%) and TikTok (69%).

2. But organic brand engagement on Facebook is a fraction of Instagram’s in almost every industry we have data for — fashion brands sat near 0.07% on Facebook against 0.31% on Instagram — and US Facebook CPMs rose 13% in a year. “More time on Facebook” doesn’t mean “more for your brand on Facebook”.

3. Short video is the one finding every dataset agrees on. Sprout’s consumers pick it first on Facebook, Instagram, TikTok and X; Metricool’s 24M Instagram posts show Reels were the only format whose engagement rate rose in 2026.

4. Consumers want human-made content; marketers are investing in AI. That gap is real, but smaller than it looks: most social teams use AI for only a minority of posts, and they rewrite what it produces. Use is growing, though — and it may lead to greater marketer-consumer misalignment.

5. Customer service on social is the gap nobody is staffing — and the alternative data says it should be staffed by people, not a bot.

Subscribe now

What this analysis is built on

Main source: The 2026 Social Media Content Strategy Report, Sprout Social, published January 2026.

• Consumer survey: 2,305 people in the US, UK, and Australia, run by Panoplai (6–16 October 2025). All respondents had at least one social account and followed at least five brands.

• Marketer survey: 1,200 marketing leaders (manager to CMO) in the same three countries, run by Glimpse, 30 May – 20 June 2025.

• Two pulse surveys (2,270 and 2,250 consumers, July–September 2025) supply the “where will you spend more time” numbers.

Three things to hold in your head for every number below.

First, this sample is now about a year old. Some trends — such as the use of AI-generated content — have likely increased since the survey was conducted.

Second, the consumer sample is people who follow five or more brands. That’s a sample that already likes interacting with brands, so expect every “how often do you engage” figure to run high.

Third, almost everything here is self-reported — “the content type I’m most likely to interact with”, “the network I plan to spend more time on”. These are preferences and intentions, so where we found measured data on the same question, we put it next to Sprout’s.

Check all insights from Sprout Social

What we checked it against, all graded in Stacy’s Brain:

• Metricool Instagram study 2026 — 24.4M posts from 375K accounts, Jan–Feb 2025 vs Jan–Feb 2026. Evidence type: observed. Vendor interest: moderate.

• Socialinsider content trends 2026 — 23M Facebook and 33M Instagram brand posts across 16 industries, 2025. Evidence type: observed. Vendor interest: high.

• HubSpot State of Marketing 2026 — 1,505 marketers, September 2025. Evidence type: self-reported. Vendor interest: moderate.

• Tinuiti Q2 2026 digital ads benchmark — $4B+ in managed ad spend. Evidence type: observed. Vendor interest: high.

• Gartner customer chatbot survey 2026 — 3,566 B2B and B2C customers, Feb–Mar 2026. Evidence type: stated intention.

• Plus smaller pieces from Sociality.io, Promptwatch and an independent Reddit-and-AI study.

The Facebook paradox

Ask consumers where they’ll spend more time in 2026, and Facebook comes first in all three countries, according to Sprout:

• US: Facebook 46%, YouTube 37%, Instagram 31%

• UK: Facebook and Instagram tied at 34%, TikTok and WhatsApp 29%

• Australia: Facebook 37%, Instagram 27%, YouTube 25%

Ask marketers where they’ll invest more, and Facebook drops — except in Australia:

• US marketers: Instagram 72%, TikTok 69%, Facebook 59%

• UK marketers: TikTok 73%, Instagram 69%, Facebook 64%

• Australian marketers: Facebook 79%, TikTok 70%

And then, a twist:

Sorry, wtf?
Sprout’s reading is that brands are chasing the shiny networks. That likely misses some context: money invested and business impact are not the same thing.

My reading is that, for most brands in the US and UK, a Facebook presence is either already established or easily won. There may already be a brand-to-customer layer that’s taken for granted, simply because buyers have a habit of tracking brand news on Facebook.

By contrast, Instagram and TikTok may be the newer, still-untapped zones where marketers are trying to build relationships with potentially younger audiences and the next generation of customers. So it’s normal that a higher percentage of budget is invested there.

So the headline “controversy” in Sprout might actually be logical.

But let’s not rush to conclusions.

Here’s what the rest of Stacy’s Brain library adds.

Marketers elsewhere say the opposite about ROI

In HubSpot’s survey, the share calling Facebook their highest-ROI social platform fell from 54% to 43% in a year, while Instagram rose to 48% and took the lead.

It’s a different survey and a different question (“highest ROI” vs. “biggest impact”), but it’s still hard to square with Sprout’s 70%.

Organic brand engagement on Facebook is low

Socialinsider’s 2025 data on 23M Facebook brand posts puts Facebook below Instagram in almost every industry. By Q4, the leading content theme sat at roughly:

• Fashion: 0.07% on Facebook vs 0.31% on Instagram

• Tech & software: 0.14% vs ~0.40%

• Education: 0.15% vs 0.50%

• Travel: 0.11% vs ~0.40%

Hospitality was the clear exception, at about 0.33% on Facebook against 0.20% on Instagram.

Margin note: Don’t treat Socialinsider’s data as hard fact. Engagement rate isn’t calculated the same way on Facebook and Instagram—but the gap is too large and too consistent to be a rounding issue.

Plus, paid reach on Facebook is getting pricier

Tinuiti saw US Facebook CPMs rise 13% year over year in Q2 2026, the fastest of the five platforms it tracks. Instagram held flat.

How we’d read it (taking into consideration all data we have)

Consumers spending more time on Facebook is believable — but they’re telling you where they’ll be, not that they want brand posts there.

Sprout’s own data shows what they go to Facebook for: product discovery (nearly 40%) and community (52% call it their top network for community). So Facebook earns its place for groups, customer service and paid discovery, not for organic brand publishing.

The marketers shifting organic effort to Instagram aren’t necessarily wrong — as I earlier mentioned, they are likely to have a reason to invest into Instagram and TikTok.

Short video is the lead format for SMM

Sprout asked which brand content people are most likely to interact with on each network. Short-form video (under 60 seconds) wins almost everywhere:

• TikTok: 60%

• Instagram: 52%

• Facebook: 48%

• X: 37%, just edging past text posts at 36%

• YouTube: short 52%, long 49% — practically level

The exception is LinkedIn, where text posts lead at 51% and short video sits at 27%.

This is the rare self-reported finding that measured data backs up. In Metricool’s Instagram study, Reels were the only format whose engagement rate rose in 2026 (up 24.76% to 4.30%), Reels shares grew 67%, and average watch time more than doubled (+118%) while views barely moved (+2%).

Human-made vs. AI-made: a real gap, growing each month

In my opinion, Sprout’s most important finding is that consumers rank human-made content as brands’ #1 priority for 2026, while marketers — leaders especially — are making AI-generated content a top investment.

The consumer side is consistent with insights from other reports. In HubSpot’s survey, 63% of marketers say they need more unique, human-made content to stand out, and 53% say they struggle to stand out in an AI-saturated market. Nobody is arguing that AI slop works.

What the “marketers are betting on AI” framing misses is how they use it.

Since I’m the kind of marketer who automates a lot with AI, I can confidently say that content is the last thing you should try to automate. The devil is in the details, and while 100% content automation is tempting, it’s usually more damaging than helpful.

The current consensus is that AI is best used for data analysis, drafting, and expanding the range of “potential options.” Final review, reasoning, and bias checks should stay human — so you don’t lose reader attention.

In Sociality.io’s survey of social teams, AI touched under a quarter of posts for the largest group, and 78% said they edit AI output moderately or extensively before it goes out (the question didn’t offer “we publish it unedited” as an answer, so that share is unknown).

Brands are good at the thing that matters least

Sprout’s report has another finding that deserves more attention than Sprout gives it. When consumers rated how brands are doing on social:

• Keeping up with trends and cultural moments: 70% say excellent or good

• Entertaining content: 69%

• Truly original content: 56%

• Responding to crises: 51% — the weakest mark

Meanwhile marketers are pouring effort into speed on trends: 75% can get trend content approved within five hours. Only 30% say at least half their content is timely or tied to cultural moments — but among teams with 650K+ followers that rises to 44%, so the bigger the account, the more of it goes on trends.

But (!) trends are table stakes — consumers already think brands handle them fine. The gap is originality, which links straight back to the human-made point above, and crisis response, which is a process problem (who can post, how fast, with what sign-off), not a content problem.

If you have one extra hour a week, spend it on original work.

The customer service gap nobody is staffing

Sprout’s second-biggest consumer ask: personalized customer service is consumers’ #2 priority, and it doesn’t appear on marketers’ priority list at all.

Customer service is what users most want from brands on Reddit (21%) and WhatsApp (40%), and 18% of users say they turn to LinkedIn before any other network when they need help — another 18% say the same of X.

Tempting first thing? Fix this with an AI chatbot. Luckily I have a great Gartner’s 2026 customer survey I can reference:

• Only 27% of customers would try a chatbot again after a bad experience.

• 87% say it’s essential to have an option to reach a human.

• 53% say they’d consider switching to a competitor if they learned a company uses AI for customer service.

Reddit: support desk, not stage

Sprout’s Reddit data is short. Users want brands there for customer service (21%), then product information (16%). 14% don’t want brands on Reddit at all, and 54% are most likely to engage with user-generated content rather than brand posts.

Stacy’s Brain knowledge adds a reason to care about Reddit that Sprout doesn’t touch: AI search.

An independent study across 12 product categories found that the brands Reddit communities rate highest are, on average, the brands AI chatbots recommend most (rank correlation 0.554, falling to 0.430 when the brand list was built independently of the AI answers). Google’s AI Mode linked to Reddit in 70.8% of “is it worth it” and comparison questions.

I shared a detailed post on this in our Telegram community GTM in Public. If Reddit and AI citations are your topic of interest, you can find it there.

Read more on Telegram

So in my opinion, on Reddit, the value isn’t in what you post — it’s in what users say about you.

AI search optimization is the new black

Open any vendor outlook for 2026 and 2027 and the same line runs through all of them: AI answers are becoming a primary discovery surface, and brands need to be visible inside them.

The industry has converged on this, and Google perfectly pictures this trend.

The change in user behavior is real.

• In HubSpot’s 2026 marketing survey, 49% of marketers said search traffic to their site has already dropped because of AI answers (source).

• In 2025 Google’s AI Overviews appeared on roughly 31% of one tracked keyword set. A year later it is already 48% (source).

So it’s no wonder that 86% of US buy-side ad decision-makers said they are already changing, or plan to change, how they measure performance for the AI web (source). Similarweb, Kantar, and McKinsey have all shifted their main advisory and research focus to how buyer behavior is changing — and what should be done about it.

AI optimization necessity is now the consensus.
The problem sits one layer down though, in the advice.

Subscribe now

AI answer systems are being rebuilt faster than anyone can document them

In the space of a few months in 2026, ChatGPT changed how many background searches it runs per question, how it phrases them, how many sources it credits, and how often it goes back for a second round (source).

In August 2026 Reddit stopped being the #1 source for community discussions seeding aimed at increasing citations: Reddit’s share in ChatGPT Search fell by 86.4% inside just ten days (source).

Meanwhile, if you search “how to optimize for ChatGPT search” on Google, it still shows Reddit and Quora among the top results — based on a one-year-old article.

Let’s be realistic: a guide written against mid-2025 algorithm is describing a different AI engine.

Furthermore, most of these guides and articles are written by vendors — and if an article from 2024 was cited by AI, they’ll keep updating the publish date to stay relevant in search, creating a contextual trap for both LLMs and humans relying on their output.

This is one of the latest cases I encountered:

So what do we have left?

We need to apply filters and re-review what’s currently observable. That’s why we prepared this Q4 2026 snapshot — to give you an accurate picture of the algorithm you’re optimizing for.

LLMs have a shortlist before they search

When you ask ChatGPT a question, it does not pass your words to a search engine. It writes its own background queries first (called fan-out queries) and those queries decide what it will ever see.

The volume of that activity climbed sharply through 2026.

The average number of fan-out queries per prompt reached 7.5, against 2.17 in 2025. The share of prompts where the system goes back for a second round of searching rose from about 5% to 33.5%. At the same time the number of unique domains credited in a finished answer fell from 19 to 15 (source).

Meaning: The system reads considerably more and cites noticeably fewer.

The shape of those queries also changed.

ChatGPT now uses the site: operator in about 23% of its fan-out searches, up from 0.3%.

Instead of asking the open web who makes good SEO software, it frequently goes straight at a named domain to pull a specific fact — a price, a spec, an integration list. That’s why product pages have risen to 16.39% of all retrieved pages, overtaking listicles as the most-fetched page type (source).

Getting into background AI searches

The most essential question for AI optimization is understanding where the brand names in these queries come from — and the answer is not optimistic.

In testing across thirteen product categories, ChatGPT injected brand names into its own background searches in eleven of them, even though the user never typed those brands (source). This suggests the model first drafts a shortlist based on what it already associates with the category, then goes looking for evidence about the companies on that list.

That ordering appears to matter a great deal.

In one practitioner study, brands that appeared in the initial fan-out query were cited in the final answer roughly 33 times more often than brands that were merely fetched during the search (source).

I have to mention that the sample of this study is small — a two-day test on a single account, 57 conversations — so treat this study as directional. The mechanism behind it, however, is visible in every transcript: a brand absent from the opening query is rarely retrieved into the answer later.

What this means for a “normal” brand?

• The “page 1” achievement illusion
  You could rank #1 on Google for a high-value keyword, but if a user asks ChatGPT for recommendations, the AI will often generate a fan-out query targeting the 3 or 4 big brands it already knows. Your top-ranking page will sit there untouched because the AI never even searched for it.

• The trust gap
  LLMs are trained to avoid hallucinating and to provide safe, authoritative answers. They lean heavily on their pre-trained “parametric knowledge.” If your brand hasn’t been mentioned frequently alongside your target keywords across the web, the LLM won’t risk putting you in its initial fan-out query.

• Zero-click extraction
  When ChatGPT does fetch your page without naming you first, it’s usually just to scrape a factual detail (like a statistic or definition) to support a point about a bigger brand. You become the research source, but the household name gets the customer.

Bad news.
And this raises the importance of brand equity.

In fact, the shortlist of AI queries is assembled from the strength of association between a company and its category — based on what the model already knows. And that association is built mostly off your own properties.

If we check the latest numbers, 85% of brand mentions in AI search come from third-party sources. Of those third-party citations, nearly 90% come from listicles, comparison pages, and review sites. (source).

Link building and PR now have a new golden age forming.

Google ranking and AI visibility are losing correlation

For most AI assistants, the link between a site’s Google position and citation is now weak.

• When ChatGPT cites a webpage, nearly 90% of the time that page sits below position 21 in Google (source).

• Across a test of 15,000 long-tail queries run through ChatGPT, Gemini, Copilot and Perplexity, roughly 80% of cited links did not rank anywhere in Google for the original query (source).

Inside Google’s own AI Overviews, the relationship is tighter, though the trend points to lower overlap.

In August 2025, Ahrefs reported that AI Overview citations come from Google’s top 10 about 76% of the time (source).

Six months later, BrightEdge — tracking its own keyword set over a year — put the share of AI Overview citations that also rank in the organic top 10 at roughly 17%, with just over half appearing somewhere in the top 100 (source).

Where AI Overviews do appear, they’ve been reshaping the page and user behavior.

• They showed up on 48% of one monitored keyword set by early 2026 (source).

• 99.2% of the keywords that trigger them carry informational intent (source).

On click behavior, the strongest available evidence comes from Pew Research Center, which tracked the actual browsing of 900 US adults. Users clicked a link inside the summary itself on just 1% of visits.

This gap is huge — and the study dates back to 2025. I assume that today, the number of users clicking a link from search (when an AI overview is present) is even lower. AI section is getting taller, and users are getting more used to chatting with Google assistant to refine their request.

Latest AI citations insights

Furthermore, AI answers are very inconsistent

The main challenge with AI optimization is that optimization assumes a stable target. AI answers don’t currently provide one.

Only 30% of brands stay visible in an AI answer from one run to the next with an identical prompt, and only 20% remain visible across five consecutive runs. Meanwhile, 68% of brand mentions are unique to a single model, meaning a company that appears reliably in one assistant can be entirely absent from another (source).

Visibility also shifts by language and by use case, so a single blended mention rate conceals most of what’s happening (source).

The instability extends to the extracted sources as well.

When ChatGPT builds a site: query for a brand it does not know well, it may guess the domain and aim its search at an address the company doesn’t own, including parked domains (source).

The brand is on the shortlist, but the retrieval step goes to the wrong place, and whatever comes back is attributed to a company that never published it.

Your own pages can be working against you

One pattern from 2026 deserves separate attention because it inverts the intuition behind a common tactic.

When a brand publishes its own “best tools” or comparison listicle and that page is cited in a Google AI Overview, 70% of the time the publishing brand is left out of the answer. Instead, its page is used to recommend competitors. The page earns the citation; the competitors named on it collect the recommendation (ha-ha-ha).

Google catches up with the trend — but in a different way: AI Overviews now cite 38% fewer self-promotional “best of” listicles than they did in June 2026 (source).

The adjacent lesson concerns community platforms.

For a long time, Reddit was the standard answer to where AI systems learn what people actually prefer. Its share of ChatGPT Search citations held steady around 3.83%, then fell from 3.83% to 0.52% between 7 and 17 August 2026 — a drop of 86.4% in ten days (source).

Reddit still remains among the most-cited domains in absolute terms though, but the shift shows how quickly a platform’s weight inside these systems can be re-tuned.

How do we adapt: Brand equity seems to be the new cornerstone

Below are the three key trends I noted down for Q4 2026 AEO strategies.

• The retrieval step now begins with a shortlist the model writes itself, assembled from prior associations and dominated by what third parties have published about a category. So increase your focus on bran equity building.

• Ranking well on Google doesn’t translate into AI citations, so while SEO and AEO overlap, they are not the same thing.

• Some established tactics have inverted, with brand-owned comparison pages now frequently serving competitors. Mentions from third-party websites seem to be a more reliable tactic now.

What to do next: Stacy’s Brain playbook

To get AI to include your brand name in its initial background searches, you need to embed your brand into the LLM’s core understanding of your industry.

Read full playbook

1. From “Keywords” to “Entity co-occurrence”

LLMs learn by mapping connections between entities (e.g., “Brand X” ↔ “SEO software”). Your goal is to have your brand name appear as close as possible to your target category across a wide range of high-authority websites.

What we do:

Secure placements in third-party listicles, industry roundups, and comparison pages (e.g., “Top 10 alternatives to [Big Competitor]”). Classic SEO link building, with a bit of PR.

2. Flood digital PR and forums

LLMs rely heavily on structured and unstructured public discussions to understand what real humans prefer. This includes Reddit, Quora, industry-specific forums, and niche media sites.

What we do:

Step one: build team accounts on the relevant platforms and grow their karma and authority.

In 3–4 months, those accounts can start naturally discussing the brand in relevant threads, strengthening “community” consensus around it.

This is highly experimental, though. It’s important that you don’t just start “shilling”—you’ll lose both time and the account :)

3. Dominate category-defining content

If you can’t be the biggest brand in the world, be the absolute, undisputed authority in your specific sub-niche.

What we do:

• Create original research reports

• Produce proprietary data studies

• Publish highly comprehensive “ultimate guides”

When other websites reference your data and credit your brand, it creates a massive web of citations that teaches the LLM your brand is the “seed source” for that specific topic.

Another argument for founder-led growth.

4. Optimize for “Brand + Modifier” searches

Because ChatGPT often searches using the site: operator or specific brand names, make sure your own website makes it easy for the AI to extract data.

What we do:

Build dedicated, cleanly structured pages for intent-heavy searches like [Your Brand] vs [Competitor], [Your Brand] pricing, and [Your Brand] reviews. Use clean schema markup and clear headings.

Stacy’s Brain reads marketing research so you don’t have to, grades every source for methodology, sample and vendor interest. This is how we find evidence for our strategies.

Most numbers cited by LLMs and search aren’t backed by methodology or transparent sample disclosure — yet they drive marketing decisions and shape strategy. Ugly reality, isn’t it?

In this series, we’ll look at academic papers: an overlooked source of distilled marketing insight that comes with full methods and sample details — so we can build on evidence.

Subscribe now

Insight #1: Buyer confidence collapses in the middle of the journey

Lifecycle marketing usually assumes certainty accumulates: the more someone buys in a category, the surer they get.

A 2025 paper in the Journal of Consumer Research tested that against four separate review corpora:

• 1M+ wine notes from 30K CellarTracker users

• 1.4M+ BeerAdvocate reviews from 32K users

• Almost 1M RateBeer reviews from 19K users

• 229K Sephora reviews from almost 12K reviewers.

Buyers start sure on the first purchase, lose that certainty as they discover how much they did not know, and climb back as real experience accumulates. A controlled experiment with 801 participants reproduced the same dip.

The dip costs money.

When the researchers made people feel less certain, switching away to a different category rose from 43% to 53%.

One thing shortened the period of lower confidence though: people who reviewed products analytically — comparing features, naming attributes, and explaining their choice — recovered their confidence sooner.

Insight #2: Buyers hand judgement to AI and rarely audit it

Wharton ran three preregistered experiments with 1,372 participants working through reasoning problems with and without AI access.

Participants copied the AI’s answer 79.8% of the time, including the times it was wrong. Where the AI was wrong and the person used it anyway, 73.2% of those cases were verbatim copying with no attempt at correction (sad sigh).

Confidence moved the wrong way.

Access to AI lifted participants’ confidence by 11.7 percentage points while roughly half the model’s answers in the task were faulty. People who reported higher trust in AI were also 3.47 times more likely to follow its incorrect answers.

An accuracy incentive helped improve accuracy though. When users received an incentive with immediate feedback on whether their answer was correct, rejection of wrong AI advice more than doubled — to 42.3%.

Insight #3: AI is the second buyer present in the shopping flow

A working paper ran repeated simulated choice experiments aimed at understanding how AI chooses products to recommend.

The experiment was running across four models — GPT-4.1-mini, GPT-5, Gemini 2.5 Pro and Gemini 2.5 Flash Lite — over four product categories, targeting 1,000 choice iterations per product per model with eight product cards on screen each time.

• Ratings were the only promotional cue that helped every model in every category, 16 combinations out of 16.

• Social proof worked in 14 of 16.

• Higher prices consistently lowered the odds of selection for every model tested.

• Most models favoured the top-left product slot, with GPT-5 as the exception.

• Bundle badges backfired for GPT-5 in the phone category.

• Urgency, discount and scarcity badges mostly did nothing at all, and star score stood out as the clear exception.

Take notes for your shopping websites.

Insight #4: 1/4 users now see ChatGPT ads as they prompt

Alongside the model as shopper, the model also emerges as ad surface.

An August 2026 audit built 91 synthetic US ChatGPT accounts to understand how ChatGPT Ads are shown to users.

Those accounts were created in a 3×3 design across race or ethnicity signal and income tercile, ran 335 prompts and logged 127,801 conversations, capturing 3,602 ads from 191 advertisers.

• No account saw an ad during its first week, and ads began around day 14.

• Across the main window ads appeared in 5.61% of interactions

• Once an account started seeing ads, the typical exposed account saw one on 24% of later prompts, and the most heavily targeted account reached 37.1%.

• Lower-income accounts were more likely to receive ads, while race showed no significant link with delivery.

• Retail and information ads made up 57% of impressions, and the highest ad-triggering prompt in the set was a question about fixing a broken headlight, at 19.2%.

Insight #5: Word of mouth is a conversation

A 2026 paper in the Journal of the Association for Consumer Research recorded people who already knew each other talking: 192 participants across 96 pairs in the first study, 222 participants across 125 conversations in the second.

Their focus was to understand how people react to different length and linguistic structures of discussions, to understand what patters make people interested in a potential purchase.

• Longer product conversations were more persuasive and more enjoyable for the people having them.

• Concrete wording and narrow follow-up questions extended the topic.

Insight #6: Correlation of influencer engagement and number of paid posts

A multimethod study in the Journal of Marketing covered 6,869 sponsored Instagram posts from 4,908 influencers, 147,877 sponsored Douyin videos from 1,000 influencers, comment-text analysis across 2,015,022 comments, and two experiments with 168 Twitter and 208 Instagram users.

Their main goal was to understand if there is correlation between number of paid posts and influencer engagement. Common sense says sponsored content works best coming from creators who rarely take sponsorships. In fact, the data traces a U.

Engagement was highest for influencers with very low or very high sponsorship loads and worst for those in the middle, with people around twice as likely to share a sponsored Instagram post from a creator at either extreme.

Repeating the same brand shrinks the upside, and unpaid posts covering the same product type soften the penalty of a heavy ad mix.

Influencer marketing in 2026

If you ask about the reasoning behind influencer marketing, most LLM answers will revolve around the “friend” effect, high trust, content authenticity, improved brand awareness, and higher conversion rates compared to classic media buying.

All of this is relevant, but for 2024 — not 2026 — and as someone who actually runs an agency focused on influencer marketing, I probably shouldn’t tell you this (but I will, because I advocate for evidence-based marketing).

Here’s what we know about creator marketing in 2026:

• 82.7% of U.S. marketers use influencers in their creative marketing campaigns (source).

• Only 5% of U.S. consumers completely trust influencer advertising, 74% — trust to at least some extent, while 26% do not trust it at all (source).

• Still, 58% of consumers have made a purchase because of an influencer endorsement and 33% actively seek out influencer content to learn about products or topics (source).

• Influencer ROIs spread much wider than linear TV’s, with far fewer campaigns landing anywhere near the channel average (source).

• 24% of marketers say they are very confident in proving influencer ROI to the C-suite, 62% are somewhat confident, and 14% are not confident at all (source).

• 30% of marketers say wrong influencer selection is the main reason campaigns fail, but only 6% blame targeting and 3% blame lack of consumer research (source).

A few more, because the picture only gets stranger the longer you look at it:

• A net 61% of marketers plan to increase their creator investment in 2026, while only 27% of creator content ties strongly to the brand paying for it (source).

• 97% of marketers describe influencer targeting as inconsistent, immature, or essentially guessing (source).

• In the UK, influencer campaigns return a short-term ROI index of 99 — within a point of the average across all channels — while the long-term index is 151 (source).

And my favorite:

Mixed signal, right?

A channel that most of the market uses, that most consumers don’t fully trust but buy from anyway, and where the average campaign returns exactly what an average campaign returns in any other channel.

In fact, the efficiency of KOL marketing often depends on factors that are overlooked: tribe identification, social listening, consumer research, briefing, horizontal creator interaction, and even optimizing influencer content for LLMs and the attribution window.

So, let’s explore what the research says about the above.

Part 1: What we do wrong about influencer marketing

I won’t cover the obvious stuff — onboarding creators with fake metrics, or pushing everyone to use the same advertising copy. I hope that by 2026 these things are a given, and that the real barrier to efficient campaigns is far more complex.

There’s an insightful report on influencer marketing performance from Influencer Strategist, which surveyed 41 senior practitioners, 63% of them agencies and consultants — so people who actually sell creator marketing services.

Here’s what they consider to be the single most common reason influencer campaigns underperform:

The “wrong influencer” problem

In fact, wrong influencer selection is not a root cause; it’s an outcome.

If you haven’t mapped the target audience the influencer needs to reach, if you don’t know what that community discusses, values, and trusts, and if you don’t know what will convince them pay attention to your product — then influencer selection becomes a guess based on engagement rates and audience demographics.

Some teams are lucky to guess, but at least 1/3 are not. And when they are not, they say “wrong influencer” instead of “wrong targeting”, “wrong messaging”, or “wrong strategy”.

Influencer Strategists share some great supporting stats:

• 50% allocate nothing to consumer research before campaign execution.

• Only 20% use social listening data on what their target consumers actually discuss.

• Only 30% have tribe identification tech in their stack.

• 94% use engagement rate for shortlisting creators while only 25% use data on what consumers actually talk about.

Number two — poor creative or content quality — is also, in many ways, a symptom. The root cause may sit earlier in the workflow — creator shortlisting, briefing, or content review. In fact, only 40% evaluate content for resonance with the target consumer tribe during review.

It’s worth highlighting that the data I’m quoting comes from senior-level practitioners. For market averages, I’d expect even lower numbers — especially for teams running influencer marketing in-house and treating it as “we pay influencers, they post”.

That turns influencers into CPM-style media buying, and creator marketing loses on CPM by a huge margin.

The approach should be different: a longer preparation phase and deeper pre-campaign analysis.

Part 2: What top teams do to boost creator ROI

1. Tribe identification

Tribe identification means finding tight-knit groups of online communities (tribes) that share the exact same hobbies, values, or passions. It’s a much deeper level of segmentation than targeting by age, geo, or content focus.

Let’s say you’re building a service for creative writing that shows X’s social trends in your niche. With classic targeting, you might market to broad marketing or business communities.

With tribes, you’d instead target “ProductHunt solo founders” as one category, “marketing degree students graduating this year” as another, and “emerging blogger communities who struggle to reach 1,000 followers” as group number 3.

All of these are highly specific, allowing for more precise targeting and messaging — no wonder the interest in collaborations with micro- and nano-influencers is consistently growing (source).

How do you identify tribes?

Two options.

First: tools. This is the easiest path to follow — especially if you can afford to allocate additional budget to influencer marketing intelligence.

Some tools on my shortlist (in no specific order; do your own research, as they’re not universal across niches):

• Audiense

• HypeAuditor (more generalized)

• Tubular (video content intel)

Second: vibecoding + AI + in-house resources (slower, but better tailored)

If you choose this path, you’ll need to define the target user personas, map their keyword cloud, and set up social listening to identify relevant communities.

Some platforms already do this under the hood — X, for example. Their latest Radar lets you build a discovery terminal using specific keywords and filters, and it can serve as a solid first layer for creator discovery.

2. Social listening

In simple words, social listening is the strategic process of tracking, aggregating, and analyzing online conversations and sentiment across the internet to understand why people talk about a brand or industry, rather than just tallying direct mentions.

While basic monitoring tells you what is being said, social listening uncovers the underlying emotions, contextual shifts, and broader trends driving those conversations.

So, in fact, social listening supports a wider range of goals beyond shortlisting creators:

• Briefing: Source current pain points, trending topics, and messaging pillars from what users are actually saying.

• Campaign design: i.e. engage creators with polar opinions on a specific topic together to boost engagement.

• Optimisation: Detect early warning signs of dissatisfaction or concerns before a major issue escalates.

• Post-campaign analytics: Collect authentic, unprompted customer feedback to guide future feature rollouts or changes.

How popular is this tactic among top practitioners?
In fact, only 24% of marketers always run social listening before briefing influencers. The largest single group, 33%, runs it only sometimes, if time and budget allow (source).

Here’s an example taxonomy for a creator campaign, based on the case we discussed earlier — where we build a service for creative writing.

How do we do social listening?

Same as with tribes, there are two options — depending on your actual budgets and capacities.

Below is a list of tools mentioned by SproutSocial (one of the leaders in social listening). I’ll leave it here for your homework research.

3. Briefing and keyword strategy

How do most teams do creator briefs in 2026? If your answer is ChatGPT, you’re probably closest to the sad reality — that’s what we see when working with brands on their campaigns.

In the best possible case, the brief includes key messages, brand guidelines, DO’s and DONT’s, and (if the team is creative) a few potential content ideas.

Now, guess the single missing piece that only 13% of pro teams include: a structured keyword strategy covering both social search and AI discoverability (source).

This is important due to two reasons:

• Users increasingly treat TikTok, Instagram, and YouTube as search engines

• LLMs and AI search tools like ChatGPT and Perplexity also start relying on social data as user consensus

This aligns with an ROI trend highlighted by IPA: influencer campaigns deliver a short-term ROI index of 99 — within one point of the cross-channel average — while the long-term index rises to 151 (source).

Here’s a quick illustration of what I mean by keyword strategy, using an example of content centered on competitive brand comparison: [Brand A] vs [Brand B], alternatives to [Competitor], best for [Use case].

More things that increase the efficiency of KOL campaigns at the briefing stage

For this part, I selected a report from Kantar, who analysed more than 15,000 pieces of sponsored creator content across TikTok, Instagram, and YouTube.

• Saying the brand name aloud lifts branding by an average of 22 points.

• Showing the creator handling the product is both the most common lever in sponsored videos and the strongest single one for brand lift, at +15 percentile points.

• A logo placed away from the product lifts branding more than one placed on it — +20 against +16 percentile points.

• Describing what the product does beats describing how good it is. Product benefits appear in 71% of creator videos and lift brand power by 18 points; quality claims lift 10.

And there’s another pre-registered experiment I have to mention in this context which ran on 210 young adults in Vietnam: young adults rated a negative influencer review as clearly more helpful than a positive one about the same product (source). The same study found that a single negative review moved neither brand image nor purchase intention — and neither did a single positive one.

A few more recommendations from my practice at Green Dots:

• Multi-level briefings optimized for discovery, interest, and acquisition help boost end-of-funnel conversion.

• Hooks are one of the most overlooked parts of creator briefs. Most attention decisions are made in the first 1–2 seconds of interacting with content—so if a sponsored post can’t capture attention, the rest doesn’t matter. Delegating this entirely to creators (especially those with limited experience in efficient ad content) hurts efficiency.

• Ideally, briefs should be tailored to each creator, their tribe, and the influencer’s role in the campaign to maximize the value of each placement.

4. Creator strategy and horizontal creator interaction

Most KOL campaigns are isolated: Creator A never engages with sponsored content from Creator B — which, in my opinion, is one of the biggest gaps in influencer marketing — because creators continuously engage with each other outside of campaigns.

This is more about campaign architecture, and it’s one of the strongest short-term efficiency levers in our practice.

Let me quickly illustrate:

Imagine Creator A only posts your brand name. Nothing else — no CTA, no referral link.

Then Creator B quotes Creator A and adds his own argument (e.g., “Stacy’s Brain — because evidence-based marketing works”).

Creator C does the same, but his argument is: “Because AI answers are not reliable.”

Creators D–N follow the same pattern.

Meanwhile, your brand (well, mine in this example) reposts each argument to your profile, then later creates a thread compiling all of them — which creators then retweet.

This is what we call creative creator marketing. And this is just a quick example off the top of my head as I’m writing this; real-world cases are far more complex (and more interesting).

Checklist for your next influencer campaign

Before anyone gets shortlisted

• Write down the tribe. Age, geo and interest tags will not carry you here — name the community, what its members have in common, and where they already gather.

• Collect what that tribe actually says: the complaints, the comparisons it makes, the phrases it repeats.

• Run social listening before the brief is written. Only 24% of practitioners always do, and the ones who skip it are the ones who later report a “wrong influencer”.

• Decide what would have to be true for this campaign to have worked, and write it down before any money moves.

Choosing creators

• Start from overlap with the tribe you mapped. Engagement rate is a filter you apply afterwards to remove the obviously broken accounts.

• Check who the creator’s audience actually is against your tribe profile.

• Give every creator a job in the campaign — reach, credibility, demonstration, objection handling — and shortlist against that job.

• Where the tribe is narrow, several small creators inside it will usually beat one large creator sitting next to it.

Campaign architecture

• Decide how creators reference each other before the first post goes live. Isolated placements waste the thing creators already do with one another every day.

• Assign each creator a distinct argument so the set covers the objection map instead of repeating one message nine times.

• Plan what the brand account does with the output — reposts, quote replies, a thread that compiles the arguments — as part of the campaign, not as an afterthought.

The brief

• Write a separate brief per creator, built around their tribe and their role.

• Put a keyword strategy in it, covering both in-platform search and AI answers. 13% of professional teams do this today, which makes it close to free advantage.

• Specify the hook. The first two seconds decide whether the rest is seen, and handing that to a creator who has never written an ad is how the budget leaks.

• Ask for the brand name to be said out loud and for the product to be handled on camera. Also, keep the logo off the product itself.

• Brief what the product does rather than how good it is.

• Let creators name drawbacks in their own words.

Measurement

• Set the attribution window before launch, and set a long one. Short-term ROI for influencer sits at an index of 99, within a point of the cross-channel average, while the long-term index is 151. A 30-day window measures the half where nothing distinguishes the channel.

• Expect wide variance and judge the portfolio, not the placement. Influencer ROIs spread much further than linear TV’s, so one campaign tells you very little about the channel.

• Review the content against the tribe, not only against the brief. Only 40% check for resonance with the target community at review stage.

• Keep the comments. Unprompted feedback from a campaign is cheaper consumer research than anything you will commission.

• Record the tribe, the brief and the creative levers next to the result. Without that, the next campaign starts from the same guess.

Influencer marketing is far more complex — and execution-dependent — than other forms of paid advertising. It’s no surprise, then, that brands are split almost evenly on which approach drives the most ROI: Ambassador programs at 24%; UGC in paid ads at 23%; One-off campaigns at 22% (source).

Subscribe now

Keep learning, keep experimenting.
Yours,
Stacy

Introlude (before we start)

The CMO Survey, 2026 Topline Report is an online survey of 308 marketing leaders at US for-profit companies, fielded in January 2026. 97% of respondents were VP level or above.

One margin note you need before the numbers:
Everything below is self-reported. That means that when a CMO says AI improved sales productivity by 14%, they believe it; it’s not a fact.

Meet the beast

We start here, because it frames who we are comparing ourselves against.

97% of respondents we will be benchmarking against are VP or above, at established US firms, and 41.4% of them report that none of their sales happen online.

So when you read “you’re ahead of the average US company,” the honest version is: ahead of a large, established, largely offline American business. That’s still a useful mirror — a lot of the market you sell into looks exactly like this — and if your frontier is B2B, there’s no better lens to understand your customer. If not — well, you’ll at least understand how you’re different from established businesses!

Fact sheet: The “average US company”

1. The “average US company” is big and mostly offline.

2. They expect AI to do 56% of marketing work in three years. It does 24% today.

3. Traditional advertising is the only line they’re cutting — expected change of −1.5% against +10.4% for digital.

4. Their top three blockers are budget, system integration, and bandwidth — 20.1%, 19.1% and 14.1%.

5. When profits are down, 53.1% cut costs and only 26.8% push for revenue.

6. They are planning for a VERY long distance: 53.9% believe marketing’s influence on a customer lasts a year or more.

You vs. Beast: Where you’re genuinely ahead

These are the gaps that don’t depend on company size.

• AI adoption. They use AI or machine learning for 24.17% of marketing optimization and automation today, and expect 55.91% within three years. Generative AI specifically sits at 22.42%.

If you’re a solo operator or a one-person marketing team with a prompt library you passed their three-year target some time ago. So yeah — you’re well ahead of the market.

• Channel mix. Expected budget change for the next twelve months: digital +10.4%, overall marketing +7.61%, brand building +5.87%, and traditional advertising −1.5%. Traditional advertising is the only negative line in the entire list.

If you’re running a startup, you can make decisions fast. You can change your entire marketing strategy overnight — and launch a new set of creatives tomorrow. Huge advantage.

• Blockers. Asked what stops them getting value from marketing technology, they named lack of budget (20.1%), technology integration and data architecture (19.1%), and bandwidth, time and focus (14.1%).

The middle one — years of accumulated systems that don’t talk to each other — is a problem you do not have. Treat it as an advantage.

• What AI actually gets used for. Content creation 73.9%, content personalization 65.4%, and improving ROI by optimizing content and timing 49.5%. Nothing exotic.

If your AI use is more sophisticated than “write the thing, then rewrite it per segment,” you’re past the median.

• Default response to a bad quarter. When year-to-date profits come in under plan, executives prioritize cutting expenses 53.1% of the time and growing revenues 26.8%. Two to one, toward defense.

Startups don’t think in long timelines; they operate in short sprints, ready to endure several bad quarters in a row — because each bad quarter provides more data to finally make it. If the budget allows, of course.

Where the comparison breaks

We have to be honest in this part — everything below will look like a win in the numbers, but it isn’t. Still, it’s important to understand the market, right?

• Marketing spend share
  Marketing is 9.64% of total budget and 8.96% of revenue at these companies. At a business with hundreds of millions in revenue, that percentage funds a department, an agency roster, and a media plan. Your 30% of a much smaller number is not a win over their 9.6%.

  

• Where growth money goes
  56.44% of growth spending goes to existing products in existing markets, 22.89% to new products in existing markets, 14.14% to existing products in new markets, and only 6.53% to new products in new markets. That looks timid until you remember they have an installed base to defend. You don’t have one yet — which is freedom, but not superiority.

  

The one thing they have that you don’t

Asked how long marketing’s influence on a customer lasts, 24.4% said several months, 21.8% said a year, 18.6% said two years, 9.0% said five, and 4.5% said ten or more. Add up everything from a year upward and it is 53.9% — over half the room planning on a horizon most small companies never use.

That isn’t corporate slowness. It’s what you can do when you’re certain you’ll still be here.

The equivalent move for a small company is choosing one asset — a newsletter, a dataset, a founder profile on X — and refusing to judge it on a 30-day window.

It’s the one habit on this list worth copying upward.

Thank you for reading!
Stacy’s Brain reads marketing research so you don’t have to, grades every source for methodology, sample and vendor interest, and tells you how to use each number.

TL;DR: Five key insights

1. Meta is still the default channel. 91% of advertisers run paid social on Meta alone.

2. TikTok buys attention and clicks more cheaply. Its CPM is 3–4x lower across all campaigns, and a click costs about half as much.

3. Meta wins on engagement. It earns ~30% more interactions per $1 than TikTok.

4. For leads, TikTok looks about 2x cheaper, but test it. Cost per lead was $2.69 vs $6.00 (nobody measured lead quality though).

5. For sales, Meta is the proven option, but returns are shrinking. ROAS fell from 5.48 to 4.57 in a year, and the biggest spenders get the lowest return.

Deeper analysis below ↓

Marketers keep naming Instagram and TikTok among their best channels for ROI. Few studies go further and look at how campaigns actually perform, objective by objective, on a large sample. We found one that does.

At Stacy’s Brain we read marketing research so you don’t have to.

We don’t only pull out the numbers: we grade every source for methodology, sample and vendor interest, and tell you how far to trust each figure. That grading is what lets you make evidence-based decisions and grow efficiently.

Subscribe now

The sources, and how much we can trust them

• Social Ads Report 2026, Metricool.
  Observed campaign results on Meta and TikTok from 44,355 advertisers, H1 2025 vs H1 2026.
  Vendor interest: high (Metricool sells a cross-platform ads tool).
  How we use it: as the backbone of the funnel comparison, trusting the ratios between platforms more than the absolute numbers.

• Instagram Study, Metricool.
  Observed organic post performance across Metricool user accounts.
  Vendor interest: high.
  How we use it: to choose organic formats.

• 2026 State of Marketing, HubSpot.
  A survey of what marketers say works.
  Vendor interest: moderate (HubSpot sells marketing software).
  How we use it: to show what the market believes.

• CMO Spend Survey, Gartner.
  A survey of CMO budgets.
  Vendor interest: moderate (Gartner sells advisory).
  How we use it: for budget context.

• TikTok Ads Benchmarks, Triple Whale.
  Median ad costs across 5,900+ brands, Aug 2025–Jul 2026.
  Vendor interest: high (Triple Whale sells ecommerce analytics).
  How we use it: as a sanity check on Metricool’s costs.

In short, none of these sources is independent. Two of them measure what advertisers actually got (Metricool, Triple Whale), and two measure what marketers believe or budget (HubSpot, Gartner).

We treat those two kinds of evidence differently — and we’re here to compare them.

What marketers believe

HubSpot’s 2026 survey says Instagram overtook Facebook as the social platform with the highest ROI, and that TikTok’s perceived ROI rose sharply too. Overall, the perceived ROI of short-form video nearly doubled in a single year.

Note the word perceived.

These are marketers rating their own channels, not measured returns. Beliefs drive budgets, and budgets are tight: Gartner puts average marketing budgets at 7.7% of company revenue in 2025, down from 7.9%.

When money is a limited resource, betting it on a belief is expensive. So let’s see what the campaign data says.

Reality check #1: almost everyone is still on Meta

91% of advertisers in the Metricool dataset ran paid social only on Meta. About 4% used both platforms and 5% used only TikTok.

TikTok is growing inside that small group: among advertisers using both platforms, TikTok’s share of combined spend rose from 23.6% to 25.7% in a year. The smallest advertisers (under $100 a month) are the boldest, giving TikTok 44% of their budget.

Reality check #2: the numbers that look too good to be true

According to Metricool, Meta awareness campaigns cost $0.48 per 1K impressions (CPM) and Meta traffic campaigns cost $0.07 per click.

Those numbers are far below anything else we’ve seen.

Triple Whale, tracking 5,900+ mostly ecommerce brands, reports a median Meta CPM of $15.06 and a median TikTok CPM of $4.08. Even Metricool’s own overall Meta CPM across all campaign types is $2.64, five times the awareness figure.

The numbers aren’t made up though; they add up inside the report. Spend divided by impressions gives exactly $0.48. But they probably reflect who is in the sample and how it is averaged:

• A geography skewed to cheaper markets. Metricool is a Spanish company, and Latin American markets feature heavily in its data. This is our hypothesis; the report doesn’t publish a country mix.

• Objectives optimised for cheap impressions. Awareness and Reach campaigns are built to buy impressions cheaply, which pulls CPM down.

The good news (!) is that the two datasets disagree on levels but agree on direction. Across all campaigns, TikTok’s CPM is about a third of Meta’s in Metricool ($0.90 vs $2.64) and about a quarter of Meta’s in Triple Whale ($4.08 vs $15.06).

Funnel stage by stage

Awareness: TikTok buys attention more cheaply

• TikTok Reach campaigns had a 22% lower CPM than Meta Awareness campaigns ($0.38 vs $0.48).

• Advertisers spent 73% more per campaign on TikTok and got 2.2x more impressions.

• CPM rose on both platforms year over year: 8% on Meta and 5.5% on TikTok.

Traffic: TikTok again, by a wide margin

• TikTok traffic campaigns paid less than half per click ($0.03 vs $0.07).

• They got about 7.7x more clicks per campaign, on roughly 3x the budget.

• Clicks got more expensive on both platforms: CPC rose 29% on TikTok and 15% on Meta.

Engagement: Meta’s one clear win

• Meta engagement campaigns reached a slightly higher engagement rate (2.15% vs 1.99%).

• With about a third of TikTok’s budget per campaign, Meta earned 10.8 interactions per dollar vs 8.2 on TikTok.

Leads: TikTok looks stronger, but check the leads

• TikTok lead-gen campaigns paid $2.69 per lead vs $6.00 on Meta.

• TikTok’s cost per lead fell 33% year over year, while Meta’s rose 14%.

• Spend per campaign was almost identical (~$1,550). TikTok produced 2.2x more leads and converted about 6.2 leads per 100 clicks vs 4.7 on Meta, even though Meta’s click-through rate was higher (2.04% vs 0.75%).

Sales: Meta is the proven option, TikTok is unproven

Metricool reports a cost per purchase of $1.14 on TikTok vs $11.14 on Meta. You can ignore it.

The TikTok number comes from only 297 campaigns, down from 1,093 a year earlier, and the accounts still running them are probably the ones for whom it already works. Metricool itself calls it “a promising signal, not a clear win”.

Meanwhile, Meta’s figure is based on 67,025 campaigns.

The Meta number that matters is return on ad spend (ROAS). Meta sales campaigns returned $4.57 per $1 spent in 2026, down from $5.48. Advertisers more than doubled spend ($62.9M → $151.8M), but purchase value only roughly doubled.

ROAS fell in every size tier, and the biggest spenders had the lowest ROAS in both years (4.88 → 3.97), below even the smallest accounts. Spending more per dollar did not make the dollars work harder.

Where the money actually goes

Advertisers run most campaigns in the middle of the funnel but spend most of their money at the bottom:

• Meta: 67% of campaigns target traffic or engagement, yet 70.9% of spend goes to leads, sales and app promotion. Sales alone takes 42% of spend with 11% of campaigns, and awareness gets only 5.8%.

• TikTok: the split is more even, but 58.2% of spend still goes to the bottom of the funnel. One in four TikTok campaigns is optimised for video views.

That is rational: the bottom of the funnel is where revenue shows up. It also means most budgets are fighting over the same last click, while awareness, the stage where TikTok is cheapest, is underfunded.

Country matters more than platform averages

Meta sales ROAS swings widely depending on where a brand is based:

Before you panic about your US ROAS: markets with weaker purchase tracking can look better or worse than they are. The report itself warns that lower tracking coverage understates ROAS.

The organic layer: which format where

If you post organically, not only buy ads, Metricool’s Instagram study adds a few useful signals:

• Video reach favours TikTok.
  Instagram Reels got 29.9% fewer views and 41.9% less reach than TikTok videos.

• Carousels favour Instagram.
  Instagram carousels got 372% more views than TikTok carousels, and 142% more reach than single-image posts on Instagram.

• Hashtags didn’t help.
  Instagram posts with hashtags got 32% fewer views.

• Asking works.
  Posts with a call to action got 203% more comments, and posts that ask a question averaged 55.7 comments vs 40.7.

Recommended read:

The playbook

• Reach and awareness → TikTok
  Expect CPM about 20% lower on like-for-like objectives, and up to 3–4x lower overall. Confidence: medium, since two datasets agree.

• Website traffic → TikTok
  Expect a CPC about half of Meta’s and far more clicks per campaign. Confidence: low to medium, one source.

• Engagement and community → Meta
  Expect about 30% more interactions per dollar. Confidence: medium.

• Lead generation → test both
  Expect TikTok’s CPL to be about half of Meta’s, with lead quality unknown. Confidence: low to medium.

• Sales → Meta
  Expect ROAS to keep falling year over year, and to fall further as you scale. Confidence: medium.

• Sales on TikTok → a small test only
  There is no reliable benchmark yet (297 campaigns). Confidence: low.

• Organic video → TikTok
  Expect more views and reach than Reels. Confidence: medium-low.

• Organic carousels → Instagram
  Expect many more views than TikTok carousels. Confidence: medium-low.

How to test before you commit

1. Pick one goal per test
   Comparing a TikTok traffic campaign with a Meta sales campaign tells you nothing, because cost per result depends on the objective.

2. Split a fixed budget 50/50 for 4 weeks
   Use the same offer and the same creative idea, adapted natively to each platform.

3. Measure one step past the platform’s number
   For traffic, measure on-site conversion.
   For leads, measure qualified leads or booked calls.
   For sales, use your own revenue data, not platform-reported ROAS.

4. Benchmark against your market, not the global average
   Use the country table above as a rough guide.

5. Scale carefully
   The biggest Meta spenders get the lowest ROAS.
   Raise budgets step by step and watch marginal returns.

What we’d expect, and what would change our mind

If you’re an early-stage brand with a small budget, the evidence points to a simple split: TikTok for cheap attention, traffic and lead tests; Meta for engagement and proven sales. Expect TikTok to win on cost per impression and click, and Meta to win on predictability at the bottom of the funnel.

Stacy’s Brain grades marketing research so your budget decisions rest on evidence, not hype. Every number in this article links back to a graded source in our library that will go live soon.

Stay tuned!

Subscribe now

Zero budgets vs. Free acquisition

Growing with a $0 paid marketing budget is possible, but it does not mean customer acquisition is free.

A company that does not pay for advertising replaces financial spend with other resources (which are often even more scarce and valuable):

• founder time and expertise

• side-product development

• lots of BD work

And almost always it also comes with slower growth.

However, most founders and teams in 2026 have no alternatives: in many verticals, the unit economics simply don’t work once you apply modern benchmarks.

For retail fintech apps, a customer now costs ~$160. The global average cost per install (CPI) on iOS is ~$6. Even a simple follower from Instagram Ads will cost you ~$0.50–$1.50.

So yeah, most businesses are forced to operate with close to zero marketing spend — not because they don’t have money, but because paid traffic is too expensive.

This creates a growth loop instead of requiring the company to manually acquire every customer — and that’s what we will be exploring today.

Subscribe now

TL;DR for you

1. Use AI to reduce execution costs and produce higher quality things (not more of them)

2. Build founder persona. It’s one of the best investments in 2026.

3. Create content with long-term discovery value

4. Build distribution into the product (collaboration loops and casual contact loops)

5. Build a free tool around the core product

6. Create UGC mechanics

7. Replace paid acquisition with manual distribution

1. Use AI to reduce execution costs

AI makes zero-budget growth more practical because small teams can do work that previously required more headcount.

Below are a few use cases that are extremely helpful in the context of zero-budget marketing:

• Building free-to-use products with high value (e.g., calculators, comparison tools, databases)

• Many parts of SEO / AEO optimization can be done with AI

• Increasing content production speed (note: AI should increase the quantity of creative options and ideas; never let AI write reader-facing content end-to-end)

• Interpreting user behavior data and social signals

These are practical “come and grab” use cases — yet most teams limit themselves to content, which is a dead-end. Users are exhausted with AI content and want to hear humans.

So AI does not replace expertise or strategy. It reduces the cost and time required to execute and test ideas before they go live.

2. Use the founder as a distribution channel

Founder-led growth works especially well for businesses where trust and expertise influence the purchase decision:

• B2B

• SaaS

• Consulting

• Agencies

• Financial products

• New or technically complex products

Instead of paying for access to an audience, the founder builds direct distribution through LinkedIn, X, YouTube, newsletters, or other relevant channels.

Does it take long? Oh, yeah.

I personally reached my first 1K followers on X after 4 months back in 2023. These days it’s even more challenging (competition, AI automations, etc.), so I’d say it takes 5–6 months of consistent execution to build a founder persona that actively converts attention into revenue or pipeline.

And it’s not just about content and consistent posting. Other pillars — active networking, collaborations, and iterative feedback — are equally important.

Example: Marc Lou

Pieter Levels grew products such as TrustMRR and DataFast largely through his own audience.

Yeah, he’s not a newcomer. In fact, he was one of the early “founders” of the build-in-public movement, which is mainstream these days. Reaching his scale and impact takes years—not months. He’s a huge example of how powerful a strong founder brand can be.

3. Create content with long-term discovery value

In this section, I want to start with an example.

If you search for “Benchmarks for user acquisition in 2026 by industry,” you’ll end up on the same page everyone (even giants like HubSpot) is quoting. It’s a page from an SEO agency that published its benchmark lists in 2024 (so they’re not even from 2026). That page got cited, and they’ve just kept updating the year on it.

It has no methodology, no sample, and no data source.

Simply put, there could be any random number in those tables—and people would still quote them and benchmark themselves against those numbers.

Anecdote, right?

In fact, almost every LLM and related article will trace back...

This is a perfect example of why evergreen content is extremely efficient.

The marketing objective here is to create content around questions the target customer is likely to continue asking over time.

Examples:

• case studies

• original research reports

• practitioner notes

• tutorials

This creates a different type of return from short-lived engagement content, and it is a must-add for your SMM and founder growth strategy.

PS: Don’t forget to optimize for LLM search visibility though!
PPS: Ahrefs and similar tools will help you choose the right topics.

4. Build distribution into the product

The strongest zero-budget mechanism is product-led distribution.

This happens when normal product usage exposes another potential user to the product.

Examples include:

• a Typeform link

• a Calendly booking page

• a Figma file

• a Loom recording

• a shared Notion document

Here, the user is not actively promoting the product and distribution happens because sharing is part of completing the task.

This is often described as a casual contact loop.

Example: FlowMapp

FlowMapp grew to more than 180K users across 150 countries while operating with essentially no paid marketing budget.

The product naturally supports external sharing: A designer creates a sitemap or prototype and sends it to a client or teammate. The recipient is then exposed to the product and can enter its freemium funnel.

The acquisition mechanism is therefore built into the workflow.

Collaborative loops

I mention collaborative loops in the same context as they are a stronger version of product-led distribution.

The product becomes more useful when additional people join to collaborate.

Examples: Figma, Notion, Miro, all types of task trackers.

The causal mechanism is:

1. One person starts using the product.

2. They need another person to participate.

3. They invite that person.

4. The second person becomes a user.

5. The product becomes more valuable to both.

Perfect for team-based B2B products.

5. Build a free tool around the core product

If the main product is difficult to discover or explain, a company can create a smaller free product around the same customer problem.

You definitely saw these examples:

• Templates

• ROI calculators

• Paper trading tools

• Free databases and diagnostic tools

This is usually called side-project marketing.

The mechanism is:

1. The free tool solves a narrow, immediate problem.

2. Users discover it through search, communities, or sharing.

3. Some users also discover the company’s main product.

This is particularly useful for complex products where demand for the category itself may be limited.

6. Create user-generated distribution

UGC becomes a growth mechanism when the product gives users something they naturally want to share.

Let’s consider Spotify Wrapped as an example.

They create personalized data about the user’s listening behavior. Users share it because the output communicates something about their identity.

Duolingo applies a similar principle through streaks, usage statistics, and progress.

The mechanism is:

1. The product creates a personalized or socially valuable output.

2. The user shares it.

3. Other users see the product.

4. Some of them become users themselves.

Successful sharing mechanics normally provide value to the user, not only to the company.

Typical motivations include social identity, status, achievement, expertise, etc. — meaning something a user would like to show off.

7. Replace paid acquisition with manual distribution

There’s a fresh study that tracked six months of unpaid social media growth for Birla Open Minds International School.

Their strategy included zero-cost marketing:

• hyper-local community engagement

• educational events

• student-generated content

As a result, their Facebook page generated more than 122,000 views without paid promotion.

What they did? Nothing extraordinary; they just replaced paid acquisition with human effort and local distribution.

There are types of businesses cannot build product loops immediately. This is common for local businesses, professional services, early-stage startups — a classic cold start problem.

In these cases, the founder or team can replace advertising spend with manual distribution:

• participating in interviews and AMAs

• direct outreach

• inviting people for customer interviews (yeah, 90% will ignore you but there are still 10% who will reply)

• active participation in relevant communities as a talking head

• collaborations with other local or niche brands

For local and niche businesses, cross-promotion can also replace part of paid acquisition.

Examples:

• joint events

• newsletter swaps

• shared educational sessions

• product bundles

• creator or community partnerships

The mechanism is audience overlap.

Two businesses serving similar customers but selling different products can expose each other to existing audiences without paying for media. This is most effective when the audiences are relevant but the businesses are not direct competitors.

Zero-budget journey is tough

Here are a few things to help you stay focused and direct your limited resources in the right way during your journey:

Focus on one or two acquisition channels

Zero-budget companies have a limited amount of founder and team time.

Trying to maintain five social platforms at once usually creates shallow execution across all of them. A more efficient model is one primary discovery channel + one owned or semi-owned retention channel.

For example: LinkedIn + Substack / YouTube + Discord / X + Telegram.

The goal is to build enough density in one channel for distribution to compound before adding more.

Measure product response, not only traffic

Zero-budget growth can create an illusion of progress because activity is easy to measure.

Posts published, impressions, followers, and website visits do not confirm that the product has demand, and many founders get trapped spending years on dead ideas.

So here are some important metrics you still have to track:

• Activation
  If users register but do not complete the key action, acquisition is not your main problem.

• Retention
  Do users return? For a zero-budget model, weak retention is particularly damaging because every new user requires time or organic distribution to acquire.

• Organic acquisition

  Are new users arriving without direct intervention from the founder? This shows whether the distribution mechanism is beginning to work independently.

• Referral behavior

  This can be measured through referral rate, invitations, shared links, collaborative usage, or viral coefficient.

• Inbound demand

  Does content generate inbound leads, demo requests, partnerships, customer conversations, etc. This is more useful than measuring engagement alone.

Validate manually before waiting for organic growth

The main risk of zero-budget marketing is slow feedback.

Paid acquisition can generate market feedback quickly — but organic growth takes longer, which makes it easier to spend months on a product with weak demand.

The first stage should therefore rely on manual validation.

During the first weeks, founders can (and should) use direct outreach, customer interviews, manual onboarding and direct product demos to understand if there’s a fit.

Only after this signal exists does it make sense to invest heavily in content, SEO, product loops, or marketplace distribution.

Expect slower growth — it’s fine

Paid marketing primarily buys speed. It allows a company to generate traffic and test acquisition at scale immediately.

Without paid distribution, growth depends on mechanisms that take longer to compound like search rankings, founder audience, referrals.

This means zero-budget growth normally requires more time before acquisition becomes predictable.

The trade-off is less financial risk, but slower feedback and slower scale.

A practical $0 growth framework

So now let us simplify the sequence to five stages.

Stage 1: Validate manually

Confirm that real users understand the problem and want the product.

Stage 2: Identify the natural distribution mechanism

Determine whether growth can come from:

— product sharing mechanics
— founder content
— search
— UGC
— free tools you develop
— partnerships

Stage 3: Build the loop

Connect usage or content to the acquisition of the next user.

Stage 4: Measure activation and retention

Confirm that acquired users actually receive value and return.

Stage 5: Add paid marketing only after the organic mechanism works

Paid acquisition can then increase the speed of a system that already converts and retains users.

So yeah — $0 marketing is not a set of free promotional tactics. It is the design of acquisition mechanisms that use product behavior, founder expertise, content, communities, and existing platforms instead of paid media.

Five key facts to grab your attention first

1. Only about 5% of Meta ads become winners. Roughly half get little or no spend at all.

2. Volume, not intuition, is what finds winners. Advertisers testing 2.8 ads a week hit 4.03%; those testing 18.85 a week hit 8.80%.

3. Cheap formats win more often than expensive ones. Text-only assets had the highest hit rate (11.6%) and took nearly twice their share of budget (1.97x), while high-production assets took less than their share (0.79x).

4. The format that travels best across industries is the plain “letter” ad — it lands in the top 10 by hit rate in 10 of 15 industries.

5. Industry beats benchmark. Street interviews hit 30% in automotive and don’t make the top 10 in most other industries. Industry-specific insights are the key.

Now, let’s start our creative journey

What this study is built on:

• Creative Benchmarks 2026, Motion — 550K+ Facebook and Instagram ads from 6K+ advertisers, about $1.3B in spend, launched between September 2025 and early January 2026.
  Vendor interest: high. Motion sells creative analytics, and the report’s conclusion (”test more”) is also its sales pitch.

• ViASNet: A Video Ad Saliency Network (arXiv, May 2026) — 151 video ads, each watched by about 20 people with eye tracking.
  Academic, no vendor interest, but small sample.

• The 2026 State of Marketing, HubSpot — survey of 1,500+ marketers.
  Vendor interest: high. I use it only to analyze what marketers believe.

A couple of side notes so you can interpret the data below

Motion calls an ad a winner when it spent at least 10x its account’s median ad spend. In other words a “winner” here means an ad the platform spent heavily on, not an ad that made money.

Hit rate is the share of ads in a group that became winners. So “high hit rate” reads as: when advertisers made this kind of ad, the it got scaled more often than average.

The uncomfortable truth about base rate

About 5% of ads reach winner status, and roughly half get little or no spend.

The share of winners also rises with budget: 3.7% at micro advertisers (under $10K a month) against 8.2% at enterprises ($1M+). The interpretation of “big teams make better ads” is actually misleading; the data suggests something duller:

• Micro advertisers test 2.8 ads a week and hit 4.03%. Enterprises test 18.85 a week and hit 8.80%.

• Enterprises put 63.7% of monthly spend behind winning ads. Micro advertisers put 23%, and 31.5% of their spend goes to ads classed as losers.

More tests mean more chances for the algorithm to find something to scale — and a bigger share of budget ends up behind the ads that work.

Visual formats: what gets scaled

Across all industries, the highest hit rates by visual format were:

• Unboxing — 9.83% (1,229 winners)

• Offer-first banner — 8.68% (11,006 winners, by far the most in absolute terms)

• Behind the scenes — 8.64%

• Founder — 8.57%

• POV — 8.28%

• Demo — 8.11%

Two highlights:

• Offer-first banners and demos are used constantly, and they produce most winners in absolute numbers.

• Unboxing, behind-the-scenes and founder ads are the high-percentage plays: used far less, but they scale more often per attempt.

Though, budget follows a different logic. It is measured by spend use ratio — share of budget divided by share of ads.

In other words:
Imagine you have $10K and launch 100 different creatives. Five of those creatives are text-only and get $1K of your budget. That means their spend ratio is 10% / 2% = 5x.

The formats that got disproportionately more money were celebrity (2.17x) and letter-style ads (1.97x), while offer-first banners took only 1.34x.

Asset types: production complexity doesn’t matter

In our opinion this is the most actionable part of the report, because it is the part that costs money. Let’s see what types of creatives got the best hit rate:

• Text only — 11.6% hit rate, the highest of any asset type, from just 2.57% of creatives.

• Product image with text — 8.75%.

• UGC — 7.56%, and the most winners overall (11,374), because it is also the most used asset type (26% of creatives).

• High production — 6.97%.

What does that mean?
Text-only assets took 1.97x their share of spend, product images with text 1.23x, UGC exactly 1.00x, and high production 0.79x — less than its share of creatives.

Margin note: in the report we’re reviewing text-only is 2.57% of all creatives. Since it’s used rarely it can look great on a rate and still be fragile at scale. Treat it as a cheap bet worth testing though.

Hooks: what you say in the first line

Highest hit rates by hook headlines:

• Newness — 11.37% (83 winners)

• Sale announcement — 11.35% (16 winners — based on a very small base)

• Price anchor — 10.89% (71 winners)

• Urgency — 9.73% (4,855 winners)

• Offer only — 9.29% (4,798 winners)

• FOMO — 9.20%, confession — 8.74%, exclusivity — 8.44%

And the tactics that produced the most winners in absolute terms: urgency (4,855), offer only (4,798), relatability (3,377), curiosity (2,402), and bold claim (2,308).

Notice which hooks sit near the bottom on rate while still being everywhere: relatability 6.85%, contrast 6.78%, authority 6.40%. These are the “brand voice” hooks. They are used heavily and scale less often than others.

The details: What are the winning creatives by industry?

Now you know which creative types worked well for an average brand — but there’s no guarantee they’ll be top performers in your industry.

Luckily, we have some data from the report. Here are a few examples of the top formats by hit rate:

• Automotive: street interview — 30.0%

• Fashion & apparel: post it — 27.4%, quiz — 25.99%

• Pets: nostalgia — 26.09%, unconventional text placement — 20.93%

• Parenting & family: unboxing — 25.53%, expert explainer — 25.0%

• Home & lifestyle: reaction video — 22.54%

• Entertainment & media: how-to — 19.51%

• Professional services: notes app — 16.67%, grid swap — 16.07%

• Technology: letter — 17.57%, slideshow — 15.22%

• Beauty & personal care: trend — 12.32%, letter — 12.27%

None of these leaders appear in the global top six. If you take “unboxing and offer-first banners win” from the headline and apply it to a B2B services account, you will most likely miss it.

One format does travel from one industry to another though — the letter. A plain, text-forward ad written like a note — lands in the top 10 by hit rate in 10 of the 15 industries, and takes an outsized share of budget in beauty (2.00x), finance (3.00x) and health & wellness (3.40x).

If you want one cheap format to test everywhere, that is the candidate the data supports.

Great study overall – but we have to highlight:

• A winner is not a profitable ad. It is an ad the platform kept funding. Nobody here measured ROAS, margin or return rates.

• The window is short and seasonal. Four months across the holiday peak.

• The publisher sells creative analytics.

• Some rates derive from tiny data samples. For example, a sale-announcement hook at 11.35% comes from 16 winning ads.

• Industry tables have thinner data than the global one — treat a single striking number (street interviews at 30%) as a prompt to test, not a fact about your market.

• Everything here is correlation. No one ran an experiment where the same offer was shot as five formats and rotated at random.

What the one non-vendor source adds

ViASNet is the only source here with no commercial stake — and that’s the reason why we want to add it here.

The researchers tracked the eyes of about 20 viewers each across 151 video ads and modelled where attention goes. One of their findings: scene cuts cause a spike in attention entropy — viewer focus scatters after each cut and has to be regained.

Why do we highlight this?

This finding gives a mechanism for what the Motion data shows from the outside: formats like text-only, letters and founder talking-head ads have very few cuts and one clear focal point. Nothing has to be re-hooked.

The playbook for winning creatives

• Set a testing quota before you argue about creative.
  Match the tier you are in: under $10K a month means roughly 3 new ads a week, $10–50K means 4–5, and above that scale up towards 7+. Volume is the variable with the clearest link to finding winners.

• Build a cheap-format core.
  Text-only cards, product images with text, letter-style ads. They are the assets with the best hit rates and the fastest turnaround, which is probably the same fact twice.

• Keep one expensive format, not four.
  High production had a below-average hit rate and took less budget than its share. It can earn its place, but it should be a deliberate bet.

• Lead with a concrete offer, newness or a price anchor in the first line, and treat relatability or authority hooks as brand support, not as your scaling engine.

• Pull your industry’s table before you commit.
  If your category’s top format is a street interview or a notes-app screenshot, that is a cheaper test than anything in the global list.

• Cut the clutter in video.
  Fewer cuts, one focal point, the claim early. The eye-tracking evidence says every cut costs you attention you then have to win back.

• Judge on your own numbers.
  Track your own hit rate (share of new ads that pass 10x your median spend) and your own cost per purchase. The benchmark is for direction; your account is the evidence.

Stacy’s Brain reads marketing research so you don’t have to, grades every source for methodology, sample and vendor interest, and tells you how far to trust each number.

See you in the next episode!

In 2026 LinkedIn Ads B2B Benchmarks Report, Dreamdata shares aggregated data from thousands of B2B customers, covering 66M+ sessions and 3.5M+ customer journeys. If you’re part of a B2B company and came across this report, you’ll walk away with just one conclusion: LinkedIn is where I should invest money.

But wait.
It’s just one dataset from one vendor.

We checked each claim in this report against other vendors focused on the same channels, as well as independent research, to understand how much we can trust this source, and whether we should act on it.

Spoiler alert: Some of the numbers hold up under the pressure of comparison. The headline ROAS figure does not.

Subscribe now

Let’s look at the source first

• Methodology disclosed: yes.

• Sample disclosed: yes.

• Vendor interest: high.
  Dreamdata sells B2B attribution and activation software, and in fact every finding here argues for more attribution and more LinkedIn.

Two structural limits on the sample we need to know about:

• Only customers who agreed to benchmark use are included. These are companies that already bought an attribution platform, which means already spending on LinkedIn.

• Dreamdata’s new IP-to-company engine, which identifies up to 15 times more companies than before, and LinkedIn’s Company Intelligence API were released within the comparison window — so the factors aren’t isolated.

Claim 1: LinkedIn has the highest ROAS

The reported figures for 2025:

• LinkedIn 121%

• Google Search 67%

• Meta 51%

Here’s what that means: Two of the three channels return less money than they cost.

Plus (!) the winner clears its own media cost only by 21% BEFORE we deduct cost of goods, sales salaries, or the software the deal runs on. At a typical B2B SaaS gross margin, 121% ROAS = loss.

Though, B2B revenue is usually multi-year while the research looks into 12 months only, and we have to count this in as well (not just for LinkedIn; for Google and Meta as well).

What does not survive is treating these findings as a universal benchmark.

There is a canonic study from 2015: Blake, Nosko and Tadelis (2015) in Econometrica, who switched eBay’s paid search off across randomly selected US regions for 60 days.

Results:

• On branded keywords, 99.5% of the click traffic lost to disabled ads came straight back through organic search.

• On non-branded keywords the experimental return was −63%.

What does this give us?

Their own conclusion: “typical regressions of sales on advertising spend result in astronomical ROI estimates that vastly overestimate the true ROI, which can only be estimated using controlled experiments.”

Claim 2: LinkedIn is the cheapest way to reach a company

Next benchmarks from the report:

• Cost per company influenced, 2025: LinkedIn €70.11, Google Search €110.37, Meta €128.70.

• Cost per contact influenced, same table: LinkedIn €1,033.29, Meta €231.77, Google Search €362.39.

LinkedIn is the cheapest per company and the most expensive per person by a wide margin: 2.9x Google Search and 4.5x Meta.

Which one it more important? Depends on how many people have to be reached inside an account.

Gartner surveyed 632 B2B buyers in August–September 2024 and found buying groups “ranging from 5 to 16 people across as many as four functions”.

• Need to reach 5 people = LinkedIn is fine.

• Need to reach 16 people = it might be too expensive.

A channel that reaches one person per account cheaply, and the other four to fifteen expensively, is a different proposition than “cheapest per company” suggests.

Claim 3: LinkedIn takes the largest share of B2B ad budget

Share of total ad budget among Dreamdata customers, 2025: LinkedIn 41%, Google Search non-brand 33%, Facebook 8%, Google Search brand 8%, Google Smart 3%, Bing non-brand 3%, and 1% each for Google Display, YouTube, Google Other and Capterra.

This one is certainly accurate for the sample: the population is companies that bought B2B attribution software. Grouped together, Google’s properties (search brand and non-brand, Smart, Display, YouTube, other) take 47% against LinkedIn’s 41%, by the way.

Now let’s compare Dreamdata vs. a second vendor, measuring the same channel

HockeyStack Labs’ LinkedIn Ads benchmark report covers 70+ B2B SaaS companies between $5M and $1B ARR, with $28M of ad spend over three years, and a position-based attribution model (40% first touch, 40% last touch, 20% distributed).

Where the two disagree:

• CPC.
  Dreamdata: €5.98.
  HockeyStack: $10.48 in Q1 rising to $15.72 in Q3.
  Currency accounts for part of the gap and not most of it.

• CTR.
  Dreamdata: 0.57%.
  HockeyStack: 0.82% to 0.96% by quarter.

• Return. Dreamdata: 121% ROAS.
  HockeyStack: pipeline ROI of 2.44x to 6.01x, revenue ROI of 2.46x in Q4.

Same platform, overlapping years, two datasets built from real spend, and a headline ROI that differs by roughly double.

And no — neither vendor is lying. They have different customers, different attribution windows, different models, and different definitions of what counts as revenue. The important thing: a single vendor’s benchmark tells you what LinkedIn did for that vendor’s customers under that vendor’s model.

The 272-day journey

Dreamdata reports an average B2B customer journey of 272 days, split 81% marketing (first touch to sales pipeline) and 19% sales (pipeline to closed deal).

We don’t have an independent source in our database to check this against — so you can treat the 272 days and the 81/19 split as indicative figures.

Margin note worth noting alongside:
The Ehrenberg-Bass estimate, produced by John Dawes for the LinkedIn B2B Institute, claims that up to 95% of businesses are not in the market for most goods and services at any one time. That means that most people you target are not really into buying anything you’re trying to sell them. It is a derived figure though, calculated from ~5y supplier switching cycles rather than measured directly.

What else is worth highlighting?

Paid search is getting more expensive and less clicked in B2B

Google Search non-brand CPC rose from $4.13 in August 2024 to $5.34 in July 2025, up 29.36%, while CTR fell from 5.47% to 4.04%, down 26.16%. In the report, this is a within-account trend on the same measurement basis, which makes it more reliable than any cross-channel comparison in the report.

CPM benchmarks

CPM of €34.33 on LinkedIn against €9.18 on Meta confirms what every B2B advertiser already pays. Google Search at €234.64 is not comparable and should not sit in the same column — search impressions are scarce and intent-loaded by design.

LinkedIn might be even less efficient than Meta

LinkedIn’s CTR of 0.57% sits below Meta’s 0.61% while costing 3.7x more per thousand impressions. The report publishes this without spinning it.

The practical part:
If you are buying LinkedIn ads on Monday

1. We’d use the report for CPM and CPC expectations, and for the direction of Google Search costs. Both are measured consistently within one system.

2. Do not use the 121% ROAS to justify a budget.
   Run a holdout instead: pick a matched set of target accounts, stop LinkedIn spend against them for a full quarter, and compare pipeline creation. A 272-day journey means you will wait, and the wait is the price of an answer that attribution cannot give you.

3. Track cost per contact influenced alongside cost per company influenced. With buying groups of five to sixteen people, the €70 figure and the €1,033 figure describe two very different campaigns.

See you in the next episode!

Stacy’s Brain reads marketing research so you don’t have to, grades every source for methodology, sample and vendor interest, and tells you how far to trust each number.

Subscribe now

Sources

• 2026 LinkedIn Ads B2B Benchmarks Report — Dreamdata — vendor report, high vendor interest, methodology and sample disclosed

• LinkedIn Ads Benchmark Report — HockeyStack Labs — vendor report, high vendor interest, sample and attribution model disclosed

• Blake, Nosko & Tadelis (2015), Econometrica — randomized field experiment, no vendor interest

• Gartner B2B buyer survey, May 2025 — 632 buyers, fieldwork Aug–Sep 2024

• Ehrenberg-Bass / John Dawes for the LinkedIn B2B Institute — derived estimate, commissioned by LinkedIn