More than a third of in-house marketers are still putting money into channels they privately believe are not working, because the numbers look too good to cut. That is one of the findings in Playing defense, a report published on October 29, 2026 by Funnel, the Stockholm-based marketing data company, and Ravn Research. Based on a survey of 233 marketers and interviews with five agency leaders, it argues that pressure to prove value is steering budgets toward what can be defended rather than what is expected to perform.

In Short

A marketing software company asked 233 marketers how they make decisions, and many admitted they pick the safest, easiest-to-prove options instead of the ones they think would work best. That matters because it means some advertising money keeps going to channels that look good in reports but that the people running them doubt, while new ideas get turned down. The survey suggests the problem is less about bravery and more about how few teams can actually show what their marketing achieves.

What the survey covers, and what it leaves out

According to the report, Funnel surveyed 233 "in-house marketing professionals and agency leaders globally" and interviewed five agency leaders "from high-performing, data-savvy shops in the U.S. and Europe." Ravn Research, a research services company founded by Clare McDermott in 2021, managed the work. Funnel sponsored it. The report calls this its fifth annual study.

Where the report uses the word marketers, it means in-house marketers at brands only. Agency respondents are labelled separately, and Funnel says they will feature in a follow-up report. The press release describes the sample as "233 predominantly in-house marketers."

Several basic details are missing from both the report and the release. Neither gives fieldwork dates. Neither lists the countries respondents work in, beyond the word "globally." Neither says how many of the 233 work in-house and how many at agencies, how respondents were recruited, whether results were weighted, or what size of company they work for. The report closes with an invitation to join the Funnel Marketing Research Panel, but it does not say whether this year's respondents came from that panel or from Funnel's own customer base.

Those gaps matter for interpretation. For a random sample of 233, the conventional 95% margin of error would be roughly plus or minus 6.4 percentage points, by PPC Land's calculation. That figure only holds for random samples, and in-house-only findings rest on fewer than 233 answers. Comparisons between agency and in-house respondents, or between the report's "clear-signal" and "muddled-signal" groups, rest on subgroups whose sizes are not published.

There are also labelling inconsistencies. The report carries the footer "Funnel Marketing Intelligence Research 2027," although it was released in October 2026. It describes the agency respondents both as "agency professionals" and "agency leaders."

Pressure, and what it does to decisions

The report's starting point is a question about personal pressure. According to Funnel, 80% of in-house marketers feel at least moderate pressure to justify their own value to the business. A narrower group, 31%, report significant pressure and say their value is regularly questioned. Some 58% say that pressure has increased compared with a year ago.

The release headline, "80% of marketers under pressure to justify their value," therefore folds moderate and significant pressure into a single number.

Among those reporting at least moderate pressure, 48% say it makes them less willing to take risks in their marketing decisions. "Marketers are afraid of making mistakes. They're afraid of being blamed if a strategy doesn't work. And let's be honest ... they're afraid of losing their jobs," said Natalia Hernandez, managing director of data at Brainlabs, one of the five interviewees. One anonymous survey respondent put it more bluntly: "I don't want to rock the boat. Get paid. Go home."

The defensive choices

The survey asked whether teams had done a series of things in the past two years. The answers form the core of the report:

  • 60% chose a safer, more provable strategy over a riskier one they suspected would work better
  • 59% prioritized marketing activities because they were easier to measure, not because they were the best option
  • 58% rejected a marketing idea because they could not estimate its impact in advance
  • 36% continued investing in a channel they believed was not working because the numbers looked too good to cut

Separately, 61% said they had chosen a metric because it told a more favorable story. The report's own text describes this as having "cherry-picked a metric." And 29% said they had stayed silent in a meeting when they knew the data was wrong or misleading.

On experimentation, 84% say they never experiment or do so only occasionally, rarely committing real budget to unproven strategies. The report does not break down how many chose never and how many chose occasionally.

The 36% figure is the most concrete of these. It describes budget that the people managing it doubt, sustained by reported performance they do not believe. The report does not say which channels are involved, or whether the numbers that looked too good were platform-reported figures, though elsewhere it notes that 49% of in-house teams rely primarily on platform-reported data and simple attribution.

Costs up, returns flat

The report sets the defensive behavior against a cost squeeze. According to the survey, customer acquisition costs rose over the past 12 months for 56% of marketers, stayed about the same for 31% and fell for 13%. Budgets have not kept pace: only 30% say their paid media budget has grown significantly over the past two years, while the rest report modest increases, no change or cuts.

Return on ad spend has largely stalled. Over the past 12 months, 35% report that their ROAS increased, 41% say it stayed flat and 24% say it declined.

Paid media takes 41% of overall marketing budgets on average, according to the report. Of that paid media money, 47% goes to Google and Meta combined. The report does not split the figure between the two companies, nor does it say whether the averages are weighted by budget size.

"Whenever there's uncertainty or pressure for in-quarter results, marketers tend to flock to low-funnel tactics. It's the safe, familiar move, but it's not necessarily the one that drives long-term growth," said Spencer Lian-Thornton, vice president of growth and partnerships at Mindgruve.

Funnel's own platform data offers a different view of where the money goes. In July, the company reported that US search spend across its customer base rose 122% to $100 million a month between January 2024 and April 2026. Other datasets cut differently. Billy Grace, which measures a cohort of European advertisers through its own model, found in August that Google's spend-weighted share fell from 62.1% to 57.3% in the first half of 2026, while median cost per click dropped 10.8%. The methods are not comparable with a self-reported survey, but together they show how much depends on who is counting and how.

Zero-click search and the measurement gap

Organic search is where the report finds the most visible strain. According to Funnel, 86% of marketers say at least one aspect of their marketing is underperforming, and organic search is the most commonly cited, at 42%. Some 39% say zero-click search, which the report describes as "a byproduct of AI-powered search results," is hurting their organic reach.

Respondents estimate that 22% of their organic traffic has been lost to zero-click search. The report is candid about the quality of that number: 68% have not attempted to quantify the impact at all. "Put plainly: The majority of marketers don't actually know how much traffic they're losing; they only know that it's disappearing," the report says.

The report's own trend data complicates that line. Asked how organic search traffic has moved over the past 12 months, 38% said it decreased, 27% said it stayed about the same and 35% said it increased. The share reporting growth is almost as large as the share reporting decline. Neither the release nor the report's written analysis discusses that split.

Alex Keller, senior vice president of performance marketing at Optimal, framed the shift as a visibility problem rather than a collapse in demand. "Zero-click is really just a loss of visibility, similar to what we saw with cookie deprecation," he said. "People are not stopping searching; they're expanding. They're going to Reddit, YouTube, TikTok, trying to find real opinions from people who've tried the brand. It's not Doomsday, but it is definitely the next replay of cookie deprecation."

John Gough, senior vice president of strategy and services at Element Three, went further on what it could mean for websites. "Google and the other big players aren't going to roll out changes that undermine their own business model. Still, the more AI lets people bypass search, the more it raises real questions about whether you even need a website. I don't think people are prepared for a world where search turns into mostly view-through conversions," he said.

Click-level data supports the concern about visibility, if not the specific 22% estimate. Ahrefs found in February that AI Overviews cut position-one click-through rates by 58% for affected keywords, measured across 300,000 keywords against a December 2023 baseline. Surveys of marketers have produced less uniform answers. An NP Digital survey of 600 US marketers in July 2025 found that 55.5% reported increased traffic since AI Overviews launched, with 8.3% reporting declines. Funnel's 38% against 35% split on organic traffic sits between those positions.

AI and job security

The report asked how concerned respondents are that their current role could be eliminated or significantly reduced by AI in the next three years. Among in-house marketers, 35% were very or somewhat concerned. Among agency marketers, the figure was 42%.

Seniority produced a wider gap than employer type. According to the report, 47% of non-leaders voiced that concern, against 25% of those in leadership roles. The release reports the agency and in-house split but not the seniority split. Because the number of agency respondents is not disclosed, it is not possible to judge from the published material whether the seven-point agency gap is larger than sampling noise.

The report also asked about AI-generated recommendations. Some 67% say they understand the reasoning behind an AI recommendation "always" or "often." The remaining 33%, a figure the release rounds to "exactly a third," act on AI insights they do not reliably understand. "Half of us are trying to figure it out. Others are like, 'I'm too old for this,'" Hernandez said.

Dr. Isabelle Kes, senior consultant for data strategy and marketing measurement at FELD M, warned about what happens when stakeholders cannot interrogate those outputs. "If the knowledge of the stakeholders using AI-generated insights is not high enough to really challenge what the insight is, and why it could be this way or another way, this is really dangerous. It might simply lead to wrong assumptions," she said.

The survey does not ask what marketers expect AI to do to overall headcount, only to their own role. PPC Land examined the counterargument in December 2025, when Box chief executive Aaron Levie applied the Jevons paradox to marketing work, arguing that cheaper output could raise demand for marketing labor rather than cut it.

Measurement: widely valued, rarely mastered

Asked what separates the most effective marketing organizations from the rest, 48% of all respondents, agency and in-house combined, chose superior data and measurement capabilities. That ranked ahead of willingness to invest in brand and creative (44%); bold, decisive leadership (43%); sophisticated use of AI (38%); a culture of honest reporting and accountability (36%); strong agency-client relationships (31%); and the ability to prove marketing's impact to the CFO (27%). The percentages add up to well over 100, so respondents evidently could pick more than one answer. Unsure responses were excluded.

Practice lags well behind. Asked how mature their company's measurement program is, 49% of in-house marketers said they rely primarily on platform-reported data and simple attribution. Another 42% said they are expanding beyond the platforms with multi-touch attribution or early independent measurement. Only 9% described triangulated measurement, which combines incrementality testing, marketing mix modeling and attribution. Agency respondents described more developed setups, at 32%, 46% and 22% respectively.

Trust follows the same pattern. Only 10% of in-house marketers say they have high trust in their attribution model; 60% report moderate trust and 30% low trust. And just 18% say they can clearly determine the impact of each channel.

The report also describes the 9% as the share who call their measurement program "advanced," a label that does not appear in the answer options it reproduces.

Lian-Thornton described the professional cost of moving away from familiar models. "Brand marketers or even performance marketers are nervous because they've spent years optimizing their teams and systems ... around that framework," he said, referring to last-click, ROAS-based thinking.

Hernandez pointed to a different problem: what happens when methods disagree. "Clients often want to measure everything, but they're sometimes scared of what the data may show, because sometimes it's not going to tell you what you want to see or hear. That adds complexity. For example, what if I'm running an experiment and MMM is giving me a completely different perspective? Should I trust the one test I've done, or the historical trend I've been using for a year? It's never an easy answer, especially when results are inconclusive," she said.

Other recent studies have reached similar numbers by different routes. Winterberry Group, in a report sponsored by measurement vendor Lifesight, found in September that only 9% of 121 US marketing executives say they can practice unified measurement extremely well, while 93% use platform-native reporting. In a study of 197 senior marketers, CIMM and the 4As found that 43% see cross-platform measurement gaps as a major or severe barrier. And the conflict Hernandez describes, between an experiment and a model, is the subject of the IAB's draft Cross-Channel Measurement Decision Guide, released for public comment on October 20 with a deadline of November 4.

Funnel's previous edition, published in December 2025 and based on 238 marketers, reported that 86% of in-house marketers could not determine each channel's impact on overall performance. This year's report says 18% can clearly do so. Whether the two editions used the same question and answer options is not stated, so the figures cannot be compared directly.

Self-assessed grades

The grades are middling. In-house marketers gave their own teams a B- on average, with 9% awarding an A. Agency marketers graded themselves a B, and 20% gave their own agency an A. In-house marketers were harsher on their agencies, at a C+ average: only 4% awarded an A, and 12% were not sure how to grade them.

Funnel's release presents the B- differently, stating that marketers "grade their own team's measurement performance on average as a 'B-'." The question reproduced in the report asks something broader: "If you had to give your internal marketing team (or agency) a grade for how well they hit goals and drive results, what would it be?" The grade measures perceived results, not measurement specifically.

Clear signal, muddled signal

The report's main analytical device splits respondents by their answer to one question: "How well can your team determine the impact of each marketing channel on overall performance?" Those who chose "We have a strong understanding of each channel's contribution" or "We have a general sense, but it's not always precise" are classed as clear-signal teams. Those who chose "We struggle to assess each channel's contribution" are muddled-signal teams. Respondents who answered "Unsure" were left out.

According to the report, the two groups differ sharply:

  • 18% of clear-signal teams report low trust in their attribution model, against 63% of muddled-signal teams
  • 14% of clear-signal teams report high trust, against none of the muddled-signal teams
  • 15% of clear-signal teams describe their strategy as primarily defensive, against 52% of muddled-signal teams
  • 85% of clear-signal teams describe their CMO's vision as at least somewhat bold, against 48%
  • 76% of clear-signal teams grade their own team an A or B, against 41%

The report does not say how many respondents fall into each group. It is careful about causation: "The findings do not suggest that advanced measurement automatically creates bold marketers," it states, adding that shared confidence in evidence and less defensive decisions "can reinforce one another."

There is a further limit. All of these measures are self-reported by the same respondents in the same survey. A team that says it understands channel contribution is likely, almost by definition, to say it trusts its attribution model. The comparison shows that the answers move together; it cannot show which comes first.

What marketers say they would do with job security

One question asked what respondents would do differently if their job were guaranteed for the next five years. The answers, according to the report:

  • Invest more in brand-building versus short-term performance: 56%
  • Advocate for different marketing technology or tools: 48%
  • Experiment with channels or tactics they currently avoid: 46%
  • Push back on unrealistic ROI expectations from leadership: 45%
  • Take a longer-term view on campaign performance: 44%
  • Challenge the current attribution model or measurement approach: 37%
  • Question the credibility of certain data sources: 29%
  • Nothing, already operating with the freedom needed: 15%

The report reads the top answer as evidence that "the amount of time marketers believe they need to achieve growth is in direct conflict with the time periods over which their performance is judged." Lian-Thornton put it more simply: "You can't just retarget your way to growth."

The finding fits a long-running argument about short-term measurement. A TransUnion and MMA Global study, reported in October 2025, found that Kroger's long-term sales lift was six times its short-term lift, and that traditional measurement tools captured between 17% and 56% of the full impact across three brands. That study relied on three case studies rather than a survey.

The report's only worked example comes from Tallink Silja Line, a European cruise and shipping company. According to the report, the company moved away from last-click reporting, which it says overvalued lower-funnel channels, and adopted a framework based on marketing mix modeling. The report credits the change with a 50% improvement in return on ad spend and a 26% increase in revenue from paid digital media, while digital media spending fell 16%.

"Google Analytics would heavily credit paid search with conversions while suggesting that Meta was not converting at all," said Tiina Lips, digital analytics lead at Tallink Silja Line. "We knew intuitively this was not true, but we were stuck with an attribution model that did not reflect the actual customer journey."

The report gives no time period for those results, no baseline figures and no description of how the improvement was measured. It also does not state whether Tallink Silja Line is a Funnel customer.

Last-click attribution, the model Tallink Silja Line says it abandoned, has become a commercial theme in its own right. In November 2025, Silverback Strategies chief executive Neil Welsh issued a public apology for inflated ROAS metrics as part of a campaign against last-click thinking. In May 2026, TikTok launched an Attribution Portfolio aimed at the last-click debate. Each of those efforts, like Funnel's report, comes from a company with a commercial stake in how credit gets reassigned.

The report's prescriptions

The final section sets out five steps, each framed by the report as a route out of defensive measurement: build trusted, queryable data; match the measurement method to the decision; make uncertainty and assumptions explicit; define in advance what a test result will commit the team to; and give teams permission to act on and challenge results, including AI-generated ones.

Gough, quoted in that section, argued that many teams have built dashboards rather than tables. "If your data isn't queryable, that's where you need to begin. By queryable, I mean SQL, agents, an MCP connection to your AI, and so on. It takes a foundational layer that, frankly, many marketing teams haven't built yet," he said.

On testing, the report uses connected TV as its example, describing an approach in which teams agree before a test what result would justify more spending and what result would mean cutting back. On zero-click search, it proposes looking at branded search, direct visits, conversions, customer surveys and business outcomes alongside traffic, while conceding that "no single measure will provide a complete answer."

Kes cautioned against sophistication for its own sake. "A super complex and highly sophisticated marketing measurement system in a not-very-mature company would lead to complete confusion and overwhelm. If you increase the complexity and sophistication of the marketing measurement, you at the same time need to improve the governance, improve the literacy, and increase the maturity overall, not only the technical or model-wise maturity," she said.

The companies behind the report

Funnel was founded in Stockholm in 2014 by Fredrik Skantze, its chief executive, and Per Made, its chief marketing officer. According to the company, it handles more than 10% of global advertising spend, works with more than 5,000 brands including Uber, Adidas, Sony, Havas and Publicis Groupe, and collects data from more than 600 sources. It sells software that centralizes marketing data for reporting, AI workflows and marketing mix modeling. The 10% figure is the company's own claim; neither the report nor the release explains how it is calculated.

That commercial context is relevant. The report's central argument, that better measurement is associated with more confident and less defensive marketing, aligns with the products Funnel sells. The research was run by a third party, but it was sponsored by the vendor, and its methodology section is brief.

"This research shows what happens when marketers can't fully see what's working and what's not, retreating to the metrics that are easiest to defend as the squeeze from zero-click search and AI hits closer to home. Clean, trusted data and accurate measurement methods give teams the confidence to challenge their own spending and act on it," said Carl Ronander, vice president of marketing at Funnel.

Why this matters

For advertisers, the report's most useful numbers are not the headline percentages on pressure but the ones on how budgets are actually judged. Nearly half of in-house teams, 49%, still rely primarily on platform-reported data and simple attribution. Almost half of paid media money, 47%, goes to Google and Meta, the same companies that supply much of the reporting used to evaluate it. And 36% say they have kept funding a channel they believed was failing because its numbers looked too good to cut. Why would a team keep paying for something it does not believe in? The report's answer is that, in the current climate, "it's easier to punish a failed experiment than change a failing status quo."

The pattern is not new. A TransUnion and EMARKETER survey of 196 US marketers, published in October 2025, found that 60.2% said internal stakeholders question their metrics at least sometimes. Incubeta's May 2026 study of retail and ecommerce leaders found that 70.4% were confident their budgets were deployed effectively, while 41.6% admitted some investment was not delivering full value. Funnel's figures add a behavioral layer: they describe what marketers say they do when the evidence is weak, not just how confident they feel.

The limits are equally plain. Every finding is self-reported, and the sample is small and only partly described. None of that makes the findings wrong. It does mean the 36% and 61% figures are best read as a description of how a few hundred marketers characterize their own decisions, rather than a measured rate across the industry.

Timeline

Summary

Who: Funnel, a Stockholm-based marketing data company, and Ravn Research, which managed the survey on Funnel's behalf. Respondents were 233 in-house marketers and agency professionals; five agency and consultancy leaders from Brainlabs, Optimal, Mindgruve, Element Three and FELD M were interviewed.

What: A report titled Playing defense, which finds that 80% of in-house marketers feel at least moderate pressure to justify their value, 36% have kept investing in channels they believed were not working, 61% have chosen metrics because they told a more favorable story, and only 9% use triangulated measurement.

When: The report and its press release were published on October 29, 2026. Neither discloses when the survey was in the field.

Where: The survey is described as global, without a country breakdown; the five interviews were with leaders in the United States and Europe.

Why: The report argues that pressure to prove value, rising acquisition costs, zero-click search and AI are pushing marketers toward defensible rather than effective choices, and that teams with clearer measurement report less defensive behavior. The findings are self-reported, the sample is small, and the sponsor sells the measurement infrastructure the report recommends.