The short answerMarketing's share of ARR is flat in 2026: private SaaS companies spend a median 8%, unchanged from last year. Buyers moved instead: Forrester puts B2B buyers using AI in their buying process at 94%. In one week of LoudFace tracking in October 2026, ChatGPT, Perplexity and Google AI Overviews rarely named the same tracked agency first on generic buyer questions.
That last point is the one most 2027 plans will get wrong.
The 2026 numbers at a glance
| Finding | Number | Source, sample and window |
|---|---|---|
| Marketing budget as a share of company revenue | 7.8% in 2026, from 7.7% in 2025 | Gartner CMO Spend Survey, 401 CMOs and other marketing leaders, January to March 2026, mostly companies over $1B revenue |
| Median marketing spend, private B2B SaaS | 8% of ARR, unchanged from the previous year | SaaS Capital, more than 1,000 SaaS companies, survey completed March 2026 |
| Share of marketing budget going to AI initiatives | 15.3% | Gartner CMO Spend Survey 2026 |
| CMOs who say they lack the budget for their 2026 strategy | 56% | Gartner 2026 CMO Spend Survey press release |
| B2B buyers using AI in their buying process | 94% (Forrester); 45% used GenAI in a recent purchase (Gartner) | Forrester, January 2026 post; Gartner, 645 buyers, August to September 2025 |
| Buyers who sourced software recommendations from AI chatbots | Eight in ten, over the last 24 months | G2 2026 Buyer Behavior Report, 1,038 B2B decision-makers, June 2026 |
| How often buyers choose a vendor from their Day One shortlist | 95% of the time | 6sense 2025 Buyer Experience Report, nearly 4,000 buyers |
| Clicks on a search result when an AI summary appears | 8% of visits, against 15% without one | Pew Research Center, 900 U.S. adults, March 2025 |
| Of generic buyer questions where all three engines named a tracked agency, the share where they most often named the same one first that week | 13.8% (17 of 123); 13.0% to 16.3% depending on how ties are counted | LoudFace, Peec AI, 140 questions that name no agency, 16 to 22 October 2026 |
| Same first-named agency as the day before, when both days named one | Perplexity 68.9%, ChatGPT 40.4%, Google AI Overviews 31.4% | LoudFace, same 140 generic questions |
| Share of non-brand Google impressions from prompt-length queries | Median 7.1% across seven sites (range 0.8% to 28.0%) | LoudFace, Search Console, 1 to 21 October 2026 |
| Click rate of prompt-length queries against other queries | About 5 times lower (0.085% against 0.448%) | LoudFace, same seven sites |
How we gathered the data
We used two first-party datasets, plus public surveys and documentation read on each publisher's own page.
AI answers. We track 150 buyer questions in Peec AI, all about choosing marketing, SEO and AI search help, mostly for B2B SaaS companies. Peec asks each one once a day on ChatGPT, Perplexity and Google AI Overviews. For the week of 16 to 22 October 2026 we pulled every answer, 3,150 in all (1,050 per engine), with the sources each answer drew on (Peec lists pages an engine retrieved as well as the ones it cited inline), the agencies it named and, for ChatGPT, the web searches it ran. Agency measures cover the 53 agencies tracked in that project, so read them as "among the agencies we track". Ten of the 150 questions name a specific agency in their own text (four name LoudFace, six name another agency), so we leave them out of the agreement and stability figures, which cover the other 140 generic questions.
Google Search Console. We read the query report for loudface.co and for every other site we run search for, 11 sites in all, 1 to 21 October 2026, with brand queries removed. We call a query prompt-length when it runs 10 or more words or carries an instruction or persona block. Seven of those sites had at least 1,000 non-brand impressions in the window. Client sites appear only as percentages, a median and a direction. No site is named and no raw count is published. Search Console hides rare queries, so these shares cover only the queries it reports.
What we did not do. We did not compare impressions year over year. Google's own data-anomalies log says a logging error over-reported impressions from 13 May 2025 to 27 April 2026, and every year-ago window Search Console still holds sits inside that period. Our trend lines use months after the fix.
Third-party numbers. Every survey figure below was read on the publisher's own page on 23 October 2026 and carries its sample and fieldwork dates where the publisher states them. Where two surveys disagree, we show both.
Budgets: flat, and 8% of ARR is the SaaS median
Gartner's 2026 CMO Spend Survey has marketing budgets "rising only slightly to 7.8% of company revenue in 2026 from 7.7% in 2025". Its article on the survey adds that 7.8% is "18% lower than the mean budget allocation just four years ago". The survey's press release reports that 56% of CMOs say their marketing organization lacks the budget to deliver their 2026 strategy.
Read the sample before you copy that number into a board deck. Gartner surveyed 401 CMOs and other marketing leaders, and the vast majority of respondents work at companies with more than $1 billion in revenue. If you run a venture-backed SaaS company, 7.8% is a large enterprise's ratio, not yours.
The closer benchmark is SaaS Capital's 15th annual survey of more than 1,000 private SaaS companies, completed in March 2026: "The median percent of annual recurring revenue spent on marketing is 8%, unchanged from the previous year." For a company with $3 million to $5 million in ARR, the median is also 8% on marketing. SaaS Capital also found equity-backed companies spend 100% more on marketing than bootstrapped ones. If you have raised venture money, treat 8% as a floor.
Forrester's 2024 benchmark, from nearly 500 organizations, sits near the same level: "the average B2B firm invests 8% of annual revenue in marketing". Gartner and SaaS Capital both show the share flat against last year. Forrester gives a single 2024 reading with no trend, and it lands at the same 8%.
Where the money goes: AI and paid media, out of agencies
Gartner puts AI at "an average of 15.3% of marketing budgets", yet "only 30% report mature or fully developed AI readiness capabilities". The companies Gartner calls AI-ready put 21.3% of their marketing budgets into AI and spend 8.9% of revenue on marketing.
Paid media is the other winner. Gartner's article says CMOs are shifting spend toward paid media, now 31.4% of budgets, "funded by cuts to agencies". A year earlier, 39% of CMOs said they planned to cut agency budgets. Awareness and conversion now take 62.6% of media spend, and labor's share of the marketing budget rose from 21.9% to 24.5%.
Gartner's public releases do not break out search or SEO as a line, so nobody can tell you from them what share goes to organic. HubSpot's 2026 State of Marketing, from more than 1,500 B2B and B2C marketers, gets closest. "Updating SEO for search changes" is the fourth-ranked trend at 40.6%. On where marketers plan to increase investment most, AI chatbots like ChatGPT, Perplexity, Gemini and Claude lead at 37.7%. Website, blog and SEO comes fifth at 35.4%, behind paid social at 37.4%, video at 37.1% and content marketing at 36.9%. The Content Marketing Institute, reporting on 1,015 B2B marketers surveyed June to August 2025, found 95% of B2B marketers say their organizations use AI-powered applications.
So in Gartner's sample, mostly companies with more than $1 billion in revenue, the money is moving to AI, paid media and labor, and out of agencies. The public surveys cannot tell you how much of it reaches the AI answers your buyers now read.
How B2B buyers find software now
Forrester's buyer research says the share of business buyers using AI in their buying process "grew five percentage points, to 94%", and twice as many buyers named generative AI or conversational search as a more meaningful source than any other. 6sense, with nearly 4,000 buyers in its main survey, also reports 94%, for buyers using LLMs. G2's 2026 report, from 1,038 B2B decision-makers surveyed in June 2026, says "Eight in ten buyers used AI chatbots to source software recommendations in the last 24 months".
Gartner gets a much lower number. In its survey of 645 B2B buyers, run in August and September 2025, 45% said they used GenAI in a recent purchase, mostly to gather information on vendors and products. Asked whether they use AI in their buying process, Forrester's buyers say 94%. Asked by Gartner about GenAI in one recent purchase, buyers say 45%. The surveys also asked different people, with different samples and fieldwork dates, so read each number with its own question and population. Neither is a reason to wait.
The shortlist is where deals are decided, and AI now helps shape it:
- The Day One shortlist wins. 6sense: "Buyers choose from one of the four vendors on their Day One Shortlist 95% of the time (up from 85% last year)."
- AI and review sites shape shortlists. In G2's survey the top shortlist influences are "review sites (38%) and AI chatbots (37%)". Buyers who sourced recommendations from AI chatbots bought from their initial shortlist more often: 80% did so in at least three of their last five purchases, against 65% of buyers who did not.
- Buyers still check with vendors. In Gartner's buyer survey, "Sixty-nine percent of B2B buyers prefer to validate" AI-generated insights with sales reps. Gartner also reports that 70% prefer a completely digital, self-service experience.
- Contact comes earlier, on the buyer's terms. 6sense found first contact moved from 69% of the journey to 61%, and buyers initiated 79% of engagements. 6sense ties the earlier contact to buyers checking how vendors build AI into their products: 58% said that need made them engage earlier.
By the time a buyer fills in your demo form, an AI answer has often already shaped the list you are or are not on.
What the AI engines do with a buyer's question (our data)
No survey can tell you this part, because it measures the engines rather than the buyers. We asked the same 150 buyer questions every day for a week.
The engines disagree about which agency to name first. Leaving out the 10 questions that name an agency, all three engines named at least one tracked agency on 123 questions. On only 17 of them (13.8%) did ChatGPT, Perplexity and Google AI Overviews name the same agency first most often during the week. Counting tied weekly leaders differently moves that between 16 and 20 questions (13.0% to 16.3%). On 50 (40.7%), the three engines each had a different first name.
The first name changes day to day. On the same generic question and engine, the first-named agency matched the previous day in 68.9% of day pairs on Perplexity, 40.4% on ChatGPT and 31.4% on Google AI Overviews, counting only day pairs where both days named a tracked agency. A screenshot of one ChatGPT answer has worse odds of holding tomorrow than a coin toss.
They read different parts of the web. The three engines drew on 2,853 distinct domains that week. Only 220 of them (7.7%) appeared for all three.
| What we measured (16 to 22 October 2026) | ChatGPT | Perplexity | Google AI Overviews |
|---|---|---|---|
| Median sources per answer | 18 | 10 | 8 |
| Answers drawing on a community or social site (Reddit, Quora, LinkedIn, YouTube, Medium, X) | 6.3% | 29.2% | 62.4% |
| Answers drawing on Reddit | 3.1% | 1.1% | 17.7% |
| Answers drawing on a review site (G2, Clutch, Capterra and others) | 6.2% | 3.0% | 2.4% |
| Share of source slots held by the ten most frequent domains | 20.6% | 30.5% | 34.3% |
| Answers naming at least one tracked agency | 91.7% | 77.5% | 82.8% |
Google AI Overviews leans hardest on community content, drawing on a community or social site in 62.4% of its answers. youtube.com and linkedin.com were the two domains it drew on most. ChatGPT reads widest, with a median of 18 sources an answer.
ChatGPT searches like a person typing a prompt. It ran at least one web search in 70.3% of its answers, a median of 2 searches each, and those searches ran a median of 11 words. In 40.6% of its answers, at least one search included "2026". OpenAI says ChatGPT "typically rewrites your query into one or more targeted queries" when it sends them to search providers. Undated pages are weaker matches for the "2026" in those rewritten searches.
There is no single AI ranking to win. Each engine pulls from its own corpus, so AI search is a distribution plan, engine by engine.
What Search Console now counts as an impression (our data)
Google says sites appearing in AI Overviews and AI Mode "are included in the overall search traffic in" Search Console, inside the Web search type. Since 31 August 2026 Google has rolled out a separate report for impressions in those AI features to all websites, though a site with too few of those impressions usually will not see it. Google's documentation lists impressions by page, country, device and date, and no clicks or queries.
Here is what that looks like in a query report.
- On loudface.co, prompt-length queries took 9.7% of non-brand impressions in May 2026, 14.0% in July and 28.0% in October (1 to 21 of each month, all after Google's logging fix). In October those 1,996 queries drew zero clicks. The average non-brand query, weighted by impressions, ran 10.18 words.
- Across seven sites with at least 1,000 non-brand impressions in October, prompt-length queries took a median 7.1% of non-brand impressions, from 0.8% to 28.0%. Of the four sites with enough data in both May and October, the share rose on three and fell on one.
- On all seven sites, prompt-length queries clicked less often than the site's other non-brand queries. Pooled, they clicked at 0.085% against 0.448%, about 5 times lower.
Some of these queries were plainly written for a machine. loudface.co's October data holds 43 queries that open with a persona block, one beginning "about me: marketing manager/director series a to series c, $5m–$50m revenue" and ending "question: best cro agencies for b2b saas and tech companies". Those read like an AI visibility tracker's prompts, and few people type a persona block into Google.
We cannot tell you what share is tools. A long query with impressions and no clicks fits three explanations, and no Google page lets you separate them:
- A person typing a full question into AI Mode or under an AI Overview. Pew found 53% of searches with 10 or more words produced an AI summary, against 8% of one- or two-word searches.
- A tool sending automated queries. Google's spam policies treat "sending automated queries to Google" as a violation, but Google says nothing about how such queries appear in Search Console.
- A reporting artefact. Google has already corrected one impressions error this cycle.
The direction of the click side is not in doubt. Pew tracked 900 U.S. adults in March 2025: "Users who encountered an AI summary clicked on a traditional search result link in 8% of all visits", against 15% without one, and clicked a link inside the summary on just 1% of visits. Ahrefs, across 300,000 keywords, found an AI Overview now correlates with a 58% lower average click-through rate for the top-ranking page. Seer Interactive's data from 42 client organizations shows organic CTR down 65.2% year over year where an AI Overview appears and the brand is not cited, down 49.4% where it is cited, and down 46.2% even where no AI Overview appears. Seer's year-over-year figures set Q3 2025, inside the months when Search Console over-reported impressions, against Q3 2024, before them. Over-reported impressions push CTR down, so our reading is that part of Seer's decline may come from that error. Ahrefs sets AI Overview keywords against other keywords from the same months, so the error weighs on both of its groups.
Impressions stopped meaning demand. If your 2027 plan sets an impressions target, part of that number may be queries no person typed, and Search Console gives you no way to tell which.
What to do with your 2027 plan
- Set the budget from SaaS numbers. Start from SaaS Capital's 8% of ARR median, not Gartner's 7.8% of a billion-dollar company's revenue, and remember that equity-backed companies spend twice as much on marketing as bootstrapped ones. Then decide what share goes to the channels where your buyers build the shortlist.
- Fund AI search as distribution, per engine. Google AI Overviews drew on a community or social site in 62.4% of answers. ChatGPT's searches run a median 11 words, and 40.6% of its answers included a search with "2026" in it. One owned page answers neither on its own: you need dated, specific pages plus a presence on the sites each engine reads.
- Measure AI visibility as a rate over weeks, per engine. With the first name changing day to day on ChatGPT and AI Overviews, a single check proves nothing. Track the share of answers that name you, per engine, over at least a few weeks. Never blend the engines into one number.
- Split your Search Console reporting. Put prompt-length queries in their own bucket, then judge organic on clicks, sign-ups and pipeline. A rising impressions line can be a rising share of machine-written queries.
- Get on the Day One shortlist. 6sense found buyers pick from their Day One shortlist 95% of the time, and in G2's survey review sites and AI chatbots are the two biggest shortlist influences.
For more on the mechanics, see what AI engines actually cite, where ChatGPT gets its information, what Google AI Mode is and what an AEO program costs.
We run SaaS search programs this way at LoudFace: per-engine AI visibility across ChatGPT, Perplexity, Gemini and Google AI Overviews, judged against the revenue it brings in. If you want to see where you stand, run a free AI visibility audit. It checks where your brand stands across ChatGPT, Claude, Gemini and Perplexity searches, versus your top competitors.
Sources
- Gartner, 2026 CMO Spend Survey press release (11 May 2026) and CMO Spend Survey article
- Gartner, 2025 CMO Spend Survey press release (12 May 2025)
- Gartner, media spend release (8 June 2026) and B2B buyer survey release (20 May 2026)
- SaaS Capital, Spending Benchmarks for Private B2B SaaS Companies (June 2026)
- Forrester, B2B Buyers Make Zero-Click Buying Number One (22 January 2026) and The Average B2B Firm Invests 8% of Revenue in Marketing (17 December 2024)
- G2, 2026 Buyer Behavior Report
- 6sense, 2025 B2B Buyer Experience Report
- Pew Research Center, Google users are less likely to click on links when an AI summary appears (22 July 2025)
- Ahrefs, AI Overviews reduce clicks, update; Seer Interactive, AIO impact on Google CTR
- HubSpot, 2026 State of Marketing data; Content Marketing Institute, B2B content marketing research
- Google, AI features and your website, Search Console performance report help, generative AI performance reports, data anomalies and spam policies; OpenAI, ChatGPT search
