On November 18, 2026, DoubleVerify published "Proof Points: Decoding the Path to Purchase," the closing report of its 2026 Global Insights series, and its consumer figures put AI chat well behind older habits. According to the report, 17% of consumers use AI chat to research products before they buy, against 39% who use search engines and 40% who read online reviews. The seven-page document combines DoubleVerify's annual surveys of 22,000 consumers and 2,020 marketers with ad spend data from DV Rockerbox, the company's attribution business.

DoubleVerify announced the report by email the same day under the subject line "Just released: Global Insights report on the consumer journey," describing it as "The final Global Insights report of 2026." It is the third report built on the same survey to appear in five weeks, after a walled garden edition in October and a study of AI in media buying at the end of that month.

In Short

DoubleVerify, a company that checks and measures online ads, asked 22,000 shoppers and about 2,000 marketers how people find and buy products, and how marketers track it. Search engines and online reviews still lead shoppers' research while AI chat sits far behind at 17%, and more than half of marketers judge results with the ad platforms' own reports. The report ends by promoting DoubleVerify's own measurement products, so its warnings about the limits of platform reporting come from a company that sells an alternative.

What the report contains

The report runs to seven pages, four of which carry data: when consumers shop, how consumers research and buy, when marketers advertise, and how marketers measure. Its opening sets the frame with a single number, $1.46 trillion in projected global ad spend for 2027, sourced to WARC. That matches the figure in WARC's October forecast, which put 2026 spending at $1.34 trillion, up 11.9%, and 2027 at $1.46 trillion, up 8.4%.

"With more than a trillion ad dollars at stake, it's critical for marketers to understand their customers' path to purchase," the report says. The email makes the same argument in budget terms: "Ad spend on all channels is projected to be $1.46T in 2027, and marketers must understand their customers' path to purchase in order to allocate budget effectively."

The report closes with product material on DV MAP and DV Rockerbox, followed by the company's "Verify. Optimize. Prove." tagline.

How the survey was run

The report's methodology note is brief. "Data for this report comes from DV proprietary metrics, DV's annual commissioned global surveys of 22,000 consumers and 2,020 marketers and, where noted, external sources," it says, and links to a separate methodology page. According to that page, Sapio Research ran the fieldwork for DoubleVerify in February 2026, using online partner panels with an email invitation and an online questionnaire.

Item Consumers Marketers
Sample 22,000 2,020 "marketing and advertising decision-makers"
Countries 21 25
Fieldwork February 2026 February 2026
Research firm Sapio Research Sapio Research
Margin of error (95% confidence) plus or minus 0.7 points plus or minus 2.2 points

The 21 consumer markets are Australia, India, Indonesia, Japan, the Philippines, Singapore, Thailand and Vietnam in Asia-Pacific; France, Germany, Italy, the Netherlands, Saudi Arabia, Spain, the UAE and the UK in Europe, the Middle East and Africa; Brazil, Colombia and Mexico in Latin America; and Canada and the US. The marketer survey adds New Zealand, Austria, Poland and Switzerland. The methodology page does not define what makes a respondent a decision-maker, and it does not say how results were weighted across countries.

This is the same survey PPC Land drew on when DoubleVerify released its walled garden report on October 14 and its AI media buying report. The overlap shows. The 25% of consumers who bought on social media in the past 12 months, and the 26% who research products there, are the same figures the walled garden report published five weeks ago.

When consumers shop

"Demand doesn't follow the traditional retail calendar," according to the report. Its headline figure is that 74% of consumers buy throughout the year, and it says that "only about a quarter" make most of their purchases around events, holidays or seasons. Older respondents are the steadiest buyers; younger ones are the most likely to time their spending.

The age chart offers three answers rather than two, and that matters for reading the 74%.

Age Buy throughout the year, no significant change for events Concentrate purchases at certain times but also shop out of season Make most purchases at certain times of year
18 to 24 22% 45% 33%
25 to 34 23% 41% 36%
35 to 44 30% 41% 29%
45 to 54 41% 38% 22%
55 to 64 52% 32% 16%
65 and over 64% 25% 11%

No age group reaches 74% on the strictest answer, "I purchase throughout the year with no significant change for events, holidays or seasons." The highest share is 64%, among those 65 and over. The headline figure therefore cannot refer to that answer alone, and the report does not explain how it was derived. Among 18 to 24 year olds, only 22% describe their buying as unaffected by the calendar, while 45% say they bunch purchases around particular moments and also shop out of season. The report does not publish the all-ages split of the three answers.

How consumers research and where they buy

The research chart groups nine sources into four families, which the report labels IRL, Online, Social and AI. Respondents could pick more than one.

Research source Family Share of consumers
Online reviews Online 40%
Search engines Online 39%
Physical store IRL 30%
Product comparison sites and apps Online 26%
Friends and family IRL 26%
Social media Social 26%
Video reviews or tutorials Online 24%
AI chat AI 17%
Influencers Social 11%

AI chat ranks eighth of nine, ahead only of influencers. The report's text nonetheless says "AI chat is increasingly part of these conversations, especially for older Gen Z and younger Millennial consumers." Neither half of that sentence is backed by a number in the report: there is no earlier figure to show an increase and no age breakdown for AI chat.

Personal recommendations do better. "Consumers are more than twice as likely to seek recommendations from people they know as they are to take advice from influencers," according to the report, and the chart bears that out: 26% against 11%, a ratio of about 2.4 by PPC Land's calculation. Research in a physical store, at 30%, beats both.

The purchase question asked where respondents had bought something in the past 12 months.

Purchase location, past 12 months Share of consumers
Physical store 52%
Online marketplace 47%
Grocery, big box or other retail site and app 36%
Brand or manufacturer site or app 29%
Social media 25%

"When it's time to convert, consumers flock to physical stores and online marketplaces such as eBay and Amazon," the report says. The physical store figure, 52%, is a narrow majority, and the report does not say how the question was framed or whether it covered a particular category of goods.

The research findings sit alongside other surveys PPC Land has covered. A study by Smarty Marketing of 1,295 US residents, published in May, found that 7.26% named ChatGPT as their starting point for product searches, against 56.68% for Google. The questions differ, and DoubleVerify asked about all research sources rather than the first one, but both point the same way: AI chat is in the mix and nowhere near the top.

When marketers advertise

The report's third data section switches from survey answers to DV Rockerbox customer data. "Advertisers increase their spend in bursts around shopping holidays and special events, which can differ by vertical," according to the report. It says advertising for home goods such as furniture follows the broader retail sales calendar, and that "Apparel ad spend rises around the winter holidays and plunges right after the new year, which is when health and wellness spend increases because consumers have resolutions on their mind."

The chart that accompanies those claims, titled "Average Daily Ad Spend by DV Rockerbox Customers, Q4 2025 to Q3 2026," shows a single line for all customers combined. It carries no values on its vertical axis and does not break spending out by vertical, so none of the category patterns described in the text can be checked against it.

What the line does show is shape. By PPC Land's reading of where it falls along the time axis, the highest point lands at the end of November 2025, the week of Thanksgiving and Black Friday, and stands well above anything else in the 12 months. A second, lower rise runs from just before Christmas into the first days of January. Spending runs at its lowest and flattest through the second quarter of 2026, then climbs through July and August before easing toward the end of September. The chart says nothing about how many advertisers are included, in which markets, or whether the figures are adjusted for customers joining or leaving DV Rockerbox during the period.

Placed next to the consumer survey, the chart shows advertiser spending concentrated around a handful of events, while DoubleVerify's headline consumer figure is that 74% of people buy throughout the year. The report does not connect the two findings. They also measure two different things in two different ways: what consumers say in a survey taken in February 2026, and what DV Rockerbox clients spent across four quarters.

How marketers measure

"Virtually all marketers measure and track path to purchase for their campaigns," according to the report, though it gives no figure for that statement. The chart answers a narrower question: "Which methods do you use when measuring and tracking the path to purchase for your campaigns?"

The report presents the methods as levels of detail. About half of marketers use marketing mix modeling to look at the many factors that affect a business, "a little over a third" use multi-touch attribution to follow the shopper across channels, and more than half rely on the reporting supplied by the platform that sold the ads. View-through attribution, which credits an ad that was seen but not clicked, sits at 40%, slightly ahead of multi-touch attribution.

Measurement method Share of marketers
Platform native reporting 55%
Marketing mix modeling 48%
View-through attribution 40%
Multi-touch attribution 38%
Last-click attribution 35%

The five shares add up to 216%, which means the average respondent named a little over two methods, by PPC Land's calculation.

On platform reporting the report is direct. "By nature, however, a platform's data is the most limited, since it covers only how that platform performed," it says.

Two features of the chart stand out. Incrementality testing, one of the three methods DoubleVerify sells through DV Rockerbox, is not among the five options shown. And the ranking matches a much smaller US survey that PPC Land covered in October: Winterberry Group, in research sponsored by Lifesight, found that platform-native reporting was the most used method among 121 senior US marketers at 93%, followed by marketing mix modeling at 71% and multi-touch attribution at 69%, with last-click at 32%. The levels differ widely between the two studies, which used different samples and questions, but in both platform-native reporting comes first and marketing mix modeling second.

How DoubleVerify describes its measurement products

The report's closing section sets out what DoubleVerify sells for this problem. It describes DV MAP as a product that "pulls together DV's vast verification, optimization and measurement capabilities," powered by DV Neura, and calls it "the digital advertising industry's only all-in-one planning and execution platform." The report offers no evidence for the claim to be the only such platform. DoubleVerify launched DV Neura on June 17, 2026 as an AI engine running across its products.

DV Rockerbox carries what the company calls its Prove offering, which "evaluates the real return on social, streaming TV and open web campaigns using deduplicated data sourced directly from our platform and publisher partners." According to the report, it rests on three methods:

  1. Multi-touch attribution. "Analyzes daily performance and engagement patterns across deterministic touchpoint events."
  2. Marketing mix modeling. "Delivers strategic planning and forecasting insights for informed campaign management and budget allocation."
  3. Incrementality testing. "Validates data-driven hypotheses with causal evidence and experimentation, and further informs MTA and MMM methodologies."

The section points to a product page at doubleverify.com/products/advertisers/prove. It does not give prices, minimum spend levels or the list of publisher partners whose data feeds the deduplication.

Why it matters for marketers

Measurement is where the report's argument lands. The survey shows that more marketers lean on the platforms' own numbers than on any cross-channel method, and the report's text points out the weakness in that. The difference between the two can be large. A guide from Funnel that PPC Land reported on November 9 cited a retailer holdout test in which a catalogue programme delivered a 14% incremental revenue lift against a 40% contribution claimed by vendor attribution, and showed how a channel with a return on ad spend of 3.0 can add only $0.40 for each extra dollar.

The timing pattern matters for budgets too. Adobe's forecast for the 2026 US holiday season, covered by PPC Land in September, put Cyber Week online spending at $47.5 billion and projected traffic from AI chat and browser services to US retail sites up 130%. DoubleVerify's spend chart shows its customers concentrating money in the same late-November window in 2025, while the headline figure from its survey is that 74% of consumers buy throughout the year.

AI chat is also where measurement firms are positioning themselves. OpenAI listed DV Rockerbox among the attribution partners for ChatGPT Ads on October 5, and two days later DoubleVerify said Weight Watchers' cost per acquisition on ChatGPT Ads ran 15.3% below benchmark in DV Rockerbox data. The new report's 17% figure for AI chat in product research gives a sense of how much of the consumer journey that channel touches at present, at least by consumers' own account.

The company behind the numbers

DoubleVerify is both the author of this research and a seller of the remedies it describes. Its critique of platform reporting, its emphasis on cross-channel methods and its closing section on DV MAP and DV Rockerbox all point toward the same commercial offer. That does not make the survey figures wrong, but it explains the framing. DoubleVerify completed its purchase of Rockerbox on March 13, 2025, and its measurement revenue reached $69.1 million in the third quarter of 2026, up 8%, out of total revenue of $205.6 million.

The company is also being sold. Nielsen agreed on August 6, 2026 to acquire DoubleVerify for $13.60 a share in cash, an enterprise value of about $2.15 billion, with closing expected in the first quarter of 2027. The report does not mention the deal.

The report leaves several things out. It carries no fieldwork dates or country list of its own, no all-ages split for the timing question, no age or trend data behind its AI chat claim, no values or vertical breakdown on the spend chart, no figure behind "virtually all marketers," and no incrementality option in the measurement question.

Key facts

  • Report title: Proof Points: Decoding the Path to Purchase
  • Series: DoubleVerify 2026 Global Insights, described by the company as "The final Global Insights report of 2026"
  • Published: November 18, 2026, announced by email under the subject "Just released: Global Insights report on the consumer journey"
  • Length: seven pages, four of them with data
  • Survey: 22,000 consumers in 21 countries and 2,020 marketing and advertising decision-makers in 25 countries, fieldwork by Sapio Research in February 2026, margins of error of plus or minus 0.7 points (consumers) and 2.2 points (marketers) at 95% confidence
  • Spend data: average daily ad spend by DV Rockerbox customers, Q4 2025 to Q3 2026, no axis values published
  • External figure: $1.46 trillion projected global ad spend in 2027, sourced to WARC
  • Products named: DV MAP, DV Neura, DV Rockerbox (Prove offering: multi-touch attribution, marketing mix modeling, incrementality testing)
  • Methodology page: https://doubleverify.com/2026-dv-global-insights-report-methodology
  • Product page: https://doubleverify.com/products/advertisers/prove
  • Company: DoubleVerify Inc., 28 Crosby Street, New York, subject to a pending acquisition by Nielsen

Timeline

Summary

Who: DoubleVerify, the ad verification and measurement company that Nielsen has agreed to buy, together with its DV Rockerbox attribution business. The survey was run by Sapio Research among 22,000 consumers and 2,020 marketers.

What: A seven-page report, "Proof Points: Decoding the Path to Purchase," finding that 17% of consumers use AI chat to research products against 39% for search engines and 40% for online reviews, that 74% say they buy throughout the year, and that 55% of marketers rely on platform-native reporting to measure the path to purchase, ahead of marketing mix modeling at 48% and multi-touch attribution at 38%.

When: DoubleVerify published the report on November 18, 2026. The survey ran in February 2026, and the ad spend data covers the fourth quarter of 2025 through the third quarter of 2026.

Where: Consumers in 21 countries and marketers in 25 countries across Asia-Pacific, Europe, the Middle East, Latin America and North America. The ad spend data comes from DV Rockerbox customers, with no markets specified.

Why: DoubleVerify argues that marketers need a clearer picture of the path to purchase to allocate a global ad market WARC puts at $1.46 trillion in 2027. The report closes by presenting DoubleVerify's own planning and measurement products as the answer.