Kochava on October 20, 2026 sent a promotional email to the PPC Land newsroom under the subject line "CTV has an attribution problem, not a performance problem [EXCLUSIVE GUIDE]", pointing to a 26-page document titled "CTV Growth Guide 2026: From Reach to Results." The guide is co-branded by Jampp, Kochava and YouAppi and carries no publication date. Its central data claim, attributed to Kochava's multi-touch attribution analysis, is that when an app install journey contains both a connected TV touchpoint and a search touchpoint, CTV comes first 96% of the time.

In Short

Three ad tech companies published a guide saying TV ads shown on internet-connected televisions often start the path that ends with someone installing an app on a phone, even when a search or social ad gets the credit. That matters to app marketers deciding whether TV ads pay off, because the usual way of giving all credit to the last ad clicked can make TV look weak. The numbers come from the companies selling CTV ads and CTV measurement, and the guide does not publish the baselines or methods needed to check them.

What arrived and who is behind it

The email, sent from Kochava's announcements address on October 20, 2026, opens with a short argument: "Connected TV's contribution to performance marketing is systematically undervalued. The reason is clear: CTV generates intent, but conversions often happen elsewhere." The capture supplied to PPC Land breaks off mid-sentence after "Kochava's multi-touch attribution data reveals that CTV exposure", so whatever figure the email went on to cite is not visible. Both links in the email point to Kochava's tracked redirect domain, go2.kochava.com, rather than to a public landing page.

The guide itself is a different object. It is a co-marketing document. Every page carries the Jampp, Kochava and YouAppi logos, and the Jampp and YouAppi marks each sit above a small "powered by affle" line. The guide describes YouAppi as "a performance-driven DSP (an Affle company)" and cites "Aggregated internal data, Affle, 2026" for its CPM ranges. Jampp is described as a "programmatic advertising platform" for mobile and CTV. Kochava, a measurement company, contributes the section titled "Proving Performance Across Screens," and its own boilerplate on the last page calls the company "the premier provider of real-time data solutions." The closing three pages are sales pages, each ending with a "Contact Us" link to the respective company's contact form.

Who wrote what matters for reading the numbers. Two of the three companies sell CTV inventory to app advertisers, and the third sells the attribution that credits it. None of the figures in the guide has been audited by a third party, and the guide does not describe its methodology beyond phrases such as "tens of millions of installs measured by Kochava."

No date on the document

The guide does not state when it was published. The PDF file's internal metadata records creation on February 26, 2026, which fits its framing: the introduction speaks of "heading into 2026," and market figures for 2025 are written as projections. By the time it reached the PPC Land inbox in October 2026, the most recent Kochava data window it cites - Q2 2024 to Q2 2025 - was more than a year old.

The 96% and 94% figures

The guide's argument for CTV rests on sequencing. "According to recent multi-touch attribution analyses, when an app install includes both a CTV and a search touchpoint, CTV comes first 96% of the time," the guide states. "When CTV and social are both part of the user journey, CTV precedes social exposure 94% of the time."

The conditions in those sentences carry much of the weight. The 96% applies only to installs where both a CTV impression and a search touchpoint were recorded. The guide does not say what share of all installs fall into that group, how many installs had a CTV touch at all, or how the search and social touches were captured. The key findings page restates the figures with the same qualifier: "Kochava data shows CTV ad exposure precedes search 96% and social 94% of the time when those channels appear together in the same install journey."

There is also a structural reason why TV tends to come first in such paths. A CTV ad reaches a household passively; a search query is something a person does when already interested. A journey that contains both will often start with the passive exposure. That pattern is consistent with the guide's argument that TV generates intent, but ordering alone does not establish that the TV ad caused the search. The guide's own conclusion is more careful in places: "CTV is frequently the true origin point of user intent. Failing to measure it holistically risks undervaluing one of the most influential channels in the mix."

The argument closely tracks one Kochava has made before. In September 2025, the company published research finding that marketing mix modeling credited TikTok with about 35% more incremental impact than last-touch attribution, on the reasoning that ads which start journeys get undercounted when the closing click takes all credit. The CTV guide applies that reasoning to the television set.

Category lift: comparisons without baselines

A table on page ten lists what the guide calls "Incremental Lift by Category." According to the guide, users served CTV alongside other media channels "outperformed their mobile-only counterparts post-install" across "tens of millions of installs measured by Kochava." The reported differences:

  • QSR (quick-service restaurants): 14% more food and drink purchases
  • Streaming: 9% more videos watched
  • Finance: 4% more transactions
  • Banking: 3% more signups
  • CPG: 3% more purchases

The heading uses the word incremental, but the comparison described is between two groups of users - those who were served CTV and those reached only by mobile campaigns. The guide does not say whether those groups were randomised, matched or simply observed. Households reached by CTV campaigns are selected by targeting, and a household chosen because its IP address matches an existing customer list, for example, would plausibly behave differently from a mobile-only audience regardless of the TV ad. The guide also gives no base rates, no time period and no sample size per category, so a 14% lift cannot be translated into an absolute number of purchases.

Three of the five categories show gains of 3% or 4%. The key findings page highlights a different number again: "When mobile follows CTV, transactions rise 9%." That 9% figure does not appear in the category table under transactions (finance shows 4%) and matches only the streaming row, which measures videos watched. The guide does not explain which dataset the key findings figure comes from.

Conversion rates by exposure frequency

Page eleven contains the guide's most detailed chart, titled "Conversion Rate by Exposure Frequency." It plots device counts in millions against the share of devices that completed a post-install conversion "(e.g., purchase, subscription, first-time deposit)" at one to ten CTV exposures.

Read off the chart, roughly 245 million devices received a single exposure, about 90 million received two, about 45 million received three, and the counts fall into single-digit millions from six exposures onwards before a small rise at ten. The conversion line starts near 0.04% at one exposure and peaks near 0.23% at eight exposures. It is not a smooth climb: the rate falls between four and five exposures, and again between eight and nine. Every value on the chart sits below 0.25%.

According to Kochava's text, "As users receive more CTV exposures, their likelihood of completing a post-install action increases. While most impressions occur at low frequencies (1-3 exposures), conversion rates rise with additional touchpoints." The chart supports a general upward trend. It also shows that the large majority of measured devices sit at one or two exposures, where post-install conversion is lowest, and that the frequency at which conversion peaks applies to a few million devices. The chart does not say whether the tenth bar groups all devices with ten or more exposures, which would explain its uptick, and it does not state the period or markets covered.

The same page claims that "Between Q2 2024 and Q2 2025, the volume of total conversions attributed to CTV campaigns grew by 10% across Kochava's global dataset," and that "CTV is the fastest-growing source of attributed conversions." No growth rate is given for any other channel, so the second claim cannot be checked against the first.

Measurement mechanics: IP matching and windows

The guide is more specific about how CTV-to-mobile attribution works than about its own results. Two diagrams describe the flow. For user acquisition, a mobile measurement partner - the guide uses the abbreviation MMP - creates a tracker for the CTV campaign, the TV ad runs and "the system captures the household's internet address (IP) and sends it to the MMP through the tracker," the viewer installs or reopens the app on a phone, and the MMP "connects the dots between the CTV impression and the mobile action." The retargeting version starts with a seed audience of IP addresses or device IDs supplied by the advertiser.

Kochava's section lists three attribution routes. Cross-screen conversion from CTV to mobile and from CTV to web are measured through SDKs matching "on Primary IP or Household Graph." Same-screen conversion, from CTV to CTV, is matched "on CTV Device ID or Primary IP."

That reliance on IP addresses is the weak point other studies have probed. In February, FreeWheel cited research by Truthset for CIMM and Go Addressable that found IP-to-postal linkages accurate only 13% of the time. In July, a study by Adstra and InterMedia found that only 23% of residential IP addresses hit their intended geographic target, and that device-level CTV identifiers stayed stable month over month 71% of the time against 57% for IPs. The guide does not discuss match rates, shared networks, mobile devices that leave the home network, or households with several phones.

Three pitfalls, and conflicting windows

The guide lists three common setup errors. The first is "Missing data postbacks," where conversion touchpoints are not connected to the attribution system. The second is "Limited cross-screen attribution settings," where conversions influenced by CTV "may not be credited correctly" unless IP-based matching is switched on. The third is "Overly short lookback windows": according to the guide, "Default impression windows (i.e., 24-hour view-throughs common in mobile) are too short for CTV campaigns," and the stated solution is to extend lookback windows "to 3-7 days."

Seven pages later the guide gives different numbers. Under "Use Fair Attribution Windows," it recommends "a 24-hour window for same-device conversions, 72 hours for cross-screen, and a 6-hour equal-priority window to balance credit with mobile's last-click logic." The 72-hour cross-screen figure sits at the bottom of the earlier three-to-seven-day range, and the 24-hour same-device window is the same length the pitfalls page calls too short. The guide does not reconcile the two passages or explain how the six-hour equal-priority window operates.

Window length is not a technicality for CTV. A view-through conversion has no click to anchor it, so the window determines how much activity gets credited to a TV ad at all. PPC Land reported in September that MNTN and AppsFlyer, linking TV impressions to app installs in a US integration, apply a 30-day default window and a 72-hour maximum respectively, so the same campaign can produce different install counts depending on which platform's report is read.

Market figures and their sources

The guide's market section draws on outside sources, linked in the PDF's annotations though not visible in a plain text extraction. Global CTV ad spend was "projected to hit $48B" in 2025, a 33% increase from 2023, citing Southern California News Group. US CTV ad spend "grew 16.8% in 2025 to $33.48B," citing eMarketer. By 2026, 89.2% of US households will be CTV households, "surpassing 90% by 2027," with 121 million CTV households by 2027, "more than double linear TV," citing Statista. US adults spend 17.9% of their media time on CTV while brands allocate 7.4% of budgets to it, citing a 2025 Innovid report. The guide calls that last gap "pure upside for app marketers."

The 7.4% budget figure sits uneasily with other published measures. IAB Europe research reported by PPC Land in March put CTV's share of media budgets at 28% in 2025, up from 14% in 2023. The two figures measure different populations and probably define the budget base differently, but the guide presents its number without noting that other surveys put the share several times higher.

CPM ranges, from Affle's internal data, are listed as $7 to $15 in the United States, $4 to $6 in Latin America, $3 to $8 in APAC and $10 to $25 in EMEA. The guide attributes the claim that "Up to 80% of CTV viewing happens with others in the room" to eMarketer, and the figures that "88% of Americans use a smartphone or tablet while watching TV" and that 94-95% of Gen Z and Millennials do so to Arena, an audience engagement software company whose blog post is the linked source. That second screen behaviour is the premise of the whole CTV-to-mobile pitch.

Two internal inconsistencies

Two pairs of figures in the guide do not sit together comfortably. On page three, "70-75% of transactions" are "expected to be programmatic by 2025," citing Southern California News Group. On page twenty-one, "Nearly 90% of CTV ad spend is programmatic," citing a 2025 blog post by MediaSmart, a DSP. Share of transactions and share of spend are different measures, but the guide presents both as the state of the market without explaining the gap.

The Latin America page is harder to reconcile. It states that LATAM digital ad spend "passes $40B in 2025" and that the "South America" programmatic market grows from $70.22 billion to $116.12 billion by 2030, a 10.58% compound annual growth rate, citing Mordor Intelligence. A programmatic market of $70 billion would be larger than the total digital spend reported in the same panel. The linked Mordor report is titled as a Latin America programmatic advertising market study; the guide does not say whether the $70.22 billion is a different geography, a different base year, or a misreading. The growth arithmetic itself holds: $70.22 billion compounding at 10.58% for five years reaches roughly $116 billion. Brazil, according to the guide, has 89 million monthly CTV viewers and 27% year-over-year CTV growth, and APAC CTV ad spend grows 14.4% year over year, with 83% CTV usage in Australia, citing The Current.

The partner list includes products that no longer exist

Page nineteen, "Partner Ecosystem," lists key players by category. Some entries are out of date. The DSP list includes Xandr, whose successor product, Microsoft Invest, Microsoft scheduled for discontinuation on February 28, 2026. The data providers list includes Oracle Data Cloud, though Oracle announced its exit from advertising in June 2024. SpotX appears in the SSP list next to Magnite, the company that acquired it in 2021 and later folded the brand into its own. Amobee, also listed among DSPs, was acquired by Tremor International in 2022 and subsequently rebranded; Nexxen, its successor, appears among the "Top SSPs" logos on page five. The list names Jampp and YouAppi first among DSPs, ahead of The Trade Desk and Google DV360.

What the guide recommends

Much of the remaining material is creative and campaign guidance aimed at app advertisers: QR codes visible for the whole ad, the app icon on screen throughout, subtitles for sound-off viewing, large text because "viewers generally sit 5-11 feet away from the screen," and deep links that open an app store page rather than a homepage. A nine-point campaign checklist asks advertisers to sync creative across screens, test QR placement and frequency caps, use multi-touch attribution and "assisted conversion metrics, not just last-touch," and "Treat CTV as a performance channel, not a manual allowlist" by starting with broad supply so the bidding model can learn. That final recommendation, from companies whose business is algorithmic buying, would move more control over where ads run to the DSP.

Jason Hicks, identified in the guide as GM, Measurement Solutions, in the section produced with Kochava, is quoted twice. "CTV shouldn't be measured in a silo," reads the first. "The cross-screen, multi-channel synergies of this potent big-screen medium can only be truly understood and optimized with integrated and holistic measurement." On the key findings page: "Cross-screen is where performance happens. CTV earns attention on the biggest screen; mobile turns that attention into installs, orders, and loyalty. When you plan and measure them together, you unlock outsized outcomes."

Why this matters for marketers

App marketers have heard versions of this argument from several directions in 2026. Moloco launched a performance CTV product in April with attribution through advertisers' MMPs, claiming that about two-thirds of post-view installs happen within six hours of exposure. Google, Meta, Moloco and Unity took minority stakes in AppsFlyer in June in a round valuing the measurement firm at $2.7 billion, on the stated aim of keeping attribution neutral. The question underneath all of it is the same one Kochava's email raises: whether TV underperforms in app campaigns, or whether the measurement systems built for mobile simply fail to see it.

The guide offers a coherent answer from the sellers' side. It does not offer the evidence an independent reader would need to accept it: no control groups, no definition of the category lift baseline, no match rate for the IP addresses on which the cross-screen attribution depends, and two different recommendations on how long the attribution window should be. Kochava itself has a record of co-publishing measurement research - it co-sponsored the CIMM and 4As study published in March that found 43% of US marketers saw cross-platform measurement as a major or severe barrier. That study's finding, that confidence in measurement has not kept pace with tooling, applies to documents like this one as much as to the advertisers reading them.

The company's data practices have also been under scrutiny in a different context. In May, the Federal Trade Commission filed a stipulated order resolving its 2022 lawsuit over Kochava's sale of precise location data, barring the sale of sensitive location data without affirmative express consent. The CTV guide does not discuss consent for the household IP matching it describes.

Timeline

Summary

Who: Kochava, a mobile and cross-channel measurement company, together with Jampp and YouAppi, two app-focused DSPs that carry Affle branding. Jason Hicks, GM, Measurement Solutions, is quoted in the guide.

What: A 26-page co-branded "CTV Growth Guide 2026: From Reach to Results," promoted by Kochava as evidence that CTV has an attribution problem rather than a performance problem. It claims CTV precedes search in 96% and social in 94% of install journeys containing both, category lifts of 3% to 14% for users served CTV versus mobile-only users, and 10% growth in CTV-attributed conversions between Q2 2024 and Q2 2025. The figures are the vendors' own, without published baselines or methods, and the guide gives two conflicting recommendations on attribution windows.

When: Kochava emailed the guide on October 20, 2026. The guide is undated; its file metadata shows creation on February 26, 2026.

Where: The data is described as global; market figures cover the United States, Latin America, APAC and EMEA. The email went to the PPC Land newsroom.

Why: App marketers using last-click attribution tend to give TV little credit for installs that happen on phones. The guide argues for IP-based cross-screen matching, longer windows and multi-touch attribution, which would credit more conversions to CTV - the inventory two of the authors sell and the measurement the third provides.