Platform
Use Cases
Many Possibilities. One Platform.
AI and Automation
The Always-on Incrementality Platform
Solutions
Teams
Built for your whole team.
Industries
Trusted by all verticals.
Mediums
Measure any type of ad spend
Use Cases
Many Possibilities. One Platform.
AI and Automation
The Always-on Incrementality Platform
Teams
Built for your whole team.
Industries
Trusted by all verticals.
Mediums
Measure any type of ad spend

I got just over half of my 2026 predictions right. That's my lowest score since I started this ritual, and I'm oddly proud of it.
Every January I publish my predictions and premonitions about what I believe will be happening in our industry. In 2024 I got 90% of my predictions right. In 2025 it was 87.5%.
For 2026 I decided to take some bold bets. This time around – I only got around 60% of it correct – and that’s awesome. So – before I publish my 2027 predictions – I want to share with you a complete retrospective over each of my predictions!
I made 11 predictions, and I got 6.5 of them correct (we will get to the half…)
| The Prediction | Verdict | In one line… |
|---|---|---|
| Fake CTV, Real Fraud | Half correct | Fraud got worse; the clean-up started late and "fake CTV" never caught on |
| DOOH Goes Performance | Correct | Our clients took OOH from under 1% to ~6% of spend |
| Incrementality Goes Mainstream | Correct | Most trusted method; Smartly bought us ;-) |
| AI Everything, Everywhere | Correct | "Agentic" won the year; proof of value didn't |
| Podcast Ads Awaken | Wrong | Growth roughly halved; the money moved to video |
| Ads Come for the LLMs | Correct | OpenAI, Google and Microsoft all sell ads in AI answers |
| AI-Powered Mega-Publishers | Wrong | OpenAI shut Sora down in March |
| Retail Media Everywhere | Correct | Walmart bought Vibe; DoorDash is buying Instacart |
I also made a couple of quick takes, which were out-there: Mini-apps on iOS (Too early, no adoption data available), AI Browser Wars (wrong, OpenAI Atlas shut down – Chrome won), Privacy U-Turn (correct, privacy is no longer news, and even the EU is loosening), App Web-Stores (Correct, the largest gaming companies are reporting significant revenue gains from their own D2C sales, Google is trying to go after these revenues as app-revenues eligible to be taxed)

The prediction: A major crackdown on CTV supply chain shenanigans.
Outcome: The "worse" part landed hard. DoubleVerify reported CTV fraud schemes up 140% year on year in Q1 2026. Pixalate looked at about 2 billion open programmatic CTV impressions from March 2026. It flagged 74% of LG bundle IDs, 65% of Apple TV and 60% of Samsung as malformed, unidentified or fraudulent.
The clean-up came late in the year and from the botnet side. In July, a research group found NewsJunkie, a CTV scheme that produced nearly 2 billion invalid bid requests per day for a single seller. In August, Samsung banned residential proxy SDKs from Tizen TVs. In November, Google banned Videostrong after its backdoor pulled more than 1.3 million TV boxes into an ad fraud botnet.
Off the record, the conversation is everywhere: on trade show floors, in back rooms, in whispers. Almost everyone I talk to agrees that if someone sells you a $3 CPM CTV ad on "premium" inventory, you're buying air.
Right on worse-before-better, early on the crackdown. Half correct.
The prediction: Performance DOOH budgets rise in 2026 as measurement improves and advertisers expect real ROI.
Outcome: Our own data says yes, loudly. Across INCRMNTAL's clients, all of them performance advertisers, out-of-home went from under 1% of share of wallet in late 2025 to almost 6% at its September 2026 peak. Monthly OOH spend grew from about $4.8M in October 2025 to nearly $58M in September 2026.

INCRMNTAL client data (performance savvy advertisers) · Oct 2025 - Oct 2026
The wider market points the same way. In VIOOH's 2026 survey of 1,050 advertisers and agencies, US marketers expect Performance DOOH investment to grow 49% over the next 18 months.
Billboards went performance, and performance marketers paid for them. Correct.

The prediction: Incrementality goes fully mainstream. There's a culture clash between attribution vendors bolting it onto last-click and always-on models. New "incrementality" vendors keep popping up. "What's the incremental ROAS?" might replace "What's the ROAS?"
Outcome: More than half of US brand and agency marketers (52%) now run incrementality tests, and 36% plan to spend more on them. Google cut its incrementality testing minimum from about $100K to $5K, which puts lift tests within reach of almost any advertiser. Retailers joined in: Albertsons Media Collective launched an in-store incrementality framework, and Kroger runs holdouts that report in under two weeks.
Our own story was part of the same trend: in May, Smartly acquired INCRMNTAL to bring always-on incrementality into its platform.
The culture clash and the vendor flood both happened. Today everyone sells "incrementality," from geo-test shops to MMPs. Where I overshot: incremental ROAS (iROAS) hasn't replaced ROAS in every meeting. Three in four marketers still say their measurement doesn't deliver the speed, accuracy or trust they need.
So - Mainstream, yes. Fully mainstream, not yet.

The prediction: Peak AI hype, with plenty of snake oil. Real impact from AI creative and personalization. 2026 forces the "AI-everything" companies to prove their value or fade.
Outcome: "Agentic" was the word of the year. On January 6, IAB Tech Lab launched an agentic roadmap built on MCP and Agent2Agent. Meanwhile, the rival AdCP standard went from launch to version 3.2 in under a year. The industry now has two competing standards for robots buying media from other robots. If that isn't peak hype, I don't know what is.
Every major platform shipped an "agent": Amazon Ads Agent, Google Ads Advisor, The Trade Desk's Koa Agents, Yahoo DSP, PubMatic AgenticOS, and even our very own AURORA, an AI growth analyst!

Hype peaked, and the bill for proving value is now due. Correct.

The prediction: Podcast advertising hits a new level of growth in 2026.
Outcome: 2025 was strong: US podcast ad revenue grew 17.6% to $2.86 billion (IAB/PwC), But 2026 growth shows that the growth is happening, but that it’s is slowing down. Our own data didn't show a breakout either.
Interestingly enough – Podcast grew much faster than the stats above – but Podcast are no longer Podcasts – they are went somewhere else: Video.
71% of US podcast creators now make video versions, and YouTube took 32% of podcast listening time in Q4 2025.
I could argue the giant woke up and moved to YouTube. But the call was podcast ad growth, and it roughly halved. Wrong.

The prediction: "By the end of 2026, most major LLM platforms will have launched some sort of advertising, or at least an ad-supported tier." The format looks like search: a sponsored answer ranked by bid, next to the organic one. LLM advertising becomes the hot topic at marketing conferences.
Outcome: OpenAI announced ads on January 16. They went live in the US on February 9 for the Free and Go tiers. Self-serve Ads Manager opened in May, and the minimum spend dropped from $200K to $50K. By the end of August, ChatGPT ads were reportedly at a $1 billion annualized run rate across 40+ countries. The format is a "Sponsored" block below the organic answer, which is pretty much what I described. We even measured that they work.
Google followed. At Google Marketing Live in May, it introduced new ad formats built with Gemini that run inside AI Mode answers, including Conversational Discovery ads and Highlighted Answers. Microsoft launched Offer Highlights inside Copilot conversations. So all three of the biggest AI players, OpenAI, Google and Microsoft, sell ads in AI answers before the year is out. Elon Musk promised the same for Grok…
The holdouts are Anthropic, which bought a Super Bowl spot to promise Claude stays ad-free, and Perplexity, which stepped back from ads in February. I’m betting that Anthropic will live to regret that ad…
The right platforms, the right format, the right year, and a billion-dollar run rate. Correct.
The prediction: AI-only apps from Sora and Veo take off: endless feeds of AI-generated videos personalized to you, owned and operated by the platforms themselves, with no creators to pay. At least one gets nine-figure spending to grab market share.
Outcome: Sora hit #1 on the US App Store after its launch in autumn 2025. New downloads then fell 32% from November to December. On March 24, 2026, OpenAI shut down the Sora app and API to refocus on coding and enterprise ahead of a possible IPO. Disney's planned $1 billion investment, tied to licensing its characters for Sora, went away with it. No nine-figure UA campaign. No new mega-publisher.
AI video didn't get its own platform. It got poured into the feeds people already scroll. Lesson learned: distribution beats novelty. Wrong.

Maor is the CEO & Co-Founder at INCRMNTAL. With over 20 years of experience in the adtech and marketing technology space, Maor is well known as a thought leader in the areas of marketing measurement. Previously acting as Managing Director International at inneractive (acquired by Fyber), and as CEO at Applift (acquired by MGI/Verve Group)