AdTech AI Files

#31 | Attention Drives Intention

ATTENTION drives Intention.Think about the last time you walked through a store. Hundreds, maybe thousands, of products sat on shelves within arm’s reach. You walked past most of them without even noticing. Not because they were bad products. Not because you had something against them.They simply never earned your Attention.Only the items that interrupt your autopilot, that visually stand out, that connect to a need or spark curiosity, have a chance to influence what ends up in your cart. ATTENTION is the gateway. Without it, INTENTION never forms.Now think about the media landscape we operate in every day.Endless scroll on a phone.Ads below the fold on a desktop page.Muted autoplay video competing with five other open tabs.Technically, those impressions exist. But are they truly seen, let alone have your attention?Compare that to the experience of watching television. A large screen. Leaned back. Sound on. Fewer competing stimuli. The content fills your field of vision. Naturally, attention is higher.This is why CTV is so powerful. It commands presence. It creates an environment where your message is not fighting for a single thumb swipe. It earns a moment. And that moment is the foundation of intention.But placement alone is not enough.Once you have the screen, what are you doing with it?Is your creative designed to capture and hold attention?Or does it the signal for someone to pick up their phone during the commercial break?Great CTV creative understands the stakes. The first seconds matter. Visuals matter. Story matters. Relevance matters. If you fail to hold Attention, you lose the opportunity to shape Intention.Attention drives Intention.Not impressions.Not frequency.Not viewability percentages on a dashboard.Attention.When your media environment earns it and your creative sustains it, intention follows. And intention is what ultimately moves behavior.

Transcribe →

#31 | Outcomes & AI

Feb 400:04:51Tap to summarize

In this episode of AdTech AI Files, we unpack the hottest topics from the past few months and explain why they all point to the same conclusion: outcomes matter more than ever, and most AI failures start with weak data foundations.We break down what it actually takes to make “chat with data” work in an enterprise environment. Spoiler: it’s not the model. Live data connectivity, a true semantic layer, and proper governance determine whether AI becomes a decisioning engine or just a confident demo.We also explore why most optimization problems in advertising are really measurement problems, how autonomous systems depend on trusted outcome signals, and why automation without outcome visibility simply scales assumptions faster.From platform measurability and operationalizing intelligence to why outcomes turn automation into learning, this episode connects the dots between AI, data infrastructure, and modern programmatic advertising.If you’re thinking about autonomous advertising, AI-driven analytics, or preparing your data stack for what’s coming next, this episode is for you.

Transcribe →

#30 | Viant’s Q1 2025 Earnings Recap

May 7, 202500:11:18Tap to summarize

Based solely on Viant’s Q1 2025 earnings transcript.Revenue YoY: 32%CxT: 25%

Transcribe →

#29 | Rethinking CTV Measurement for the Streaming Age

May 2, 202500:08:55Tap to summarize

𝗟𝗶𝘃𝗲 𝗦𝗽𝗼𝗿𝘁𝘀 𝗶𝘀 𝗔𝗱𝘃𝗲𝗿𝘁𝗶𝘀𝗶𝗻𝗴’𝘀 𝗡𝗲𝘅𝘁 𝗕𝗶𝗴 𝗔𝗿𝗲𝗻𝗮 | Jon Schulzhttps://www.linkedin.com/posts/jon-schulz-4780ba4_upfronts-ctv-livesports-activity-7318655564027215873-qDIg?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAOMQksB5Fgbm_qk5Eeo-7X6VF7Vp2IJNg8Bold #Upfronts Predictions from an #IndustryVeteran — With a Twist | Tom Wolfehttps://www.linkedin.com/posts/tomhwolfe_upfronts-industryveteran-upfronts-activity-7315732919719784448-b_Ct?utm_source=share&utm_medium=member_desktop&rcm=ACoAAAOMQksB5Fgbm_qk5Eeo-7X6VF7Vp2IJNg8Rethinking CTV Measurement for the Streaming Age | Steven Ohrnstein

Transcribe →

#28 | The Man, The Myth, The Legend

Mar 26, 202500:10:43Tap to summarize

Pod dedicated to the one and only, Mikhail BerylantPeople, Process, and ToolsFor any company to be successful, it takes these three things. Rarely is there a person that embodies all 3, Mikhail Berlyant is one of the greats that does.I wanted to take a moment to honor Mikhail for the #contribution he has had in my professional career. It was through Mikhail creating Potens.io, a software that sits on top of Google BigQuery, that makes it easy for someone like myself, a non-engineer who knows the business but otherwise no way to bring data to life, to automate & create solutions that empower internal employees to do their jobs and make the lives of our clients a bit easier. This Tool has been used across Viant for the last decade plus, not just Product and Engineering, but nearly every other department across the org, including Operations, Marketing, Accounting, Finance, and Sales. This same tool is the underpinning of Viant's Data Platform, which nobody in our industry can come even come close to competing with. Mikhail is on my shortlist of unsung heroes from which Viant would not exist today without. Without him, I assuredly would have attempted taking on a full-time blogging career a long time ago.Let's continue...When I first started reading Google BigQuery books around 2014-2015, the very first one I purchased was co-authored by Mikhail to my surprise. I’ve never met anyone so PASSIONATE about what they do and so DRIVEN towards excellence. His unrelenting enthusiasm on every call or meeting is unlike anything I’ve ever experienced, and I believe that’s a big part of his success.Mikhail is one of the top contributors on Stack Overflow, with a reputation score of 173,230 (no idea what this means, but for comparison, my reputation is 29). He is one of the top contributors in the world in his field which speaks volumes to his expertise.But more than anything, the only thing that gets him more excited than his work is his family, especially when he talks about his grandchildren. And that is a beautiful thing. While we work in different offices and don’t get to see each other all that often, he’s one of the few people that requires a bear hug every time I do. Thank you, Mikhail, for everything you do. I am forever indebted to your life's work.CALL TO ACTION: For anyone else that was impacted my Mikhail, whether a colleague, friend, or one of the 14.9 million people he's helped on Stack Overflow, share your gratitude in the comments and we'll memorialize this post with an AdTech AI Files pod cast a week from today.Mikhail's Stack Overflow link for the record: https://lnkd.in/gS_sZQDC#AdTech #BigData #DataEngineering #AI #Innovation #TechVisionary #GoogleBigQuery #Viant #Passion #Gratitude

Transcribe →

#27 | Conversion Rate Benchmarks

Mar 24, 202500:04:30Tap to summarize

“What’s a Good Conversion Rate?” 🤔 The Benchmark MythFor as long as I can remember, marketers have asked me:“What’s a good conversion rate for my campaign?”And here’s the honest answer… it depends.Conversion rate benchmarks are tricky because different brands have different goals. A KPI could be downloading a coupon, signing up for an email, opening a bank account, or even buying a car, all of which have vastly different intent levels and friction points.The head of Marketing Analytics for one of the largest global OEMs once told me “it’s all relative to the dashboard you’re looking at. There’s a million different options and providers of data, so measuring across providers doesn’t have apples to apples comparisons. Focus on the report you trust and make everything relative to what’s within the given report.”With that, when it comes to wanting benchmarks, try these:✅ Instead of chasing an arbitrary benchmark, focus on relative performance.✅ Compare Campaign A vs. Campaign B, not your brand vs. a generic industry average.✅ If one tactic is driving a 20% higher Conversion rate, that’s an insight you can actually act on.✅ Focus on Incrementality, if your spend is driving net new sales at a lower incremental cost per acquisition than your marketing cost, keep goingThink about it this way:Would you rather hit an “industry average” conversion rate, or actually improve what’s working for your business?Benchmarks can be a useful reference, but real optimization happens when you focus on your own data, your own customers, and your own goals.#ConversionRate #MarketingMetrics #DigitalMarketing #GrowthMarketing #Advertising #Attribution #AdTech

Transcribe →

#26 | Foot Traffic 101

Mar 23, 202500:11:47Tap to summarize

The Who, What, and How foot traffic measurement works.

Transcribe →

#25 | Google or Bust

Mar 21, 202500:05:04Tap to summarize

Will Google’s Ad Empire Be Broken Up?When I first started in programmatic, 2nd price auctions ensured advertisers paid a fair market rate. You could bid confidently knowing you’d only pay a cent more than the next-highest bid.Then, about 8-9 years ago, Google led the shift to 1st price auctions (where you pay exactly what you bid) & enabling their publishers to create dynamic bid floors (to artifically increase the clearing price). This fundamentally changed how media was bought, increasing advertiser costs while boosting Google’s profits.My First Exposure to Googles dominance was in 2016.A programmatic trader in Google’s DSP meant to raise their bid from $5.00 to $6.25 but accidentally entered $625. No big deal, they caught and fixed the mistake just 5 seconds later.In that 5 seconds, the remainder of their entire 200,000 daily impression budget was spent at a CPM of over $200. What should have cost $1,200 ended up costing more than $40k.And where did that ad spend go? Straight into Google’s ecosystem.Keep in mind Google has Unmatched ControlToday, Google dominates every part of digital advertising:🔎 Search ads (Google Search)📺 Video ads (YouTube)📈 Ad buying (DV360, the largest DSP)📡 Ad selling (Google AdX, the largest SSP)✅ Verification & measurement (Google Analytics, Campaign Manager 360)🌐 The largest browser in Chrome in the US and the largest Global smart phone operating system in Android.They own the buyer, the seller, the marketplace, devices, browser, and the referee, a level of control no other company has ever had.Will Google’s Dominance End? We’ll get an answer this summer…In the meantime, what can you do about it?PSA for the day: if you’re a marketer in any DSP, don’t think twice about using AI Bidding, if you can get screwed 9 years ago in 5 seconds, just imagine what the technology is today across any entity representing publishers. Pick a DSP that only represents the buyer, and let them work for you.🤔Anyone else have a similar experience?#GoogleAds #AdTech #ProgrammaticAdvertising #Antitrust

Transcribe →

#24 | Lift Off: The Truth About Ad Impact

Mar 18, 202500:05:50Tap to summarize

The Illusion of High Lift: Why Longer Lookback Windows Matter in Incrementality TestingMarketers love seeing high lift in incrementality studies, it feels like proof that campaigns are driving real impact. But if you’re only looking at short lookback windows (1-7 days), you might be chasing last-touch credit and overestimating effectiveness.The Problem: Short Lookback Windows Inflate LiftWhen we measure lift between 1-7 days, results look strong because ads do work, they drive awareness and may accelerate conversions. But here’s the catch:✅ Many buyers were already on the path to purchase.✅ Your ad may have just nudged them to buy sooner.✅ If you extend the lookback window, the control group converts too.Cutting off analysis too early makes it seem like your ad drove the entire conversion, when in reality, it just influenced timing.Real-Life Example: Travel & CTVA major travel brand ran a CTV and audio, with display retargeting campaign to drive bookings. Their 1-7 day lift study showed a big conversion increase. But at 90 days, they saw: • People go on summer vacations, the control group converted too, just at a later time. • Many “incremental” conversions in the short window were people who would have booked anyway. • The true lift, those booking only because of the ad, was much lower than first thought.This insight changed how they measured success, shifting to 60-90 day lookback windows to separate actual lift from simple purchase acceleration.Why Longer Lookback Windows Give a More Honest ViewThis is critical for CTV, audio, and upper-to-mid funnel media, where the goal is awareness and long-term consideration, not instant conversion.In high-ticket industries like travel, auto, and luxury retail, brands sometimes use 60, 90, or even 180-day windows for true impact measurement. If you only measure at 1-7 days, you’re likely capturing last-touch credit rather than real incremental lift.What This Means for Your Marketing Strategy🔹 Short lookback = overestimated lift. Your ad worked, but not as much as you think.🔹 Longer lookback = a more accurate read. It shows how many would have converted anyway.🔹 Test across different time horizons. Compare short and long-term effects.🚀 Key Takeaway: Marketing isn’t just about driving faster conversions, it’s about driving incremental ones. Ensure your lookback windows are long enough to separate the two.#MarketingMeasurement #Incrementality #Attribution #MediaStrategy

Transcribe →

#23 | Relative vs Incremental Lift | Pizza Talk

Mar 16, 202500:04:55Tap to summarize

Relative vs. Incremental Lift: How you can Leverage Both Let’s say you own a pizza shop, and you’re running an ad campaign to get more people ordering your delicious pies. You set up a test: • Exposed Group: They see your ad and start craving pizza. • Control Group: No ad seenAt the end of the campaign, let’s say your control group only sold 10 pizzas and the exposed group sold 25, for a total of 35 pizzas sold.🍕 What can Relative vs. Incremental Lift tell you?➡ Incremental Lift answers: “How much of my total sales actually came from the ad?”📊 (Exposed Sales - Control Sales) / (Exposed Sales + Control Sales) • This number is always between 0 and 1. In this case 15 incremental pizzas because of your ad out of 35 total pizzas sold, which means 43% of your total sales were actually influenced by your ad.➡ Relative Lift answers: “How much more did the exposed group buy compared to the control?”📊 (Exposed Sales - Control Sales) / Control Sales • This number can be over 100%, in this case your control group only sold 10 pizzas and the exposed group sold 25, your relative lift is 150%!🎯 The Purpose of Both Metrics🔍 Use Relative Lift to optimize performance between tactics:✅ Which audience, targeting, geo, or creative messaging works better?✅ Which channels are more effective within my media mix?✅ How do I allocate spend efficiently?📢 Use Incremental Lift to measure true impact:✅ Is this campaign actually driving new demand?✅ What % of sales came because of my marketing efforts?✅ Would these sales still happen if I didn’t run this ad?📌 Why It MattersIn the very first iteration of measurement many years ago within Viant’s Advanced Reporting, we focused solely on Relative Lift. We learned that if you focus only on relative lift, you might be optimizing for efficiency, without knowing total impact. We then added incremental lift, which if you focus only on incremental lift, you might miss ways to fine-tune performance. We found our customers appreciate both in the same UI. Your DSP should have both as well. 📢 Bottom LineGreat marketing isn’t just about improving results, it’s about ensuring those results actually matter.#MarketingAnalytics #Incrementality #LiftAnalysis #DigitalMarketing #AdTech #GrowthMarketing #Advertising #MarketingStrategy #Lift #AdTechAIFiles

Transcribe →

All episodes

#31 | Attention Drives Intention

#31 | Outcomes & AI

#30 | Viant’s Q1 2025 Earnings Recap

#29 | Rethinking CTV Measurement for the Streaming Age

#28 | The Man, The Myth, The Legend

#27 | Conversion Rate Benchmarks

#26 | Foot Traffic 101

#25 | Google or Bust

#24 | Lift Off: The Truth About Ad Impact

#23 | Relative vs Incremental Lift | Pizza Talk