If you’ve spent any time in marketing, analytics, or, frankly, on the internet in the last decade, you already know the name Avinash Kaushik. He’s one of the rare voices who not only sees around corners but somehow makes the rest of us feel like we can, too.
From his sixteen-year run shaping the future at Google, to his current strategy leadership at Human Made Machine and Tapestry, Avinash has been the industry’s north star for elevating how we think, and just as importantly, how we act, in the age of AI, marketing complexity, and measurement upheaval.
Avinash is not only a bestselling author who rewired how we understand analytics, his critical writings have become required reading for any marketer serious about incrementality, accountability, and actually driving profit. He’s advised iconic brands, lit up stages across the world (including our very own MeasureUp Summit), taught at top universities, and collected a shelf of industry awards along the way.
I’m thrilled to sit down with Avinash for a conversation that digs into why last-click is still the “kiss of death,” and what it really takes to build an incrementality-centric culture. We also explore how to escape pure “data-puking” and move into applied analytics, and why concepts like IAbI and “win before you spend” are the difference between brands that merely report performance and brands that actually grow.
Let’s dig in!
Question #1
Your definition of incrementality: “If you fire everybody in marketing, how long until the company misses the team?” is provocative and clear. Given that many organizations still default to last-click-attribution-led thinking, what are the top three organizational blockers that keep brands stuck in attribution mode, and how can marketing leaders address them?
Avinash: Companies stay stuck on last-click attribution because it’s been the default for such a long time. But it is unproductive for two simple reasons: Except for non-trivial purchases, no one shows up and buys immediately, and it incentivizes renting traffic rather than owning consumer relationships over time.
Smart companies make the leap to Multi-Touch Attribution (MTA). This allows them to understand the more complex days and visits to conversion behavior, and make owned, earned, paid media optimizations accordingly. This reflects consumer behavior, and it gives deserved oxygen to non-paid initiatives.
Super smart companies, however, make the strategic leap to measuring incrementality. It's the only way to prove to the CFO: What would the company lose if we did no marketing? That in turn helps inform the size of the marketing budget and channel allocation.
Each measurement step is ever more valuable, but it also requires both technical effort (which you can build/buy/rent) and culture transformation. I wrote about this in a recent newsletter (here’s a whiteboard version).
Here are the top three insights for brands ready to make the leap to incrementality-centric marketing:
- Ensure incrementality initiatives come with a “trust but verify” mindset built-in to prove accuracy of analysis.
- Ensure your culture welcomes bad news as lessons learned vs. a path to punishment. (Initially there will be a lot of bad news, trust me, your marketing is currently not as incremental as you think.)
- Don’t judge a fish by its ability to climb a tree. Incrementality is a high-altitude initiative. Use it to make those decisions. Don’t ask: Which creative version worked better according to the Media Mix Modeling (MMM)? Which bidding strategy inside Meta in-feed ads is more XYZ? Be prepared to make big, strategic moves.
One immense benefit of working with Measured is that you get both attribution and incrementality measurement in one platform, with the super critical “trust but verify” pre-built into the platform via the capability to do large-scale experimentation.
Question #2
You emphasize the importance of moving from analytics to applied analytics and using results to change behavior, not just report it. For a marketing leader today, what is the one “small but high-leverage” change they can make right now to shift their team toward applied analytics rather than, as you termed it, “data-puking”?
Avinash: Reporting is the act of identifying what happened. The report could be a simple three-KPI dashboard or a glorious 27-tab, font-size 5, Excel data puke. Analysis is the act of identifying why. This is almost always missing from our “reports” because it is not always easy to automate the why. Yet, it is the why that drives change. It's silly to assume that just puking data on the organization translates into every recipient being able to figure out the why.
I recommend the following: NEVER send data without the IAbI:
- Insight: The why. The non-obvious patterns in the data that explain the results.
- Action: What should the organization do, given the patterns identified? Action A. B. C.
- Business Impact: If we activate A, B, C, what will the impact be on the business? Profit. Customer satisfaction. Conversion rate. Offline store visits.
The “bI” is super important because it is the one thing that creates urgency for action. Data without action is people, tools, consultants eating into your profits.
Question #3
You often emphasize creative and media-plan analysis, especially creative pre-testing and ensuring media has sufficient weight before spending. Given the rise of platforms like TikTok and algorithm-optimized creative swaps, how should legacy brands adapt their creative process to this new environment while still safeguarding brand equity and long-term growth?
Avinash: We often look at data through the rearview mirror: What was our campaign performance last month? Did the latest change we made deliver higher profits? Maybe things worked well, maybe they didn't. The challenge is the money is already spent and the opportunity is already gone.
I’ve found that using data BEFORE we spend the money, effectively looking out of the windshield, is so much more productive. I call this strategy "win before you spend." In a marketing context, my meta-analysis has identified that win before you spend comprises two critical actions:
- Pre-Test your creatives (ideally with Human Made Machine!). Your creative is responsible for 60-70% of your ultimate success.
- Ensure your media plans meet channel-level media minimums. This is responsible for 30-40% of the ultimate success.
In both instances, you're using data upfront to influence success. Then, a month or two later when you do your analysis, your dashboards are full of greens.
I obsess about the creative because of its extraordinary importance. One important evolution has been that on TV, we needed to have one or two ads, and that was enough for the whole campaign (say three to six months). On channels like YouTube or CTV choices like Disney+, you need to have four to six creative executions (of the same concept). On channels like TikTok, you need to have 25-30 creative executions!
This trend creates a big challenge for many teams. My recommendation is simple:
- Stress about the creative concept. Pre-test it with a tool like HMM Pro. Read the qualitative feedback, understand the talent analysis, deep dive into attention variations, etc. If you get the concept wrong, you're dead.
- Once you have a high-scoring concept on channels where you have high media weight on each creative, still use HMM Pro, which uses human test and control groups to pre-test. On channels like TikTok where you have low media weight on each creative, use HMM AI.
You get the best concept for your purpose (and remember, great creative concepts don’t change over years). You get creative executions that scale depending on the platform. Nailing that ensures you cover the 60-70% of the success driving factor. Winning after that will come naturally.
Reach out to a Measured expert today to schedule a demo.