Editorial Team Let’s be real, attributing campaign success is still a mess. That’s the problem IAB Project Eidos is trying to solve by pushing for a single, unified approach to ad measurement. It’s on us as marketers to stop looking at data in silos and start using interoperable solutions that give us the full story. This takes new tools, sure, but it also requires a complete rethink of what “success” even means on different channels. The big question is, are we actually ready to pull all these different data streams together?
Key Takeaways
- Get your data out of platform silos. You need one unified view of campaign performance to know what’s actually working.
- A real measurement framework, like the one IAB Project Eidos is pushing, is the only way to get attribution right and prove your actual ROAS.
- As third-party cookies go away, your entire measurement capability will depend on having a solid first-party data strategy and using privacy-safe identifiers.
- If you’re not constantly A/B testing and optimizing with granular data, you’re just wasting money and will never hit your target cost per conversion.
The “Connect & Convert” Campaign: A Case Study in Unified Measurement
Back in mid-2025, my team at a mid-sized e-commerce retailer kicked off the “Connect & Convert” campaign to find new customers for a sustainable home goods line. The goal was straightforward: hit a return on ad spend (ROAS) of 3.5x in three months. Our target was eco-conscious folks aged 25-45 in major US cities, and we put a $150,000 budget behind it, spreading the cash between programmatic display, social media, and search advertising platforms.
We knew from the start that we couldn’t just look at the standard, siloed platform reports. Our strategy had to align with the IAB Project Eidos principles. For us, this meant using a single common identifier, a hashed email we got from a pre-qualification quiz, so we could actually connect a user’s journey across all the places they saw our ads. Creatively, we ran authentic, UGC-style videos on social and clean, minimalist display ads that showed off the products’ sustainability certifications. The tech stack was Adform’s demand-side platform (DSP) for programmatic, Meta Business Suite for social, and of course Google Ads for search.
Initial Performance: A Disconnected Picture
The campaign ran from July 1 to September 30, 2025. In the first four weeks, the numbers coming out of each platform looked okay on their own, but they didn’t tell a cohesive story. Google Ads was showing a cost per click (CPC) of $1.15 and a solid 4.2% conversion rate on our branded search terms. Over on Meta, the video ads were getting a strong 1.8% click-through rate (CTR), and our quiz completions were coming in at a cost per lead (CPL) of $12.50. Programmatic display, meanwhile, served over 50 million impressions but had a lower CTR of 0.35% and a CPL of $21.00 when you just looked at its own dashboard.
This is exactly why the IAB Project Eidos principles were so important for this campaign. Just looking at those isolated metrics would have given us a completely wrong idea of what was working. Our analytics team, using a customer data platform (CDP) from Segment, started pulling all this data together. What we found was a huge amount of audience overlap and cross-channel influence that the individual platforms were completely blind to. For instance, a typical path might be someone seeing a programmatic ad, then a social video a day later, and then finally doing a brand search on Google to convert. In a siloed world, the programmatic ad’s contribution is zero. Invisible.
Data Synthesis and Attribution Modeling
Our job was to get away from the default last-click model that every platform pushes on you. We fired up a data-driven attribution model in our CDP that gave fractional credit to every single touchpoint along the customer’s path. The model used machine learning to sort through the more than 25,000 conversions from the first month of the campaign, and the results were a real eye-opener.
| Channel | Platform-Reported Conversions (Last Click) | Data-Driven Attributed Conversions | Attribution Delta (%) |
|---|---|---|---|
| Google Search | 1,800 | 1,250 | -30.6% |
| Meta Social | 1,100 | 1,680 | +52.7% |
| Programmatic Display | 250 | 720 | +188.0% |
| Direct/Organic | 700 | 200 | -71.4% |
The table shows just how misleading last-click is. Programmatic display, which looked like a weak performer at first glance, was actually a beast in the upper funnel, assisting in almost three times the conversions it got credit for (+188.0%!). Google Search, on the other hand, was getting way too much credit because it was often just the last, easy step for users who were already sold. This analysis totally changed our understanding of the true cost per conversion. Based on the data-driven model, programmatic’s effective cost dropped from a painful $75.00 to a much more respectable $26.00. Meta Social improved from $45.00 to $29.00, while Google Search’s effective cost went up from $30.00 to $43.00.
Optimization and Iteration
Once we had this unified view, we started making changes in month two. The first move was to reallocate budget, shifting about $20,000 from Google Search over to programmatic display and Meta Social, where we beefed up spending on video and retargeting. For the programmatic side, we built lookalike audiences from our best converters and dialed back the frequency caps on our awareness ads to stop burning people out with ad fatigue. On social, we tested more interactive formats like polls to get people engaged earlier.
Of course, the creative needed work too. We A/B tested CTAs on our display ads and found “Discover Sustainable Living” beat “Shop Now” by 15% in post-click engagement. For our social videos, we tried out shorter, more focused narratives that hammered a single product benefit, and that bumped our completion rates by 10%. We didn’t stop there. We were constantly optimizing our landing pages, using heatmaps from Hotjar to find and fix friction points in the checkout flow, which squeezed out another 0.5 percentage points on our overall site conversion rate.
These tweaks paid off. By the campaign’s end, we hit an overall ROAS of 3.8x, beating our 3.5x goal. The blended cost per acquisition (CPA) for a new customer dropped from an initial $38.00 (what we thought it was with last-click) to a true $31.00 (based on our data-driven model). We ended up with 4,838 total conversions, way more than the 3,900 we would have projected if we’d stuck with the siloed platform data.
The big lesson for us was about data governance. The IAB Project Eidos framework is right to stress having consistent data taxonomies and clear conversion definitions across every single platform. Without that grunt work upfront, even the fanciest attribution model is just guessing. Collecting data is easy. Making sure it’s clean and consistent is the real job. This means getting on the phone with platform reps to figure out exactly what their data exports can and can’t do (a task people always underestimate). And as privacy rules get tighter, our dependence on first-party data and tech like differential privacy will only grow. The bottom line is that privacy and measurement have to work together, and anyone who isn’t prepared for that is going to be left behind.
What Didn’t Work and Future Considerations
Not everything we tried was a home run. We dipped our toes into audio advertising with a small $5,000 test budget, but it gave us almost no direct conversions and was a nightmare to integrate into our attribution model because we couldn’t match identifiers. It might have helped with brand awareness, but we couldn’t prove its impact on the bottom line, so we couldn’t justify more spend on it for this campaign. We also ran into headaches with a few programmatic partners whose data was a pain to get into our CDP, requiring a lot of manual cleanup. This just reinforced that we need to be much tougher on data integration requirements when vetting new partners.
Looking forward, the marketer’s job in measurement is only getting harder and more important. The death of third-party cookies in Chrome by 2024 has already made everyone rethink tracking. We’re now looking hard at solutions like Google’s Privacy Sandbox Topics API for future programmatic work and spending a lot more time and money on our first-party data collection through better loyalty programs and content. The future of ad measurement depends on how quickly we can adapt and adopt these privacy-first, interoperable tools, the exact vision that initiatives like IAB Project Eidos promote.
Getting ad measurement right is a process, not a project you finish. As a marketer, you have to stay on top of the tech, keep learning, and be driven by the data if you want to actually understand where your ad dollars are going. It means you have to look past the easy vanity metrics and dig for the real insights that only unified data can give you.
What is IAB Project Eidos?
It’s an initiative by the Interactive Advertising Bureau to fix cross-platform ad measurement. The project’s goal is to create standards and interoperable tools so marketers get a unified, accurate view of how their campaigns are actually performing across all channels.
Why is unified ad measurement important for marketers?
It’s important because it’s the only way to know the real impact of every ad and touchpoint in a customer’s journey. Without it, you misattribute credit, waste money on the wrong channels, and can’t effectively maximize your return on investment.
How does data-driven attribution differ from last-click attribution?
Data-driven attribution uses machine learning to give partial credit to every ad a user interacted with on their way to converting. Last-click is a blunt instrument that gives 100% of the credit to the very last thing they clicked, which ignores all the upper-funnel work that got them there.
What role do Customer Data Platforms (CDPs) play in IAB Project Eidos principles?
CDPs are the engine room for applying the IAB’s principles. They’re the platforms that actually collect, clean, and unify customer data from all your different sources into one profile which is what lets you connect the dots of a user’s journey and do proper cross-channel measurement.
How will the deprecation of third-party cookies impact ad measurement?
It’s going to make things a lot harder by breaking traditional cross-site tracking. To keep measurement accurate, marketers have no choice but to get much better at using their own first-party data, along with new privacy-safe technologies and good old contextual targeting.
