‘It’s tricky and political’: Confessions of a marketer on getting log-level data from ad tech vendors

confession guy

This article is part of our Confessions series, in which we trade anonymity for candor to get an unvarnished look at the people, processes and problems inside the industry. More from the series →

More advertisers think log-level data is a remedy for many of programmatic’s woes. Originally, they used the data to check nothing shady was happening to their money, whereas now they’re using it to funnel their money through cheaper, better-performing ad tech vendors. But getting this data isn’t easy. In the latest edition of our Confessions series, in which we trade anonymity for candor, we speak to the head of display at a retail advertiser about how hard it is to get ad tech vendors to hand over this data. 

This conversation has been edited and shortened for clarity.

Why is log-level data so important to how you buy programmatic ads? 

It’s all data generated by a single impression. And if I can get it from the supply-side platforms I buy ads from then I can start to understand how they run auctions and from there focus more of money through the ones that give me the best performance at cost-effective prices. In a perfect world, I’d like to ingest log-level data into our demand-side platform so that we’re able to automatically bid into the most efficient SSPs. We’re not there yet but those vendors are providing some of the data.

Why are you only able to get some of the data?

It’s tricky and political because the data we’re asking was never meant to be shared. I have this problem when it comes to knowing how much of our media dollars go on tech fees when I buy impressions. For this to happen SSPs must get approval to disclose their take rate for every publisher, which is included in the log-level data. Often, when we get the log-level data back some of the data isn’t there or it’s not granular. That’s the tricky part because we’re not getting the same level of transparency from all SSPs and subsequently their publishers. There’s some complicity across the market and we haven’t had an outright no from any of our preferred SSPs except for one.

If log-data data is that hard to wrangle, will it ever be truly scalable?

We’re trying to do what we can to get better transparency into our programmatic auctions by cutting down the number of exchanges we work with. We’ve dropped 22 exchanges over the last two years. That’s been led by our in-house team. When it comes to accessing log-level data from SSPs, however, we ‘re bringing in the product and analytics team at our DSP to help us get to that next level.

Is the challenge made harder by the fact that Google won’t share that data?

Google isn’t in the business of sharing data. The fact that they’re not willing to share data to help us improve our programmatic strategy doesn’t set us up for success. We’ve started reducing the number of impressions we buy from its exchange but I can see why so few advertisers have done the same thing. I know a lot of marketers still prioritize reach and competitive CPMs, which makes it harder to turn away from the exchange with the most impressions for sale. My team is more focused on the effectiveness or ROI of what we buy, not the amount.  

Does it help when you’ve got a close relationship with SSPs?

We have a good relationship with the eight exchanges we’re working directly with and are getting a lot of the log-level data we’ve asked for.But those relationships are more than log-level data. We’ve paired buyers in our in-house team with each exchange to review everything from the data to the take rates that are disclosed as well as looking at when we need to move money between private marketplaces and the open market.

https://digiday.com/?p=367084

More in Media

Incoming teen social media ban in Australia puts focus on creator impact and targeting practices

The restriction goes into effect in 2025, but some see it as potentially setting a precedent for similar legislation in other countries.

AI Briefing: Amazon’s new Nova models boost AI model efficiency, accuracy and variety across AWS

One of the most buzzy debuts was Nova, a suite of six new AI models that include understanding and creating text, images and videos.

Q&A with Jessica Chan, Perplexity’s head of publisher partnerships

Perplexity’s new head of publisher partnerships Jessica Chan shares how the AI tech company is wooing publishers, from what the program offers now to what she hopes to add to the program next year.