‘It is not a panacea’: Why log-level data hasn’t lived up to its promise for advertisers

Log-level data was meant to be the key tool for advertisers seeking transparency into the programmatic ad bids they’ve won and lost. The reality is most advertisers aren’t asking for it. 

In vetting the programmatic supply chain for ad trade group ISBA, PwC expected it would be far more straightforward than it was to get log-level data from 11 ad tech vendors. But it took the auditor nine months and cost £1.2 million ($1.7 million) to get the job done. And even then it was only enough to track 31 million of the total 267 million impressions that were bought.

“Log-level data is really important, but in isolation, it is not a panacea,” said Sam Tomlinson, a marketing assurance partner at PwC.  “It’s just a vast amount of data that you have to store, analyze, and action. It’s necessary, but it’s not sufficient.  To be really useful, that data needs to be in a consistent format, with a unique identifier for each impression, all along the supply chain.”

Over the last two years, log-file data has been pushed as the ultimate insurance policy for those advertisers not sure about what happens to the auctions they’ve won and lost. With this data, advertisers can get valuable intel on each impression like its viewability and true cost and then use it all to prune ad tech vendors and tweak bidding strategies. Getting those answers, however, is no small feat. Some log files contain up to 150 different things about an impression. Extrapolate that across the hundreds of billions of impressions that are bought daily across a variety of different ad tech vendors and it’s easy to see why PwC struggled to make sense of it all.

“Often advertisers can be worried about the complexity of receiving raw log files and the costs that this complexity might entail,” said Cadi Jones, commercial director for Beeswax’s business across EMEA. “Volumes for win logs and bid logs can soon add up, and if you’ve chosen to work with a partner who will share your full auction logs, the volume of data can be significant.”

In fact, only a handful of advertisers like Nestle and agencies like Essence are accessing this data, and they’re going straight to ad tech vendors for it. Joey Leichman, vp of buyer development at OpenX, an ad tech vendor that shares log-level data with media buyers, explained: “Event-level data is enormous, from a volume and breadth standpoint. The potential use cases for the recipient can be limitless, but there is also a considerable investment required to produce, store and analyze the data.”.

Ad tech vendors have seen their efforts to ease those challenges barely register with advertisers. Of the 250 companies that access Index Exchange’s log-level data daily, just 10 are buyers across the U.K. and Europe, none of whom are directly advertisers. It’s a similar story at ad tech vendor TrustX where it has had no requests since the start of the year and had just one last year. It’s slightly different at Adform where there are a large number of tier one advertisers and agencies that get sent log-level data, but only the most sophisticated ones have the in-house expertise to process it. 

Related
Member Exclusive
Pitch deck: How Target is pitching its advertising network to retailers

“CMO organizations aren’t very hands-on with the technology needed to be able to crunch the data at scale, which means something like log-level data becomes a project across their team and a chief technology officer’s team, which takes up a lot of time,” said Jochen Schlosser, chief strategy officer at ad tech vendor Adform. 

While there’s an abundance of log-level data from various vendors now, they don’t always share the same thing, and when they do it’s not always in a standardized format as the head of display at a retail advertiser recently discovered. Most demand-side platforms will limit the information they share with clients to details about the impressions they have won, while SSPs aren’t always permitted to reveal the full extent of their relationships with publishers. It’s part of the reason why there’s so much disparity between what each ad tech vendor records and subsequently shares via their log-level files. At the core of this issue, is the fact that there’s no standard format for log-files. 

“If the ad tech industry wants log-files to be useful for brands then we need to have standards for how they’re reported,” said  Tom Kershaw, chief technology officer of Rubicon Project. “Without standardization, log-files aren’t a workable way to get transparency into the supply chain because of how many there are.”

There are workarounds and alternatives to the log-level data problem. None have been widely adopted yet, but they’re all pitched as more cost-effective ways for advertisers to see what happens to their programmatic bids. 

Given that many advertisers want log-file data to audit what happens to their money before it gets to the publisher, some ad tech vendors have found another way to share that information. The likes of TrustX has moved to a post-auction fee model so that advertisers know there are no buy-side fees, and publishers know they are paying a flat rate at the end of the month. Others have moved away from the log-level data approach entirely and are accessing many of the same sorts of insights using readily available APIs and standard data ports.

“I don’t believe that log-level data is the right long-term answer to this transparency problem,” said David Kohl, CEO at publisher trade body-funded private marketplace TrustX. “While we’re happy to provide the data, it often takes a lot more effort than anticipated for our clients to get a complete and accurate picture of their buys across providers. We’d much prefer to see the industry investing in a simpler and more sustainable transparency solution.”

https://digiday.com/?p=367598
Digiday Top Stories