Ads.txt errors – here’s what the data says

Adoption of Ads.txt by publishers has grown quickly over the last six months. While many publishers view this as a nuisance, they are deploying an increasing number of ad buys, and are adopting Ads.txt files to whitelist demand sources for individual campaigns. As implementation increases, what does the data show about management and upkeep of these seemingly simple files? What’s the frequency of errors we’re seeing? And what’s the cost of these errors?

To understand the dynamics, Industry Index tracks over 325,000 publishers with active Ads.txt files. While this represents a small fraction of the 25 million domains we track to identify technology installations and usage, it clearly demonstrates major acceptance of Ads.txt as a tool to improve supply transparency.

We utilize this data to help publishers better manage and optimize their programmatic buying, and to maximize the revenue potential of every impression. Each day we analyze the Ads.txt files of all publishers, identifying two major error types:

> Demand sources listed erroneously.
These pervasive issues are generally due to fat finger syndrome – human error. We see “spotexchange.com” misspelled “potexchange.com” in thousands of instances. Syntax errors also prevail, frequently with “direct/reseller” and “Publisher ID” values swapped.

> Misalignment of Ads.txt listings and active demand sources.
We commonly see this in two ways: An Ads.txt file lists a technology as an active demand source, but that technology is not actually running on site; Or the inverse – a demand source is active on the site, but not listed in the Ads.txt file.

In both cases, such errors create missed opportunities to monetize inventory. (We also identify other on-site tech as well as data leakage problems. We’ll explore these topics in an upcoming post.)

In the chart below, Industry Index reviews the 5,000 most trafficked publisher sites (based on Quantcast rankings). These 5,000 sites average 129 listed demand sources, and in over 150 instances these publisher sites contain in excess of 500 separate ads.txt listings – one site contains 1,310 separate Ads.txt listings. We also find an average of 8 errors per site and 1.8 warnings per site. These numbers are also highly variable, with one site containing 52 errors and another containing 246 warnings.

What does it all mean?

The obvious takeaway is that Ads.txt adoption is growing in prevalence and is certainly the unofficial, de facto tool for buyers and sellers seeking to increase transparency in programmatic relationships. Less obviously, the data shows such a breadth of differences in implementation quality. It’s clear that Ads.txt suffers from the hallmarks of youth. Fat finger and syntax errors showcase a lack attention – and a demand for tools that mitigate human error. While misalignment of tech stacks and ads.txt files illuminate the need for better upkeep systems, regular management, and—from anecdotal discussions with Ad Ops teams—a collectively understood value of the process. Errors that lead to missed programmatic revenue opportunities are a nearly universal data point. What a strange takeaway from an industry that has all but given itself over to a programmatic future.

Jonathan Shaevitz, President, Industry Index 

@jshaevitz 

Re-published with kind permission of Industry Index, the company that tracks the MarTech & AdTech ecosystem to deliver market research, competitive intelligence, and monitoring tools for tech stacks, data leakage, and publisher revenue.