Too often, marketers and their agencies make supply path optimization decisions based on business relationships, marketing materials, and intuition. But thanks to industry transparency initiatives like ads.txt and sellers.json, buyers can now make data-driven supply path optimization decisions. In particular, we think buyers need to use data to answer three key supply path questions:
Which exchanges provide the necessary scale to power my media buy?
How directly is each exchange integrated with its publisher partners?
Can I disable reselling, or is there resold inventory that I should continue to buy?
The charts below use publicly available data to bring facts to these three questions. These metrics are the output of a daily-updating analysis of the 10,000 highest grossing ad-supported websites, mobile apps, and CTV apps. We draw on Jounce Media’s crawl of over 200,000 ads.txt and app-ads.txt files and over 250 sellers.json files as well as daily-updating spend metrics contributed by our buy-side advertising clients.
Question 1: How big is each exchange?
There are over 400 sell side technology companies listed on publisher ads.txt files, and a single DSP campaign commonly bids into auctions operated by 50 or more of these companies. But the publishers and placements available through these exchanges are highly overlapped, and advertisers are finding that they can successfully execute their media buys through a short list of 5-10 trusted exchanges. Through this consolidation process, the marketer gains greater transparency and control of its media buy, and the short-listed exchanges gain proprietary demand.
So who should be on your short list? The chart below contains data about the 25 most widely-deployed ad exchanges. The bars in dark blue represent the percentage of global ad-supported publishers who indicate they work with each exchange (via ads.txt files). And the the bars in light blue represent the subset of those publishers whose integrations are confirmed by each exchange (via sellers.json files). We think the light blue bars are a more accurate representation of the true reach of each exchange.
There are notable specialty exchanges who do not appear in the chart below. For example, exchanges who power audio formats and custom native executions might be critical to a marketer’s media plan. But the list below is a good starting point for marketers who are looking to consolidate spend for standard display and video ad formats.
Question 2: How directly does each exchange integrate with publishers?
It turns out that the publisher reach of some exchanges is highly dependent on reselling arrangements. Publishers might make their inventory available through dozens of exchanges, but the practical day-to-day constraints of running an ad-supported website make it challenging to have direct business relationships with all of these companies. Instead, it is common for a media company to have direct relationships with a handful of programmatic ad exchanges and then authorize other exchanges to resell this inventory. This helps the publisher build auction density, and it can also help ensure every marketer in every DSP has access to the publisher’s inventory. Reselling, however, can also create redundant access to supply that is available through a more direct path.
To quantify reselling, we cross-reference publisher ads.txt files with exchange sellers.json files. This allows us to identify the entity that is paid by the exchange for each of the publisher’s authorized supply paths. We then classify each path as a direct path or a reseller path. There are three key points of subtlety that significantly affect the final analysis:
First, the direct vs. reseller labeling in publisher ads.txt files is often inaccurate. Paths that pay an intermediary are commonly labeled direct by the publisher. We don’t know why publishers do this, but they do it commonly, and so we ignore those labels.
Second, many small publishers authorize a third party company to manage their programmatic operations. In these cases, we consider paths that pay the outsourced yield management company to be direct. We do this by identifying what we call a “primary seller” for each website. For large publishers, the primary seller is almost always the publisher or the parent company of the publisher. But for small publishers, the primary seller is often a third party. We classify any supply path that pays the primary seller as direct.
Third, exchanges very commonly have multiple integrations with a single publisher. And these multiple integrations are often a mixture of direct paths and reselling paths.
The chart below is based on an analysis of all exchanges who are integrated with at least 20% of global ad-supported websites. We then rank those companies to identify the 25 exchanges that are least dependent on reselling. The dark blue bars represent the portion of each exchange’s publishers that are exclusively integrated via direct supply paths. The light blue bars represent the portion of each exchange’s publishers for which the exchange has at least 1 direct integration and at least 1 reselling integration. And the gray bars represent the portion of each exchange’s publishers that are exclusively integrated via reselling.
Question 3: Should I disable resellers?
Reselling introduces additional supply chain fees, which means that the marketer needs to pay more to win an impression. Reselling can also introduce latency, which means that the marketer’s bid (no matter how high) might arrive late and not be considered by the publisher. This economic inefficiency and technical inefficiency leads marketers to consider turning off resellers.
The challenge is that while some resold inventory is duplicative and wasteful, other resold inventory is unique and valuable. As an example, Taboola has exclusive right of sale for select publisher placements, but there are many DSPs who cannot bid directly into these auctions. Instead, the DSP might utilize a reselling supply path to participate in the Taboola auction. Turning off all resold inventory might protect the marketer from duplicative auctions for commodity ad units, but might also block the marketer’s access to unique ad units.
The heatmap below is intended to help marketers make data-driven choices about resold inventory. The rows (horizonal) list the 20 largest resellers whose inventory is in some way unique. The columns (vertical) list the 10 exchanges who provide the highest scale access to that resold inventory. And the numbers in each cell indicate the number of website and apps on which each exchange integrates with each reseller.
What do I do with this data?
There is no one-size-fits-all solution to supply path optimization. We recommend that buyers use this data to consolidate to a short list of trusted exchanges who provide scale, efficiency, and transparency. Buyers should then work closely with these exchanges and with their DSP to develop a data-driven strategy for managing resold inventory.
For access to the underlying data behind these charts and for help shaping your SPO strategy, reach out to us at contact@jouncemedia.com.