Low-Data Advertising
Case stories about no-cookie digital advertising interest me because they offer a glimpse at what a more-private open internet might look like.
Back in 2018 with GDPR coming into effect, the publishers of the website for the Dutch public broadcasting system (Nederlandse Publieke Omroep or NPO) decided to conduct an experiment that would inform their strategy about how to (a) reduce data collection and hopefully (b) grow their advertising revenue.
The test did not confirm previously-accepted market norms, instead showing that publishers can monetize content without surveilling users.
The outcome and implications are reported in WIRED at the link below, so I will limit this post to a few highlights and questions to expand the topic from publisher-focus to marketer-focus.
First, the highlights:
NPO found that disabling 3P cookie collection for all but the small percentage of users who explicitly opted-in did not decrease their salable impressions AND actually increased their revenue.
Why?
NPO posited that selling directly and avoiding audience data intermediaries like DMPs, DSPs, SSPs and Ad Exchanges left them with a greater share of total ad spend.
Other European market studies support this, such as one from the ISBA in the UK showing that publishers generally receive only ~50% of marketer’s ad spend, with the rest going to ad tech and other participants in the supply chain.
Google is open about the fact that they operate on both sides of the seller/buyer transaction, and that in situations where both sides use their technology they collect ~30% of the total.
Under the new conditions established by GDPR, the direction that European publishers are taking may indicate the way the market may go here in the US — either as a result of new regulations, technology change, consumer demand, or some combination.
If privacy wins out and NPO’s experience is predictive, then ad revenue could shift back to content creators/producers from the audience industry. That is, unless there is a more-significant shift away from advertising-supported free content on an open internet.
And yet, there remain unanswered questions based on the write-up:
For this to work for advertisers, the important question is whether the advertising was effective. Did the low-data model drive the intended effects for the advertisers involved in the test?
Another question relates to market operations. Does this indicate a return to selling directly or through sales houses? Or will there be low-data platforms that automate the new model?
What about rates? Will there be a big reset if less audience information is available? And if so, will it disproportionately harm smaller publishers?
With some of these privacy-oriented changes starting to arrive here in America (CCPA in California, Chrome deprecating 3P cookies, the Data Dividend movement) I believe we should be experimenting more with low- and no-data models too.
I say “low-data” because there are large indications that audience data will still be part of the mix of the future, but not in the unregulated, collected-by-anybody-who-can, sold-to-anybody-who-wants, for-profit, more-is-better, cross-internet way.
Why?
Because +70% of digital ad revenue is going to companies like Amazon, Facebook and Google whose offerings are based on audience-data. And those offerings have inertia and represent the market. So high-fidelity audience targeting will likely continue to exist, but within walled gardens.
And because tech-savvy, content-oriented publishers like The Washington Post and Conde Nast are already productizing 1P audience targeting for advertisers. These will represent lower-fidelity, interest based targeting options available publisher by publisher. Ideally they can get together and standardize these, to some level, to make them easier to try and buy.
Here’s the link to the article …
https://www.wired.com/story/can-killing-cookies-save-journalism/
