Despite pledges from players across the digital advertising supply chain to stop running ads on ‘made-for-advertising’ (MFA) websites, the majority of big advertisers are still having their ads show up on websites which clearly fit MFA definitions designed by industry trade groups, according to a new study from forensic ad tech business Adalytics. Adalytics’ study found that most Fortune 500 brands, as well as multinational non-profits and government bodies, are still running on MFA. And the majority of major DSPs, SSPs, verification vendors and media agencies are still playing a role in MFA transactions. This even appears to be the case for auction packages specifically advertised as being MFA free, and transaction types generally considered to be less exposed to MFA, such as private marketplaces.
The new Adalytics study was prompted by an investigation into one of its client’s digital ad spending. The client, a Fortune 500 brand, believed itself to be minimally exposed to MFA in the second half of last year. Adalytics however identified $10 million worth of that brand’s ad spend ending up on MFA websites, and as a result ran a wider study into MFA.
This wider study found instances of shocking waste resulting from MFA spend, with brands potentially paying large sums to show hundreds or thousands of ads to the same individual. Verizon for example had 781 impressions served to one person on one MFA website over a one hour period. Kroger meanwhile paid an effective cost per 1000 people reached of $5,491 to reach one individual on an MFA site.
The report doesn’t assign blame to any specific companies, since unentangling responsibility with so many players involved is impossible. But it does suggest that some products designed to help advertisers avoid MFA may not be working as intended. And it highlights how the industry is plagued by misaligned incentives across the ecosystem, resulting in massive inefficiencies for advertisers, but a lack of will or ability to really tackle the issue.
The report tells a story of how MFA has become deeply embedded in the programmatic supply chain, which tallies with conversations that VideoWeek has had with other industry participants from various parts of the supply chain. Put simply, it appears to be a system-wide failure. Brand advertisers and their procurement teams are signing contracts with agencies that require low CPMs in brand safe environments, and their bonuses often hinge on securing the lowest possible price. Then agencies are also required to deliver for the clients against these KPIs at low prices, and again there is often a bonus structure in place. Many premium publishers are often hooked on MFA in an indirect way too and they promote the links on their sites. Then there’s a commercial incentive for ad tech vendors to assist delivering all the above, and if they won’t help, their competitors most certainly will.
Still a widespread issue, despite claims to the contrary
For this study, Adalytics ran analytics across tens of thousands of websites, looking at signals related to MFA, including ad delivery, ad refresh rate, number of ad slots per page, audience traffic sources, and signs of being a cloned or templated website. From this list, the company picked out 22 specific sites which met the ANA and other trade groups’ definition of “made for advertising”. These sites all also serve over 1,000 ad impressions to a single consumer in a single short duration page view session, have paid traffic acquisitions through Taboola, Outbrain, and other sources, and were confirmed by DeepSee.io and/or Jounce Media to be MFA.
Adalytics then observed ad activity on these websites earlier this year using a panel of volunteers with a plugin installed on their web browsers. This plugin monitored which ads showed up on these sites, and looked for signals showing which companies were involved in transacting those ads.
On the advertiser front, hundreds of Association of National Advertisers (ANA) member brands were found running ads on these sites (the ANA association being relevant since the ANA has led the charge on transparency around MFA). This included Procter & Gamble, Bayer, Reckitt, AT&T, Johnson & Johnson, AbbVie, Novartis, Pfizer, State Farm, PwC, Hershey’s, Unilever, Mondelez, Mars, Haleon, Ford, NBC Universal, and Disney. Adalytics said the majority of brands which participated in the ANA’s 223 Programmatic Transparency Study, which highlighted issues around MFA, appeared on MFA websites too.
On the ad tech front, pretty much all SSPs and DSPs were seen transacting on these sites. This include Google, Xandr, Magnite, OpenX, TripleLift, PubMatic, Sharethrough, Nexxen, Index Exchange, Teads and Criteo on the SSP front, and DV360, Amazon, Xandr, Roku, Yahoo, Beeswax, AdTheorent, Adobe, StackAdapt, and Samsung Ads on the DSP side. All the major verification vendors – IAS, DoubleVerify, Pixalate, Human, Fou Analytics, and Moat were seen monitoring ads on at least some MFA sites. And all major agencies – Publicis, WPP, Omnicom, IPG, Havas, and Dentsu – were found buying ads on these sites.
Most of these companies have made various pledges and commitments around cutting down, or completely removing, MFA. Even exchanges which work with vendors specialising in identifying MFA, like Jounce Media, were seen running ads on these sites. For example, Adalytics claims the brands including Dell, Microsoft, Comcast, and Adobe had ads served on sites classified as MFA via PMP deals transacted through Sharethrough. This is despite the fact that Sharethrough is working with Jounce to remove MFA from all custom PMP deals (though this only applies to deals created after July 1st last year).
And while some argue that MFA isn’t a clear cut issue, claiming that some websites which fall under the industry’s MFA definitions aren’t necessarily ‘bad’ for advertisers, that certainly doesn’t seem to be the case with the sites identified by Adalytics. As mentioned, these sites ran very high ad loads with very high refresh rates, leading to excessive ad repetition and high prices on clearly low quality inventory.
Adalytics stressed that it’s not possible to show that any one company has dropped the ball, given the tangled supply paths each business operates in. And since it only looked at a small selection of sites over a limited period of time, Adalytics said it couldn’t draw any conclusions of whether any companies were worse offenders than others (though it’s pretty damning that most major players were found playing a role in transactions on this small range of websites). Given the massive waste associated with these websites, it’s clear something is going wrong.
An industry plagued by misaligned incentives
Given the extensive coverage given to MFA by industry associations and trade press over the past year, the obvious question which arises from this new report is why this is still happening.
Adalytics found a few examples of MFA websites using tactics to avoid detection. Some sites for example show a reduced ad load when a user navigates directly to them, and only run an excessive heavy ad load when they’re accessed via a paid traffic source. But the fact that a small company like Adalytics was able to find these sites suggests they shouldn’t be too difficult for others to identify.
One possible explanation is that ad tech tools designed to avoid MFA simply aren’t working as intended. Adalytics recommended that advertisers should be active in verifying how effective MFA solutions really are.
Another explanation is that advertisers aren’t using these tools. Some companies mentioned in the report offer MFA-free solutions, but allow brands to opt out of them if they choose to.
This would seem a strange choice for an advertiser to make, given the massive wastage associated with MFA. But while MFA doesn’t deliver real value, it can deliver on vanity metrics used by marketers to demonstrate the effectiveness of their ad spend.
For example, one unnamed Fortune 500 media executive told Adalytics that MFA performs extremely well on attribution metrics for food and beverage companies. “Most DSPs (like Amazon) use view-through attribution windows that are significantly longer than the average purchase cycle for groceries (14 days for platform reporting, and 28 days for AMC),” they said. “Consider, then, how the “flooding of the zone” on MFA sites then interacts with attribution: by ensuring as many view-through events fire as possible, they are artificially inflating ROAS results by simply “collecting” purchases versus having any evidence of causal effect.”
contracts, KPIs. we must buy for a certain low cpm and reach high viewability values with it while keeping it all brand safe. if MFA sites provide this, we are happy. our bonus depends on it.
— Malefiz (@Indecence) February 27, 2024
These misaligned incentives then spread through the ecosystem. Several SSPs for example have openly acknowledged the challenge they face in cutting out MFA sites from their supply pool. By cutting out access to MFA sites, they cut out a large amount of cheap inventory which performs well on metrics used by some advertisers. And given there’s still demand for this inventory, by switching off access to this inventory, they risk losing business to their competitors – without having any real impact on the health of the overall market.
It’s perhaps telling that the two major actions recommended by Adalytics both come back to brands. The first is for them to audit any MFA-free tools they’re working with, to see if they’re actually effective. And the second echoes the ANA’s call for advertisers to “lean in and be more active stewards of their media investments”. This might involve tough internal conversations, where marketers have to explain that their impressive delivery on marketing KPIs has been misleading. But the result will ultimately be a much healthier overall ecosystem, and better delivery for advertisers on the metrics which actually matter.
