In App Bidding
It’s been a couple of years since we heard the buzz word “in-app bidding” for mobile ad monetization as it seems to be the next big thing.
In theory, it should lead to more automation, better fill, less integration work (SDK, adapters..) and higher eCPM.
But is it going to meet the expectations?
I have a few doubts about it, let’s look into it:
- Inventory: Video ad-networks which are now delivering impressions through these solutions, apart from Facebook Audience Network and mediation-owned networks, are frankly not very relevant providers anymore. Those with high demand/supply are reluctant to step into this technology, so I’m wondering if there will even be competition in the auction as one network will be over performing all the others anyway
- Auctions: When an advertiser launches a campaign to get an install, the bidding model is usually based on CPA/CPI so in a traditional mediation the publisher will be paid only if the user watches the advertisement and the install happens. In this case the “risk” (not a real risk, but more taking responsibility) is on the advertisers as they will decide to participate in an auction based on the conversion rates they observe for a specific target audience. When it comes to “in-app bidding”, the ad networks will have to translate the CPI of the advertiser to a CPM, giving a price to a single impression and participate in the auction. In case they win it, the impression will be paid to the publisher, regardless of whether the install does or does not happen, but the advertisers will pay only in case of an install received. Are we sure that networks won’t try to contain these costs bidding proportionally less than the CPI agreed with the advertisers? Also, providers will probably not participate in all auctions, but only in the relevant ones based on the user profiles to maximize the chance of an install to occur and avoid bidding on not compatible audiences. This will leave many auctions filled by lower quality impressions (possibly from programmatic/RTB networks) and potentially lower eCPM on average
- Automation: for the reasons mentioned above about inventory and auctions, a completely automated waterfall might not guarantee the higher eCPM per segment (app/platform/geo): this is why having a hybrid solution where certain instances are defined by the ad monetization team seems to be the only reliable option, at least for the moment.
- Transparency: in every platform that comes with automation, as it happens with traditional mediation, it all comes down to trusting an algorithm and make sure it’s not biased. As it will work as an unregulated market, unlike the stock exchange, transparency will be key.
What is your opinion of in-app bidding until now?