Why everything you’ve heard about (in-app) header bidding is fake news
In his keynote speech to this year’s 2018 IAB Annual Leadership Meeting, Keith Weed, Unilever chief marketing and communications officer, stated that he will ‘only partner with organizations which are committed to creating better digital infrastructure, such as aligning around one measurement system and improving the consumer experience.’ The Drum reported that Weed pledges to cut ties with ‘platforms that create division’, attributing this decision to a loss of trust in digital advertising. Weed’s address, shows that it is increasingly important to rebuild trust between advertisers, publishers and tech providers. A step in the right direction would be to debunk all the myths surrounding header bidding solutions.
header bidding image
Header bidding continues to be a white-hot topic and for good reason: it allows publishers to find the best buyers for each and every impression, and drive ROI, allegedly. The amount of good this does cannot be overstated. Profitable publishers can stay in business and offer great content at little or no cost to consumers. Unfortunately, there’s quite a bit of fake news surrounding header bidding for in-app inventory, and it’s time to dispel the myths.
Plus ça change, plus c’est la même chose
The advances in digital advertising haven’t benefited publishers and advertisers equally. For a long time, publishers viewed programmatic as merely a way for buyers to access their premium inventory at cut-rate prices. Armed with sophisticated data, buyers could easily find their target and reach audience on the exchanges, foregoing costly direct deals.
Header bidding promised a more level playing field, allowing publishers to include many more buyers and pick the highest bidder. In this ideal situation, whenever an impression became available, the header bidding technology would solicit bids from all demand sources and classes -- buyers from direct campaigns, ad exchanges, ad networks or trading desks -- simultaneously in a super auction.
This isn’t the reality, especially for in-app inventory. What passes for header bidding is often a smokescreen. The reality is that this “fair” solution isn’t much of an improvement to what came before. Indeed, many header bidding solutions are little more than add-ons to existing desktop exchanges. This is hardly what app publishers had in mind when they sought to increase demand, as these integrations omit a whole range of demand connected via the SDKs on the client side.
The hybrid solutions offered to increase demand sources are equally disappointing, and their providers have no business calling them header bidding. First, the auctions are waterfalls by another name – or at best, supplements to them – as they require ad exchanges – usually desktop – to compete first, with the winners of those auctions going on to compete against a publisher’s classic mediation set-up. There have been cases where 'header bidding' vendors, who only represent a sliver of the overall available demand, start waterfalling their own header bidding solution, running it a few times at different floor prices. This is absurd and exposes the vulnerability of their so-called solution. Second, there is no transparency regarding details of the auction process itself, and its various steps. Is the auction first- or second-price? If it’s the latter, the bid that’s sent to part two of the auction is lower than necessary, and the publisher loses if a buyer in the mediated network is willing to pay a price that’s higher than the second-price bid.
In order to achieve a real ‘state of header bidding’ for in-app, we need a truly fair marketplace, with all demand sources able to participate in a single auction simultaneously and winners determined by the price they’re willing to pay for a user. This is a win-win situation, because it means that buyers get a fair shot at securing the users they value most, and publishers are fairly compensated for their audiences.
But how do we get there?
First and foremost, we need to allow mediated networks to compete in a real-time auction simultaneously with DSPs, foregoing the waterfall altogether. All buyers, whether they’re buying via an SDK or an exchange, must be able to see and bid for all ad requests at the same time, and not just the ones that trickle down to their spot in the sequential mediation. The entire process must be transparent, allowing publishers to see exactly what buyers are offering for their inventory, so they can better understand demand and create better pricing strategies.
Second, our industry needs to apply basic economic theory to our markets. Increased competition delivers greater efficiency, which benefits all players in the ecosystem. If we persist in obfuscating the auctions that fuel header bidding, frustrated buyers and sellers may revert back to the labour intensive direct deal sales, eschewing all of the efficiencies that programmatic delivers.
Truly fair auctions can and must emerge as the industry standard for the so called ‘header bidding’, with the backing of both SSPs and demand partners. When such auctions do become the norm, publishers will be able to enjoy higher yields and greater flexibility with the same tools they already use. And all buyers will have a fair shot at winning the impressions they value most. Then, and only then, will we truly achieve what true header bidding set out to achieve - a fair, transparent, simultaneous auction where the only factor determining the winning bid, is the bid itself.
Jonnie Byrne is Fyber's managing director, UK & France