Rubicon Project hit the headlines recently announcing its move to a hybrid auction model offering publishers the opportunity to monetize inventory using first-price auctions, a historic move away from its preference for second-price auctions.

The Drum caught up with Tom Kershaw, Rubicon Project, chief technology officer, at its recent publisher summit to discuss the under-fire adtech outfit’s quest to offer media partners greater transparency, plus other developments in the leadership team's turnaround program.

In this piece he gives details on the adtech outfit’s outlook on its balancing act between first- and second-price auctions, the shift towards take-rates of sub-20%, plus its move towards server-to-server header bidding, as well as efforts in supply path optimization (SPO).

Some in the industry have argued migrating to a first-price auction helps solve the transparency issues of second-price auctions, what are the challenges associated this and why are you going for a hybrid model?

To some extent saying first pricing is the only option is giving up on the industry, and not understanding that – from an auction perspective – second-price auctions drive liquidity into the system, it drives money into the system. It allows buyers to bid the true value as opposed to shaving and reducing their bids.

We need to have the correct balance between driving liquidity into the system, and making sure that we are valuing impressions correctly. We [still] believe in second-price auctions and think it is the way to go over the long term, but right now it is important that our buyers have access to the inventory they want. If we second-price, and everyone else first-prices, we will never win the impressions that our buyers want, so we have to make sure we do this correctly.

We have to start to differentiate which ads have value, and it really comes down to the audience segment that buyers are trying to reach. We are trying to use our machine learning algorithms to determine which ad requests have what levels of value and what types of auction we should run as a result.

Instead of treating everything the same, we will have high value types of ads based on audiences and formats that will have first-price auctions; others that will run second-price auctions. We don’t want to drive all the revenue into this top 10% of ads and everything else be can’t be monetized. That creates more problems for publishers than it solves, and it also means that buyers can’t fulfill their campaign objectives because they spent all of their money on 10% of the buy. The correct balance will allow us to achieve the best of both worlds.

Do you envision a time when the industry will move to first-price only?

I don’t think so, it would be a shame if it did because it would really only serve a small fraction of the transactions. Keep in mind advertisers’ budgets aren’t going up, they’re are fixed. What we don’t want to do is go through all this tech churn and all I am doing is taking money out of your left pocket and putting it into your right pocket, it is not helpful.

It is true that certain audiences and certain formats are more valuable to buyers, and we have to create the right dynamic for that. The most important thing is that we are open and transparent about it, that we don't just run a first-price auction in a cave where nowhere can see. We are saying right up-front: this is a first-price auction; this is a second-price auction. And then, participate if you want.

Can you expand on the issues you discussed on your latest earnings call, when you discussed cutting your take-rates?

That’s part of this auction dynamics conversation, price is going to be a big part of how we compete so the most efficient marketplace is going to win – that is what we are moving towards. Instead of the secret sauce stuff, it’s about who can be super fast, super efficient, super low cost in the market, and we are able to do that because of our scale.

So we are reducing our prices across the board, we have to get to option dynamics where those prices matter. Because today what you charge doesn’t really affect whether you win an auction or not, you can bid something into an auction and take a big giant fee out of it after you have won. That doesn’t make sense, if we are going to first-price we have to move towards a net bidding model where what you bid is what you pay. That way the most efficient lowest cost operation will win the impressions and will be the most effective at monetizing.

That is an important trend: reduce prices, let those prices compete in the market; and be super fast and efficient. Not only are we reducing our prices, we are going to be putting our prices on our invoices and our statements until everyone hears exactly what we are charging, so there is no more doubt or uncertainty.

Can you tell us more about the impending launch of your server-to-server side header bidding solution?

It is coming soon. We have one in the market already which is Amazon A9, we have a bunch of others we are looking at. Any solution that works and is transparent and has cookie match rates of more than 2% – we are going to do it. It is a matter of time and prioritization. Amazon A9, PreBid and Google will be the three big things we are working on this year. Hopefully we will get them all on the market by the end of the year.

How is the current industry trend towards 'supply path optimization' (SPO) influencing its strategic decisions?

Totally, everything we are doing is to make SPO work. The first part of that is to stop spamming our buyers. That’s why we bought nToggle, why we put in self-service tools so our buyers can tell us what they want to see. SPO is about picking the most efficient way to get to an inventory source, so that is why we have all these self service tools, why we are trying to classify data, why we tell buyers this is a first-price auction, do you want to participate or not? Give them the tools to make the right decision. To me SPO is about buyers’ decision making but its more about giving them the tools so they can make the decision rather than just hiding everything