• by Adam Soroca , January 12, 2016

The programmatic queries-per-second (QPS) arms race must end.  Isn’t the industry mature enough to stop the “mine is bigger than yours” argument?  

With access to capital challenging and an industrywide margin improvement focus, the days of buying hardware to evaluate as much of the bid stream as possible are coming to a close.  I foresee that this year programmatic buyers will compete on innovation, algorithms and superior data usage, versus the bid-request volume their systems evaluate.  

To that end, it’s time to introduce a new programmatic measurement statistic.  In all other aspects of our industry, we monitor precise metrics that detail the full-funnel analysis, from click to post-click to lifetime value.  However, little attention is paid to the efficiency of biddable systems.  

Buy-side platforms must listen to the entire bid stream or be considered inferior.

For the first innings of the programmatic game, size mattered.  That was fine.  However, a company’s coffer size and ability to spend venture capital on servers -- a commodity -- will not provide long-term marketplace differentiation.

Today even larger buying platforms experience capacity limitations as the programmatic tsunami rises.  Why?  Because the game has been to hunt through the vast growing bid stream to find bid requests that match demand.  In order to do that, as much supply as possible must be evaluated  — and quite expensive infrastructure is required to pull this off.

I have observed that the average demand-side platform (DSP) bid rate falls somewhere in the 1-3% range.  Occasionally, I bump into 5% bid rate players.  On rare occasions, I’ll talk with 10%+ bid rate DSPs.  At the higher end of the bid rate spectrum, you'll find either the largest demand sources, companies building profiles or bottoms feeders.

Out of these mostly single-digit bid rates, typical DSPs win 30%-40% of the inventory they bid on.  This means that up to ninety-nine point something percent of the traffic hitting buy-side systems is not monetized and accepted as the cost of doing business.  Meanwhile, almost every DSP wrestles with the cost necessary to run these platforms and recognizes the waste pummeling their servers.  

This means hard dollars are spent on servers and bandwidth.  Meanwhile, people, time and attention are allocated to maintain systems struggling to keep up with the traffic burden.  All of this in an industry under severe margin compression as advertisers and publishers look to claim greater share of the media dollar.

Throughout 2015, I’ve found this frustrating state of affairs is echoed by DSPs and exchanges alike.  Wouldn’t we all prefer a world where the resources (capital and people’s time) are put to work in creating differentiation?

Let me recommend adding the biddable addressable market (BAM) Index to our vernacular.  It's time to stop discussing the total QPS that bidding systems can handle, and start promoting what percentage of the addressable traffic is accessible to advertisers running on our platforms.

It’s time to change the dialogue to focus on the usable amount of traffic that our systems can handle. Let’s report what percentage of the total BAM our systems can handle, and what our bid rate is in 2016.

This would focus the conversation around innovation and customer results through better algorithms, data, pricing, service and anything new on the horizon.

BAM!  And, there it is.