In our quest to improve and empower buyers and sellers of media , we have spent a lot of time thinking about what actually happens when buyers and sellers come together. Specifically, what happens between buyers and sellers before the ad is served. We quickly realized that ad tech can be broken down into three teams: those who work to understand what has happened after the ad was served, those who work to understand what you should do when the ad is served, and those who work to understand what will happen in the future. We focus our work on understanding what buyers and sellers are telling each other they want to happen in the future. When a buyer and seller share the same desire for opposite sides of a transaction, we bring them together. So, there are three areas on the timeline: the past, the event horizon, and the future. What is the event horizon you ask? (a term borrowed from physics) "The point of no return" - the moment at which an advertising insertion decision needs to be made. In digital, it is the moment at which an ad call is made from the web server. In print, it is the moment at which files go to press. In television, it is the moment at which the broadcast file is finalized. How supply and demand are understood is different in each area on the timeline. From an analytic standpoint, what all that means is that the "math of the past" is different from the "math at the event horizon," which are both different from the "math of the future."
The event horizon is the place on the timeline where we yield optimize; the place where insertion orders meet real-time bids. Yield optimization is a critical function, but how do publishers choose the ideal combination of insertion orders that will maximize yield opportunities? How do publishers know which deals to say "yes" to and which to turn away? The answer is to supply optimize the future. Finding the best way to sell for maximum revenue, not necessarily the highest price. In the past, this was simply impossible. The reason yield optimization at the event horizon is possible is because there are multiple options to choose from when making that final sale decision; multiple insertion orders and RTB bids competing for the same impression.
When a sales team has to deal with sales for future delivery, all they know is what they have been told by the buyers they communicated with. Further, understanding the relationship between multiple buy orders and how they overlap or do not overlap on the supply landscape, in terms of revenue impact, is simply not done. Yes, there are lots of inventory forecasting tools, but that is not the same thing. To optimize supply, the sales force needs to have the tools to see the entire landscape of demand, identify the data points meaningful to the decision at hand, and understand the implications of those decisions.
So, when dealing with inventory for future delivery, sales teams need multiple options to choose from when making that final sale decision. Because MASS Exchange is for advertising futures, programmatic forwards, we can, and did, build just those tools. In the next phase of the growth of our platform. We are building supply optimization tools to answer those questions above. In reality this is not a one-sided function, if supply can more easily find demand, than the opposite is also true: demand can more easily find supply. Its not about empowering one side or the other, it is about the belief that efficiently finding someone that wants to buy what you are selling or sell what you are buying will drive the best possible outcomes for both sides. At MASS Exchange, we win only when both sides win.
In short, supply optimization occurs when sellers can efficiently map and identify demand from which the can choose to initiate a transaction. This optimization is an outgrowth of our market model: continuous two-sided (RTB exchanges are generally one-sided second price - for those among you that want something a bit wonkier, here's a handy reference Auctions and bidding: A guide for computer scientists) In other words, when buyers express their orders in a market that supports resting orders, those orders can be mathematically optimized to provide a seller the ideal combination of insertion orders that should generate the most revenue. For buyers, the increased efficiency of matching supply and demand enable publishers to sell targeted placements, on an audience attribute and ad placement attribute basis, with significantly smaller transaction minimums, expanding the efficiency of audience targeting at scale beyond the event horizon (RTB) and into the future.
In our next blog post, we will show buyers the other side of this "coin" - demand optimization.