Demand Optimization: A New Ad Tech Function

As a buyer of upfront media, understanding the optimal way to spend your money is probably your number one priority. Is there a better way to spend the money to buy upfront media, so that the outcomes are greater than those from the last time I purchased? To answer that question, a buyer needs to understand two things: the value of the media and the price of the media in the market. What's the difference? Value, is a way to understand how much separate items are worth to you as compared to each other, while the market price indicates how much the last buyer and seller agreed that a single item was worth, between the two of them. Markets are created because different buyers and different sellers each have a belief of what their value is, and they would like to find each other and transact. So, markets are created when there is a landscape of bids to buy and offers to sell that can be meaningful to both the supply side and the demand side. This way, buyers can find sellers and sellers can find buyers more easily. That is precisely what our two-sided market model for upfront media does. (explained in an older post)

An illustrative example: you are standing on the floor of the imaginary New York Advertising Futures Exchange, you want to buy a piece of inventory at $10 CPM for February flighting. You shout out "I have a buy order for Males, 18-25, in an automotive context at $10 CPM, 10 million impressions, for a February flight." It just so happens that there are a bunch of publishers, sellers, on the floor of the exchange. When you shout out your order, they can hear it. If they like your price, the publisher can offer what they have that matched the specification of what you ordered. Alternatively, none of the publishers may like your offer, so you hear a bit of silence and then a shout back " 7 million impressions, males, 18-25, in an automotive context, $10.80 CPM, for a February flight." Now there is a market! Other publishers may step up and offer at a lower price, or another buyer may find the offer priced correctly and take that inventory off the market. What that means for buyers is that they can buy the inventory they find valuable, at the offering price, or they can choose to put in bids on inventory to tell the seller that they want that inventory, just not at the price at which it is being offered. This means that supply is competing for demand, not just demand competing for supply.

For inventory from a specific publisher, the exchange manages those conversations so that an order shouted out "I have a buy order for Males, 18-25, in an automotive context at $10 CPM, 10 million impressions, for a February flight, from Yahoo!" will only be heard by Yahoo!

So, what is demand optimization? In short, it is a way for buyers to control a much larger set of conversations with sellers, some public and some private. The more information you have for decision making, the more nuance can be had in the conversation with the seller. MASS Exchange collects the bids from buyers, the offers from sellers, classifies them, stores them, applies market participant rule about who wants to do business with whom, and then makes them searchable. If one of the offers and one of the bids match, the exchange executes that trade for upfront media. This might seem like what other ad exchanges do, which is true. The difference is that MASS Exchange does it for orders for future inventory, not impressions that have already arrived at the publishers page. Hence, an advertising futures exchange.

Demand optimization is drastically different from what today's DSPs do. In short, demand optimization is all about figuring out all the upfront inventory that one could buy to meet the campaign objectives, comparing  those inventory options based on price and previous performance, and empowering media buyers to buy the best performing inventory at the best price. It's all about taking the media buyer's decision making powers to the next level. It's not machines replacing people, it's people making better decisions powered by machine driven data and analytics. Simply put, it is a tool to help organize a massive parallel conversation happening between thousands of and buyers and thousands of sellers. Demand optimization allows buyers to virtually survey the landscape of all inventory available in the market place, identify if it meets their specification, understand the clearing price, and select the best inventory to meet their needs across hundreds or even thousands of sellers, with millions of possible inventory options.