Bunch of people have been talking about the de-averaging of pricing for awhile now, thats all RTB really is… I agree that you have to ultimately de-average the value (not the price) of the impressions but I'm not sure it benefits the buy side to de-average the actual cost. Specifically, it doesn't benefit people with technology to pay de-averaged prices as the cumulative cost for the set of desirable impressions may be more expensive than buying in bulk due to increased competition for that set of impressions along with irrational buying activity. Price discrimination will always benefit the seller, not the buyer so lets be careful what we wish for.

The notion that any unsold impression is “remnant” has allowed those with technology, data expertise, organizational alignment etc. to buy these under-utilized impressions at discounted bulk prices, transform them into something more desirable with greater value and resell them to someone else at higher rates. It is important however to distinguish the players who do add value from those who just arbitrage.

When buyers start to compete for a set of desirable impressions we create upward pricing pressure on those impressions. We cannot discount irrational buyers who will drive up overall cost even further because even those with access to more complete information will not always act in a way that is obvious to the overall market because of its own motivations which are not public, its the practically of game theory discussion.

This does all point back to a post that you made a number of months ago about data valuation however, I wonder who in the industry is pretty good that stuff…

If the old media started to: sell inventory based on its true value to marketers, create virtual demographic consolidation, use quantitative marketing techniques to serve up the right content to the right reader, etc., then they might be able to overcome their legacy cost base to return to viability (although not to the days when the Hearsts et al made a fortune by owning the printing presses.)

As for the evolving famous ESPN vs. AFF argument – there will be a value associated with advertising on each and depending on which one provides enough lift in associated metrics (placement, size of unit, time of day, day of week, audience, frequency, etc) for a particular brand and campaign.

When you go into real time price observation and discovery, this is where hills start to look like mountains in terms of gains… if your system can enable + you can action it.

Exchanges have introduced a big shift in pricing information availability. With traditional direct buys, buyers and sellers only know the clearing price or market value of a transaction if they execute the contract. You only know the price of what you buy/sell yourself. Now with exchanges, you learn the clearing price even if you don't win the auction.

The market structure for direct buys allows media buyers, especially large consolidated media buying enetities, plenty of pricing info, but they don't use it because of organizational failures. Unfortunately, media buyers generally don't have good business intelligence systems. The existing pricing data isn't aggregated and made available to individual media buyers who are negotiating. The failure to leverage direct buy pricing data has less to do with market structure and more to do with transaction processing routines. When you cut paper IO's, the pricing data is locked up in unstructured data stores — uniquely formatted Excel media plans and signed PDF IOs. The data might be available in a structured form in Atlas or DART, but these companies never added an attractive business intelligence layer on their adserving platforms.

The fact that exchange based buying is processed via API's rather than email and fax, allows pricing to be aggregated, analyzed and leveraged by buyers. IMHO, this more mundane shift in transaction processing will have a bigger impact on how data is used than the shift in market structure allowing non-transactors to discover pricing.