• by Tyler Loechner , Staff Writer @mp_tyler, July 3, 2013

When it comes to big purchases, the humans like to (and should) be directly involved. This example isn't real-time bidding in the way the advertising industry thinks of it, but it's interesting nonetheless. Williams & Williams, a company that brands themselves as a "premier brand in real estate auctions," has announced that it will be auctioning off a hotel (the Ouray Hotel) and retail center located in Ouray, Colorado.

Here's the tie-in to RTM: the auction will be live both on-site and online with remote real-time bidding via AuctionNetwork.com.

Something tells me that those bidding on the Ouray Hotel and retail center will not be using algorithms to do their bidding. Using automation comes down to quality and scale, and while this probably isn't news to any of you, I thought it provided a nice argument against premium inventory being bid on by algorithms.

With millions or billions or trillions or zillions (depending on who you ask) online/mobile ad impressions available each day, it would make sense that advertisers didn't treat each individual impression like they were buying the Ouray Hotel. Some advertisers will argue that they do, noting that they only operate in premium environments and are constantly fine-tuning their algorithms to be the best. 

I'm not current with how retail auctions are run, but I would imagine this isn't the first time one is conducted with the option of bidding in real-time from a remote location. However, what is really interesting - to me, at least - is how much something like this highlights quality issues when it comes to real-time bidding for ad impressions. The issue of scale is also raised in this example, because if people were bidding on millions of hotels each day, they would rely on algorithms.

Options like programmatic direct and private marketplaces can offer viable ways to manually strike deals while having automation take care of the paper-pushing. Maybe that is the future of premium trading. Whenever scale allows you to optimize programmatically, why not do it?

However, we will only rely on the algorithms until scale and quality meet a crossroads because when the inventory is both important enough and small enough to handle manually, the humans will be hands-on.