Commentary

Another Look At Ad Exchanges

by Anonymous , March 16, 2012

In a recent article, Ari Rosenberg outlined some reasons why an exchange may not be of value to a premium publisher. The points were accurate and well-reasoned. They did not, however, cover all the potential opportunities that exchanges offer. As a counterpoint, here are some concepts that may provide value from a premium publisher standpoint.

Almost all purchases are driven by three concepts. In no particular order, these three are price, quality, and ease of use. From the start, it’s evident exchanges are designed to fulfill at least one of these concepts, ease of use. But if ease of use is the primary reason purchasers become return customers, it doesn’t necessarily imply better pricing and it certainly doesn’t assure quality.

In an exchange, the method typically used (and primarily addressed by Ari’s original article) is bulk impression purchase. If a publisher simply packages up all or most impressions and sells them on an exchange, Ari’s arguments are well-reasoned. There is little to support an exchange as a value proposition. However, what if we close the exchange somewhat? Let’s
say we create a private exchange, limit the impressions offered, break them into smaller packages (Women 18-49, Auto purchasers, Male Teens) and offer only smaller portions of the impressions we have as a premium publisher. And what if we can assure that the buyer knows it is a premium publisher rather than a random bunch of impressions from any and all sundry websites? This should create a degree of scarcity, as well as improving the assurance of another key purchase point –- quality –- while limiting quantity and potentially bringing in more bidders. This should help drive up price.

Assuming this is no sure thing, however. Even private exchanges have limitations as functional equivalents to broad-based reach purchase programs. There is still a need to make sure we are not cannibalizing from our direct sales team, nor undermining our primary ratecard. If an advertiser can buy premium impressions on an exchange at a discount, there is less need to go direct. How can we augment and avoid this, while still providing ease of use?

Each package in direct sales typically includes somewhat targeted or ROS line items. These are meant to provide some level of mildly targeted and bulk impressions at a price that brings in the CPM, while providing some additional value. Agencies recognize and accept this. As long as the pricing in the exchange does not discount these line items, there is no incentive to snipe inventory –- advertisers can’t get anything special at a discount that they aren’t already getting. Furthermore, the ease of use of the exchange (limiting the number of human hands touching a deal) reduces potential for error, and frees up the sales force to spend time actually pitching the inventory on the private exchange as “more valuable than that found elsewhere.” This actually enhances the value of a sales force, without undermining it.

Finally, as the industry moves closer to the viewable impression, or some other form of currency, the limitations this creates (as opposed to simply purchasing ads rendered) on the overall industry will place more value on all available impressions –- even those that are ROS. This should increase the bidding further, eliminating the undercurrents of low-priced, low-value
inventory that are making exchanges only appear to be glorified ad networks.

The final, additional, point that would provide value is what Online Spin columnist Cory Treffiletti recently suggested -– a futures market for advertising. While many people question this proposition, allowing publishers to offer future inventory for sale, or advertisers the chance to purchase in advance, will firm up long term revenue streams, allowing the publishers greater control of their short-term inventory, while offering advertisers the chance to jump ahead of markets and purchase at a discount.

Exchanges are not a panacea and will not solve all industry problems. But they are a good step in the right direction and worthy of greater analysis and study before simply rejecting them for the flaws they currently exhibit.

ad exchange, online publishing

4 comments about "Another Look At Ad Exchanges".

Check to receive email when comments are posted.

Brian Quinn from Triad Retail Media, March 16, 2012 at 11:32 a.m.
Why did this person feel the need to be anonymous? Are there some state secrets I missed in this piece. If the Goldman Sachs dude can say who he is, so can you!
Reply

Jimmy Bogroff from Emerge Interactive, March 16, 2012 at 2:06 p.m.

The whole premise behind the exchange concept from a publisher perspective is that it allows them to make a profit on their unused inventory. Creating a premium exchange (and by the way this is exactly what Right Media is - among others), is something that has to be publisher driven, or done in cooperation by multiple publishers; AppNexus.

It is analogous to creating Nordstrom Rack for media buyers. But here's the thing about shopping at Nordstrom rack. . you don't always find exactly what you need, or you are willing to compromise due to the price discount, and sometimes you find yourself back at Nordstrom willing to pay a bit extra to get exactly what you need. Or maybe you find yourself making a compromise on brand.

The author makes a great point when s/he points out the risk that publishers have in cannibalizing their salesforce; this is a valid concern, and its also why a premium exchange should be driven by the publisher.

The other challenge in a premium exchange is that you have to be relatively publisher agnostic. Unfortunately this is a problem as performance drastically changes by publisher.

Nike, adidas, Brooks, Reebok, Skechers and many others all make running shoes, but they are vastly different in their price-point, sizing, technology, performance, and style.

Consumers in this category are not brand agnostic.

Media buyers aren't either.

Google, Aol, Yahoo!, Microsoft, Facebook and many others may share similar audiences, or may share similar contextual placements, but each contain vastly different audience mindsets, audience segment supplies, technological capabilities, and more.

Using a current example, say you want to do an audience buy targeting males 18-34. One contextual placement that would make logistical sense in reaching this audience is via a fantasy sports buy (college brackets). The problem is that these are routinely sold in more of an upfront or sponsorship fashion. This puts tremendous pressure on the supply of available impressions.

This also forces the buyer to either take a risk on a direct placement with a second tier publisher, or to increase their bid via an exchange or DSP where the demand for this type of inventory is high at this time of year and the supply is low due to upfronts and sponsorship.

So the advertiser / media buyer / agency must ask themselves: is it worth a 25% discount / 50% discount on price when the result is a volatile delivery, or even an under-supply of their impressions, or to pay full price for the impression count that they need.

In other words, I can find a navy blue "Polo" shirt for 35% off at Nordstrom Rack, but it might not be the right size, might be improperly tailored, too long, or most importantly might not be the "Polo" brand I 'm looking for.

Because when making a purchase decision, even through an exchange / secondary market, the brand is still an important determinant in how consumers make their decisions.

Ari Rosenberg from Performance Pricing Holdings, LLC, March 16, 2012 at 2:58 p.m.

@Brian -- no cover up or big mystery here -- some folks who want to share their insights also work at very large corporations who have long approval processes in place in order to submit columns like these -- so remaining anonymous was a very honorable and selfless intention.

@Jimmy -- love the analogy -- thanks for breaking it down that way -- makes this complex process far easier to understand.

Alexander Edström from Admeta AB, March 19, 2012 at 6:37 a.m.

Hi there,

The idea must be that differentiate premium and remnant sales by clearly know who to target with the different products offerings.

Yes, they should be seen as different offerings...with the Private Ad Exchange taking care of the risk of running performance campaings by its optimization engine (per impression based) - and there by aiming its offering to e-commerce.

Premium:
Fixed price
Guaranteed impressions
Guaranteed placement
High risk with performance campaigns

AdExchange:
Auction based sales
Non-guaranteed impressions
Non-guaranteed placements (Most often Publisher RON)
Low risk with performance campaigns

In addition - it is vital to take a clear market position of who you represent when offering a AdExchange- "cannot eat from both sides of the cake" in the long run.