Are Private Exchanges Overhyped?

For publishers, private exchanges promised to combine the efficiencies of real-time bidding with the type of control and transparency they associate with their direct sales efforts. The private exchange was the driving force behind Google’s $400 million deal for Admeld.

But as the area begins to mature, some are asking questions about its long-term viability. Private exchanges, many say, are at best a half step for publishers wary of the world of programmatic buying commoditizing their inventory even further. There are also questions as to whether private exchanges were retrofitted technology for remnant ads that was promised to work for tier-one inventory.

Major publishers such as The Weather Channel and Forbes say they’re happy overall with the progress of their private marketplaces, though both admit they’re approaching them cautiously to manage issues such as channel conflict. Neither is ready to christen their private exchange as a major revenue driver. That’s no surprise: private exchanges are hardly manna from heaven for publishers.

“A very small sliver of publishers’ inventory is being sold through private exchanges,” estimated Forrester analyst Michael Greene. “I don’t see private exchanges being a huge revenue component for them.”

The Weather Channel rolled out an Admeld-powered private exchange 18 months ago. According to Bill Murray, vp of ad operations at Weather, the exchange only began really picking up steam recently: “At this point we’re happy with how it’s working out, but it took a while to get to this point. Three or six months ago I might have had a different answer.”

Murray added, “In the last 12 months we’ve definitely seen an increase in the amount of buying occurring through our exchange. Some of that’s a reflection of the efforts we’ve put into it, some is the industry becoming more comfortable with it.”

Forbes, meanwhile, says it believes investing in the technology will put it ahead of the curve as marketer adoption of real-time bidding continues to grow. Mark Howard, its svp of digital ad sales, said it expects to eventually see a “good portion” of inventory being transacted this way. “Right now clearly direct sales are still driving most of our revenue, but the market dynamic is changing and want to be proactive in accommodating that,” he said.

Howard expects between 20 percent and 30 percent of Forbes’ inventory to be sold through exchanges — both private and public — in the first half of this year. During the second half that portion will dip to five percent, he said, as its direct sales pick up in line with seasonal demand.

Currently, inventory that isn’t sold to advertisers in Forbes’ private marketplace is tipped into public ones to sell there. That’s a common practice for most publishers running private exchanges, but Howard said he hopes almost none of its ads will end up on public exchanges by the end of the year, as adoption of its private one continues to grow.

The fact remains, though, that the adoption and escalation of private exchanges hasn’t been quick. For most publishers a relatively small amount of inventory actually ends up selling through private exchanges versus public ones or direct sales efforts. Despite the attention given programmatic buying, the truth of the matter is it is doing very little for most publishers’ bottom lines.

According to John Ramey, founder and CEO of direct ad sales technology provider iSocket, the slow rate of adoption is a symptom of a larger problem with the private exchange model itself. Ultimately publishers are always going to be reluctant to trade premium ads through exchanges of any kind, he suggested, so even private exchanges end up filled with what’s fundamentally low-quality inventory.

What’s more, buyers know that unsold impressions from private marketplaces will likely end up in an open market anyway, where they can pick them up for less, Ramey added. “Buyers would spend more if it was better inventory, but its not,” he said.

Greene concurred, suggesting publishers are so concerned with protecting their direct sales teams and their efforts, that the quality of inventory in most private exchanges is, and will inevitably remain, extremely low.

Another limitation of exchange technology for publishers as outlined by Greene, and Sunil Sharma, CEO of RTB optimization company Infersystems, is the fact that it’s currently incapable of successfully evaluating the premium nature of an impression. As a result, demand-side technologies are always going to cherry-pick the cheaper impressions from which they can generate better performance, and publishers will see little value directing more premium inventory there.

“Unless algorithms can evaluate the attributes that make an impression premium in real time, and buyers want to do it, private exchanges have no plausible way to scale at premium prices because cheaper inventory will provide better CPAs and CPCs.”

According to both Murray and Howard, their respective companies are well aware of the potential for private exchanges to cannibalize their own direct revenues if they’re not managed carefully. But both claimed their introduction has been additive in terms of revenues, allowing them to attract spend they may not have done otherwise.

Advertisers that are working with their direct sales teams on brand-oriented campaigns might be invited to the exchange to tap into more direct-response appropriate inventory too, for example, hence extracting greater spend.

But Ramey disagrees that things are going as well as some publishers claim. “Some people will smile and say it’s working, but the first iteration of this technology has failed. It’s actually just taking existing revenue and putting it in the other pocket,” he said.

Even if private exchanges are attracting new revenues from advertisers, the fact that such small volume is being traded through them means those revenues are probably incremental at best. In addition, the process of onboarding new advertisers is labor-intensive from a technology perspective, bringing into question whether the entire arrangement is ultimately cost-effective for the publisher.

According to Greene, Ramey’s analysis of private exchanges having “failed” might not be too far from reality. “To describe the results for publishers as mixed is probably pretty generous,” he said.

But he emphasized that although they might not have lived up to expectations, private exchanges have still served an important role as an on-ramp to automated trading for many publishers. The fact that they offer such close control has at least encouraged many to “dip their toes” in the area, Greene said, which could be a important stepping stone for the future of automated trading, whatever it may look like.

“As they exist today I don’t see private exchanges ever being a huge revenue component for publishers, but strategically it was an important intermediary step,” he said