Quality Is Job One

Just about a year ago I wrote a column that argued “in the long run, there is no stopping the shift toward exchange-based buying, but in the near future anything less than total transparency will hinder the growth of the category.”

Well, the industry met the transparency challenge head on. Today, all of the major demand-side platforms (DSPs) offer transparency, either built directly into their interfaces or through the ability to create manual whitelists/blacklists.

But transparency has a downside I didn’t anticipate: once you see exactly what you’re buying, you might long for the days of buying blind. It’s a big Internet, and there’s a lot of junk out there.

So if transparency was last year’s key issue, I’d say the critical issue the industry needs to tackle today is inventory quality.

Just to be clear, I’m not talking about adjacencies with porn or other forms of clearly objectionable content, which can be prevented with solutions like AdSafe. I’m talking about the endless list of properties that seemingly exist for no other reason than to serve ads.

Vivek Wadhwa wrote a much-discussed piece last week on the amount of search spam within Google. The flip side of the problem Wadhwa raises is that all of these search listings lead to sites that offer consumers little value, but still serve ads. And unless you make a concerted effort to prevent it, those could be your ads.

I recently tried to wade through a “safe list” of more than 20,000 sites, and while I didn’t come across any content that would get someone fired if their ad appeared alongside it, I also saw tons of sites that I would never actively choose for any brand I’m affiliated with to appear on. To come across an ad from a “real” brand on some of these sites would be completely disconsonant, like finding a Tiffany’s on Skid Row.

Now I know the exchange space has embraced the notion of “buying audience, not content,” but here’s the thing: quality content matters. It impacts performance.

Last year the Online Publishers Association released a piece of research that found “audiences are more likely to have positive brand perceptions of advertisers on sites where they have positive content perceptions.”

Since the OPA represents premium publishers, I’ve heard people snicker that the research is self-serving, and by extension presumably not valid. I disagree. For starters, I think the research makes intuitive sense. Plus, it’s a basic finding that’s been replicated numerous times in the context of print (see this study and literature review published by researchers from Northwestern University).

That’s why I think it’s time to reject “audiences vs. content” as a false choice and start a three-part quality movement in the exchange space.

One: Treat publishers like partners so they release more quality content. My sense is that buyers have approached sellers with a pirate mentality to date. We need to take the opposite view, and view publishers as partners we want to see succeed. The more money publishers make, the more likely they’ll be to release quality inventory. At the exchange and SSP level, the more that can be done to give publishers control and a path to good business outcomes, the better.

Two: Better tools, and more practice with the ones we have. I’m not saying there isn’t quality inventory in exchanges today, just that it’s difficult to weed out the junk. I’d love to see more features that make this process easier for buyers. MediaMath is on the right track with its MathSelect 250 offering, a curated list of top properties. Other DSPs have similar offerings. Buyers also need to become more adept with the tools we have. As painful as slogging through a list of 20,000 sites is, it can be done, and custom lists can be created at an agency or advertiser level.

Three: A third-party ratings system for quality. One of my frustrations with the OPA research is that it proves that quality and audience engagement matter, but there isn’t a way to rank properties on this dimension (at least not that I’m aware of). I’d love to see a third-party quality score assigned to properties, which would provide a very meaningful signal as to the value of the inventory and appropriate bidding system. This could put an end to the conflict between audience and content, and move us toward a model where we’re always buying the right audience within trusted, quality content.

What are your thoughts about how to improve access to quality inventory, either with today’s tools or future improvements?