Is audience extension for everyone?

Published by: Gavin Dunaway , AdMonsters
Published on: January 29, 2013

Hey Monsters,

I was promoting part 2 of my audience extension - - piece when a trusted industry pro suggested that audience extension is only for a "select group" of pubs. But when I was researching this piece, I was surprised at the variety of pubs I talked to – newspapers, general info, niche info, financial news. It just seemed to me that audience extension is an opportunity for any site with an audience. Of course, sometimes this audience could be too small to make it worthwhile, which is where techniques such as lookalike or act-alike modeling become interesting – not just for targeting similar audiences, but also for building a publisher's audience (by drawing them from a third-party site to home base).

Perhaps my bigger question is what is the potential to separate audience extension efforts from direct sales? One pub we talked to has its own media-buying team – is there an opportunity here for pubs to pretty much integrate a targeting platform as a revenue stream?
I also want to mention that Kerel Cooper will be leading a session at the Publisher Forum Sonoma on Advance Digital's impressive experience with audience extension – I believe few people on the publisher side can talk about this subject quite as well as him. Read more about it:

To summarize, I put it to you guys: Is audience extension something a majority of publishers can take advantage of? Why or why not? And is there a bigger opportunity here beyond bolstering direct sales?


I won't make it to Sonoma, but wanted to respond to this - I think audience extension is indeed an opportunity for any publisher that supports a direct sales organization and has high sell through against any particular piece of inventory. At least, it's an opportunity for the near term.

I think longer term, audience extension will probably only work for publishers that have large reach against very specific and valuable audiences - think finance, health, ecommerce, and niche B2B segments for decision makers of things like IT purchasing.

The big brands are likely the only ones that can make more money selling their data themselves than offloading it to an aggregator or data broker for years to come. The reason is that buyers will demand specificity and granular targets, will want to know a lot about what drives performance, and you just won't have many folks that can support that level of sophistication.

A regional newspaper chain isn't going to be able to afford a data science team. And I don't think buyers will put up with the hassle of having someone else control the targeting, segment construction, and execution if that party doesn't have the market cornered on that audience. If there are other choices in the market, even slightly inferior ones, it'll be more efficient.

You also have to realize that data science is an amazing thing, and predictive modeling often puts up a really good fight from a cost efficiency point of view than just picking out what seems like the most obvious match to the target segment.

In the longer run, what will likely make sense for everyone is to just sell the data and let the agencies figure out how and where to buy. Pubs will set their terms and price, and they'll activate it as buyable segments in programmatic buying desks that are wired up to exchanges where they can place guaranteed buys on pubs, or just make spot buys. As long as the publishers can work out agreeable economics, I think the most efficient way to use the data will win out in the end, and that won't look like the current audience extension model.

Rocket Fuel