Retargeting: A Scam or ROI Harvester?

When affiliate marketing and online media intersect, marketers are well aware of the perils of attributing a single order (or lead or other monetized action) to more than one proximal click. What many marketers apparently don’t know is that many vendors that sell search retargeting (the ability to buy display ads targeted at former site visitors) are taking credit for orders and leads based purely on a view-though conversion, and doing so in a way that can only be called a scam.

How did this come to pass? Unfortunately, undereducated advertisers have often fallen for “performance deals” that include both a view-through and click-based conversion and insist on their own tracking methodology. Additionally, it used to be that only the largest 5 percent of search advertisers were pursued by retargeting vendors, but recently this has changed and confusion about billing, attribution, and inventory sources now makes it difficult for thousands of mid-sized search advertisers to understand the differences in what true value vendors can deliver (or are delivering).

In case you missed it, the buzz regarding search engine visitor ad retargeting (also known as search retargeting) has long been hot at many conferences and within online discussion forums. In just a few short years, the opportunities to retarget prior site visitors (particularly those who arrived via search engine listings, both organic and paid) has blossomed from a few general ad networks and a couple of specialist ad networks to dozens of providers.

The premise is clear: fewer than 5 percent of search visitors convert to a lead or sale, yet those visitors are clearly in-market, or at least likely to be highly interested (due to their search behavior). Following those visitors around with targeted advertising as they read news, check their Web-based e-mail, and visit social networks makes perfect sense given the relatively low cost of impression inventory within exchanges.

One catalyst for the explosion of retargeting providers has been the growth in inventory within the ad exchanges (Google’s AdX and Yahoo’s Right Media, in particular). Auctions within these exchanges share a lot of similarities with the auctions of paid placement search. Some exchange auctions are real time (conducted on an impression by impression basis, called RTB or real-time bidding); others allow for bids set at the segment (group of cookies) level with changes in bid for the segment made via API (define) or the user interface.

With Google’s AdX adding inventory all the time to its real-time bidding inventory and established exchange players like Right Media continuing to provide significant inventory sources, it’s no wonder ad networks have been joined by an assortment of demand-side platforms (DSPs) in offering both search retargeting and general retargeting to marketers.

What bothers me, however, is the fact that there’s a huge opportunity for non-seasoned marketers to get caught in a situation where they dramatically overpay for the retargeted inventory, all the while thinking they’re getting a great deal.

Here’s how it works (regardless of billing method). Retargeting systems (including the one you can use within Google’s self-serve AdX) all require a pixel be used on your site to capture and set a cookie (or cookies). These third-party cookies typically serve one or two purposes (often both). The cookies allow for the retargeting process; additionally, one of the cookies is associated with a required pixel placed on the thank you page of your order or registration process. Here’s where things can get dicey. It is possible to track not only orders from inbound clicks but also view-though orders (where an ad was rendered in a browser and then an order was placed).

However, if view-through orders are to be compensated on a CPA (define) basis (not just used to validate that there was lift due to direct navigation, post-impression, or media interaction effects between display and search), then the ad network or retargeting provider has an incentive to buy below-the-fold, less expensive inventory on the hope that some of the search visitors were:

1. Going to come back anyway to your site because they were early in the buy cycle
2. Going to get exposed to another form of media that drives them back
3. Good customers of yours who returned on their own
4. Responding to one of your CRM (define) initiatives and returned to purchase/transact

In each instance, the provider gets paid on faith alone.

View-throughs can be very useful for comparing media or creative effectiveness. However, I suggest that you use view-throughs only if you are willing to set aside 10 to 15 percent of the media to run:

• A PSA ad
• A non-competitive ad from your technology provider or agency’s client stable (if that non-competitive advertiser is willing to reciprocate on its testing inventory)

By comparing the view-through data of your test ad to your live ad, you’ll know what your true measured lift is. You may, of course, also have received brand impact and non-measureable response.

Use search retargeting if your site is large enough that the cookie pool can be reached through the ad exchanges, but don’t pay for results you aren’t actually getting.