Prediction: PPC Search Version 3.0, Real-Time Bidding

I’m going to go out on a limb here and predict that within three years, PPC search will look very different than today. Ads on search engine results pages (SERPs) provided by Google and perhaps Yahoo/Bing will be served via a real-time bidding algorithm.

If you aren’t familiar with real-time bidding, it’s a rapidly growing way of buying display ad inventory, particularly in Google’s AdX, AppNexus, and more recently, Yahoo’s Right Media platforms. The promise of real-time bidding is that rather than bundle impressions into a bucket (or segment) and force advertisers (and their technology partners) to pay one price for the entire bucket, it lets a technology provider bid in real time (less than 50 milliseconds) for each and every impression. While this difference may seem trivial, I can assure you that both from a theoretical and practical level there is a profound difference, particularly for marketers who want to use behavioral or profile data as part of their display strategy (for example, in search retargeting).

PPC search 1.0 was style (straight auction).

PPC search 2.0 was the hybrid auction where predicted CTR was included in AdRank.

PPC 2.5 evolved with the introduction of Quality Score.

Now we are ready for PPC search 3.0, real-time bidding.

All impressions (display or search ad impressions) do not have an equal value to us as marketers; in similar fashion, all clicks do not have the same value. Real-time bidding allows for the cherry-picking of the most valuable impressions (and, therefore, subsequent clicks) in both a display and a search environment.

However, currently within the SERP, we don’t have real-time bidding for search results, so instead, more advanced search marketers target different impression/click values by creating segments (for example, geographical segments). Segmented bidding in search makes more money for Google and Yahoo/Microsoft because each advertiser bids more for clicks that are in fact more valuable to them (in economics this concept is called utility).

To illustrate the point that publishers make more in a real-time bidding display advertising environment, Aaron Easterly, GM of advertising strategy and monetization for Microsoft, presented a scenario with assorted coins in a jar, asking how much you would pay (bid) for a random coin vs. being able to pick and pay with transparency as to the value of the coin. In real life, the benefits of real-time bidding for advertising are even more dramatic, because the utility (value) of a specific impression is based on advertiser profile/cookie information, IP address information (including all the things we can derive with reasonable accuracy from an IP address) and, potentially, third-party data.

Imagine if you will a scenario in which you could bid in real time for search impressions (showing up higher when you wanted to) based on whether the searcher:

  1. Was an existing customer or not. (Nothing predicts future transaction likelihood better than past transactions for many types of sites, not just e-commerce sites.)
  2. Had visited your site before (via search, direct navigation, an affiliate, or e-mail campaign).
  3. Viewed one page or five pages on her prior site visit.
  4. Had interacted with your social media presence (assuming that those networks allow for cookie synchronization with a real-time search bidding system).
  5. Was located within 50 miles of one of your store locations.
  6. Had, based on third-party data, visited a publisher in a category that indexed well against your target market (for example, if you were selling real estate and the user had visited Zillow and the RE section of a local newspaper site).
  7. Was on a high bandwidth connection.
  8. Was using one OS or another (assuming there was some predictability in value based on that factor).
  9. Was on an iPad or mobile device.
  10. Was capable of viewing a mini-icon (like a Favicon), mini-banner, or image you could serve up along with the PPC text ad.

Can you imagine increasing or lowering your CPC bid based on one or more of the above? If so, then you could run a more efficient – and, more importantly – a higher profit campaign in a real-time bidding environment for the SERP, assuming that Google and/or Microsoft adCenter are willing and able to open up SERP bidding in the same way they have opened up real-time bidding within their display networks. Perhaps they may even allow consumer search users to opt in/out of real-time bidding search results.

Real-time bidding of PPC search ads is great for all three participants of the online ecosystem: the searcher, the engine, and the advertiser. Let’s delve into this a bit more deeply:

Searchers: Searchers see more targeted advertising because advertisers will know if that searcher is a better fit and therefore bid more in that instance.

Search engines: The engines love relevance but the CFOs of the search engines love revenue even more (or at least as much). Yield goes up with greater transparency for a variety of reasons, even if only the top 20 percent of advertisers participate in real-time bidding (leaving the remainder to bid based on the current targeting levers).

Advertisers: Advertisers can aggressively target the most valuable searchers (to them) based on their uniquely calculated utility, saving budget for the clicks that matter most.

Lots of things stand in the way of real-time bidding for search advertising, but nearly every roadblock is surmountable, and the two big players in the space already have the assets to enable real-time bidding for search. Google has AdX and Microsoft has both its current display ad serving systems as well as a partnership with AppNexus. I could be wrong, but I believe that PPC search could look very different in the next three years and I’ve started investing resources to be ready when real-time bidding search becomes a reality.

Related reading

Screenshot shows a Google search for outdoor grills, the shopping ads shows images with “in store” showing the product is available nearby.