« What would you like to ask Avinash Kaushik? | Main | My face, on the Internet »

May 13, 2009

Does Display help Search? Or does Search help Display?

One of the topics that we didn’t get quite enough time to cover in detail in my face-off with Avinash Kaushik at last week’s eMetrics Summit (of which more in another post) was the thorny issue of conversion attribution. When I asked Avinash about it, he made the sensible point that trying to correctly “attribute” a conversion to a mix of the interactions that preceded it ends up being a very subjective process, and that adopting a more experimental approach – tweaking aspects of a campaign and seeing which tweaks result in higher conversion rates – is more sound.

I asked the question in part because conversion attribution is conspicuously absent from Google Analytics – a fact which raises an interesting question about whether it’s in Google’s interest to include a feature like this, since it may stand to lose more than it gains by doing so (since the effective ROI of search will almost certainly go down when other channels are mixed into an attribution model).

Our own Atlas Institute is quite vocal on this topic, and has published a number of white papers such as this one [PDF] about the consideration/conversion funnel, and this one [PDF], on which channels are winners and losers in the new world of Engagement Mapping (our term for multi-channel conversion attribution).

The Atlas Institute has also opined about how adding display to a search campaign can raise the effectiveness of that campaign by 22% compared to search alone – in other words, how display helps search to be better.

However, a recent study from iProspect throws some new light on this discussion. The study – a survey of 1,575 web consumers – attempted to discover how people respond to display advertising. And one of the most interesting findings from the study is that, whilst 31% of users claim to have clicked on a display ad in the last 6 months, almost as many – 27% – claimed that they responded to the ad by searching for that product or brand:

image

This raises the interesting idea that search can actually help display be better, by providing a response mechanism that differs from the traditional ad click behavior that we expect. Of course, this still doesn’t mean that search should get 100% of the credit for a conversion in this kind of scenario – in fact, it makes a stronger case for “view-through” attribution of display campaigns – something that ad networks (like, er, our own Microsoft Media Network) are keen to encourage people to do, to make performance-based campaigns look better.

All this really means that, of course, it’s not a case of display vs. search, but display and search (and a whole lot of other ways of reaching consumers). Whether you take the view that it’s your display campaign that helps your search to be more effective, or your search keywords that help your display campaign to drive more response, multi-channel online marketing – and the complexity that goes with measuring it – looks set for the big time. And by “big time”, I mean the army of small advertisers currently using systems like Google’s AdWords, or our own adCenter. So maybe we’ll see multi-channel conversion attribution in Google Analytics before long.

del.icio.usdel.icio.us diggDigg RedditReddit StumbleUponStumbleUpon

TrackBack

TrackBack URL for this entry:
http://www.typepad.com/services/trackback/6a00d8341bf6d253ef01156f8de3c3970c

Listed below are links to weblogs that reference Does Display help Search? Or does Search help Display?:

Comments

About

About me

Disclaimer

Subscribe

Enter your email address:

Delivered by FeedBurner

Subscribe