NMS Blog - What we're talking about.
Click here for all categories in our blogosphere
Last-Click Analytics and Social Media Measurement, or How Search Drinks our Milkshake
Early in my career, I heard a very senior executive make the statement “All online marketing is, at its heart, direct marketing” (he was, and is, in direct marketing). Not being in a position where I could challenge this rather bold declaration, I was appalled as I was working rather feverishly on trying to make the case to clients that online could do much, much more. Think about the alternate universe where this same approach applied to offline media. Not that I’m in love with :30s, but if every one of them was a direct TV spot judged by how many people stopped what they were doing, jumped in the car and drove to the store, the TiVo would have been invented in 1972. Similarly magazines would resemble Val-Paks, and newspapers would still be screwed (sorry, couldn’t resist).
In my last post, I went into detail about how the ability of online to measure direct response has had an adverse affect on creativity, consumer acceptance, and even brand priorities. For all intents and purposes, the focus on the click has meant most online marketing has, in fact, become direct response. Two recent articles have crystallized this challenge, and have made very convincing arguments that the click should be sidelined as the be-all-end-all for online. The first is Young-Bean Song’s “Don’t settle for any old ROI” (iMedia) which is packed with stats on how much online advertising works outside the “Last Ad”. The second is Abby Klaassen’s “Why the Click Is the Wrong Metric for Online Ads” (AdAge) which gives an overview of the different industry efforts to create more equitable attribution models for display advertising. Both hinge on reams of data that show multiple messages and exposures build to make a case for the consumer versus a single impulsive click.
They also reference data that show that search ads perform much better when preceded by exposure to display ads, usually to the tune of 20% more traffic. Young-Bean cites a statistic that 72% of all paid search clicks are navigational – brand name or URLs terms – which means that other vehicles have already persuaded the consumer, but search gets all the credit for the action. This gets even more interesting when you take into account all of the various studies that show that consumers are influenced by peer recommendations and blogs at least twice as much as they are by display advertising. Taking this further to say that WOM and social media influences search much more than display advertising is not a stretch, just common sense. There are technical and privacy limitations that prevent us from tracking individual consumer consumption of editorial in the same way as banners, but that shouldn’t prevent us connecting the dots and placing a premium on what consumers already tell us. Looking beyond the last click is long overdue, and will greatly help publishers and agencies. What is needed to help marketers, however, is going the extra step and creating qualitative measures for how editorial and consumers influence each other in the same fashion. The good news is that we have some stuff in the hopper that will potentially do just that. I know that exec was wrong, and I hope I don’t have to wait another 7 years to prove it.
Published by
Sam Huxley at NMS Blog on February 24, 2009 in New Media Strategies, Online Intelligence, Metrics, Online Advertising, Online Marketing, Trends
4 COMMENTS SO FAR
I loved Michael Lewis’s piece, and his books as well, and it’s not tangential at all. The fact that it took over 50 seasons for teams to catch on points to how hard it can be to spot what matters. I agree about the misplaced focus on the publishers, and I think that is a function of the IAB’s role in the article more than anything else. The “Moneyball” approach is definitely the way to go, but it takes a ton of data and a very sharp eye to catch the patterns. We’ve been working on this for a while, and though we’re making progress, no one is close to having this figured out across the board. The Comscore study, “Whither The Click” is definitely worth the Google and read as well.
Sam,
excellent piece - thanks very much for sharing. I wanted to offer some data on how engagement ads linked to online video (both professional and user-generated) play into the arguments you make. First point is that video is just another form of editorial content. The engagement ad units GET Interactive builds for customers are derived straight from the customer’s video content, and so these ads can be thought of as opt-in engagement extensions of that editorial. The GET platform supports capturing the “beyond the last click” data you describe because when users view the GET engagement unit, it is easy to track in the same manner as display advertising – i.e., impressions, interactions, clicks, and their corresponding averages.
We’ve been at this for a little while and have accumulated some statistics that you and others here may find interesting. Our approach to video engagement advertising has yielded an average interaction rate of 67% and a click-thru yield of almost 20%. On average, GET ads generate 5.25 product panels viewed per interaction - orders of magnitude higher than banner or search ads alone. Because GET in essence combines online branding WITH the opportunity to click thru in one self-contained system, I think the high CTRs we’re seeing really supports your point.
Importantly, our preliminary data for editorial content with GET ad units that are ‘grabbed and shared’ across social websites shows metrics in-line with the above. This means that content owners and publishers can embrace social media distribution to reach more eyeballs without (fear of) losing the connection tosponsors.
Finally, the GET data, which is all available to our customers, can be used to see which ‘categories’ of products (i.e. apparel, jewelry, food & beverage, etc.) garner more interest and in what type of scene (based on subjective evaluation), and ultimately, with which type of user (when such data are available). I’m open to and interested in public/private comments to further discuss – greg dot wallace at getinteractive.tv
As part of a search engine marketing company, I am in total agreement with the careful consideration of how measurement influences the perception of value. That thought, of course, extends to “last click”, “first click” etc. conversion attribution. And it is also true that many searches are simply at the “end of the chain”, and are over-credited for conversion influence. “Up stream media” etc., including social media, can influence “down stream” behaviors including search. But I would add that search also influences other channels—offline, direct navigation, organic search (if talking about paid search), and (yes) social media. Influences are not “unidirectional”, but complex and interactive. And there will no “simple” method of A—> B—-> C. And, as Sam points out in this post, data tracking is not the only solution to the difficulty (and is limited by privacy considerations, etc.), and that careful measurement, testing, and analytics can help fill in the picture of mutual, interdependent influences. Excellent post. Best, Pat @ impaqt.com
This link below may seem tangential, but isn’t at all. It’s a piece about basketball that ran last week in the Times, and how a completely different approach to the game reveals extraordinary value in previously under-rated players and tactics.
The salient point is that the things that are easy to measure aren’t necessarily the ones that matter, and that focusing on the obvious instead of the essential leads to a serious competitive disadvantage when you’re up against people who have abandoned the old metrics, and are deliberately operating in the blind spots that those metrics create.
The AdAge article kind of gets it, but not really, in that is says media providers, as opposed to media buyers, need to be more savvy about how they sell. This is true, but glosses over the fact that they’re having to compensate for inability on the part of the media buyers.
After all, individual outlets have no control over the other platforms in an advertiser’s mix, and it is precisely this mix that increases or decreases the effectiveness of any one buy. The fact that the buyers seem to be unaware of this - looking only at ‘performance’ on a per-platform basis, and not using filters that quantify how different platforms perform in different contexts (that they create) - says they’re going to find themselves loosing to media agencies that figure this out, and start playing like Shane Battier and Daryl Morey.
http://www.nytimes.com/2009/02/15/magazine/15Battier-t.html?_r=1&ref=magazine&pagewanted=all
Alex Bowles commented on February 24, 2009