R.E.M. Reshaping E-Metrics (or Revolutionizing E-Marketing)

Chimere1
Marshall Sponder’s blog WebMetricsGuru had a post with comments from a recent Eric Peterson’s e-consultancy interview that caught my attention. The conversation is about the present difficulties for analytic vendors and display advertisers in the face of rich-internet applications (RIAs) and feeds as they continue to proliferate on the web. Sorry if I don’t shed a tear.

Murmur
The impression has been the metric du jour since the inception of the commercial web. Advertising and reporting models, products and services have all been built around delivering and measuring impressions. The problem is that impressions (and its twin brother the page view) do not measure interests, intents or actions. To me, that’s a pretty worthless metric. Also since they occur so infrequently, click through and view through rates from display are also pretty worthless metrics. Making maters worse, display ad models are better served by (and sometimes created for) poor user experience. Thankfully for most of us these models and metrics are becoming remnants of a bye-gone era.

As long as we’re throwing out page views and killing CPM let’s make sure there’s room in the trash bag for panels as a form of measurement. This relic is from the days of that analog device known as television. There have been some incredibly valid points made the past week on this. The IAB who voiced these concerns is also at a crossroads and bears some responsibility for lack of foresight to comfort the industry (disclosure: I am a former member of the IAB Research Committee). Action in the face of changing technology is not one of the IAB’s strong suits. Committees are very large and consensus and action take far too long in a medium as dynamic as digital.

Reckoning
In stark contrast stands the cost/action/result advertising model. Once action or interest is expressed a series of events are initiated to deliver relevance matching the user’s goal with a fulfillment opportunity. As digital mediums pervade our lives our ability to track, measure and improve the context of relevance increase. The brainpower, money and technology fueling the development of this by Google, Yahoo and Microsoft is like a modern Manhattan Project…and they will not be waiting for marketers or agencies to figure out how to deliver, track and measure any of this. They’ll be building this into their ever-expanding solutions and inventory.

The digital advertising and the metrics industries may need to adapt to these technologies but there’s no reason marketers need to wait around for that to happen in order to be successful. The success and failure of RIA use can be easily ascertained with A/B testing. Multiple design patterns can be tested against each other to see which one moves users closer to their goal. Our success or failure with these strategies is easy to determine and we are and should be held accountable for the results. This level of accountability requires us to understand data and distinguishes us from traditional marketing and older mediums. What hasn’t changed, but is amplified is that good user experiences are the key to successful long-term brand building and immediate conversion and the place testing has in creating effective marketing strategies.

Fables of the Reconstruction
Still, the display world has been trying to convince advertisers that everything is going to be better with Behavioral Targeting. They claim analytic and data driven delivery is going to make display, well relevant. Yes, BT is better than just blasting out media however it is flawed because it is targeting based on past actions and previous intention assumptions. It’s nowhere nearly as effective for ROI as targeting to a current or engaged action as is done in SEM. Nor does it address the largest problem facing display, drawing the context of user focus with great creative and messaging.

Ironically one of the few places I have seen display provide a measured online success is when used in a marketing mix that includes search. Using search impression data we can measure the interest that display advertising creates. At SES New York this year I presented research collected Sony’s launch of a new a digital camera showing how a branded display campaign increased brand impressions in search. Others like Cam Balzar at Performics have presented similar findings. Clearly display can be used to increase branded keyword inventory. This can be a huge benefit for advertisers as these low cost/high click keywords are usually the most important in the portfolio.

However, the search world and the display world have and continue to reside in different universes. Why? One requires understanding of how to spend money. The other requires understanding of how to make money.

Life’s Rich Pageant
We are only scratching the surface of what RIA and metrics driven marketing will do and mean to analytic vendors and ad buyers. There will be a sea change. But businesses and marketers need not worry. As with the rest of the next-generation changes to the web progress is not only inevitable, progress is good and brings with it opportunity. If there is one thing that this new paradigm is about it is about doing more with less. Those are metrics everyone should be happy with.

*The Emetrics Summit is October 16-18 in Washington DC

3 thoughts on “R.E.M. Reshaping E-Metrics (or Revolutionizing E-Marketing)

  1. search world and display advertising world reside in different universes

    Jonathan Mendez, in OptimieandProphesize, commented on an earlier post of mine highlighting problems measuring traffic for Web 2.0. Also, Jonathan mentioned something about how online display advertising differs from search advertising. ……

    Like

  2. Jonathan: It is quite fair to take a whack at the traditional metrics, impressions & page views etc, I have engaged in that sport myself! 🙂
    [Challenges for Web Analytics Vendors:
    http://www.kaushik.net/avinash/2006/09/five-ecosystem-challenges-for-web-analytics-vendors.html%5D
    But I am not sure of what alternatives you are suggesting we measure for the great new world that some people are already in and most will be in soon.
    Yes we can do testing with RIA’s and alternatives (remind me to talk to you about how we have done a few of those already and the delightful lessons we learned). But how should we measure success? How about advertisers? Maybe the suggestion in this post is to measure “cost per outcome” (CPA).
    Or maybe source success metrics from the two other elements of my Trinity model. Something from “Outcomes” or something from “Experience” (satisfaction, task completion, likelihood to recommend etc).
    [Trinity mindset overview:
    http://www.kaushik.net/avinash/2006/08/trinity-a-mindset-strategic-approach.html%5D
    Thanks for helping exercise brain cells early in the morning! 🙂
    -Avinash.

    Like

  3. Metrics exist to facilitate buying and selling, negotiation and resource management. It’s not an intellectual exercise.
    Whether it’s IBM or P&G, my experience is that marketers would gladly pay a publisher a fixed margin on every unit sold. Avinish, I’m sure you all would too!
    And, it’s been done before – DRTV and DR Print are proof. Cost per click was an attempt to follow the model… But paying for performance is not a good model for publishers that have lots of inventory to sell, or huge revenue hurdles to reach – and we all know how tough it is to generate and keep an audience.
    Publishers, be they web or display or broadcast, make money from advertisers that gamble on that publishers delivery of an audience. “Wasted Ad Dollars” is how publishers clear the inventory and make money. And, marketers have learned to price risk into their models, which is why 1% response rates can keep our mailboxes full of trash.
    Without change in the publishing model, marketers are stuck gambling, and buying data and tools which help them hedge their bets. Outsell estimates that over $25 Billion will be spent this year on marketing info from Nielsen, TNS and the like.
    So long as publishers are unwilling to provide transparent data or even agree to standardized metrics – you’re stuck with inferential statistics, mediocre techniques based on biased sample data, custom panels and lots of media provided market research that demonstrates (coincidentally) how that unique publisher property reaches your target best.
    Meanwhile the big data firms are more than happy to keep feeding IBM and P&G something that might just give them an edge…with $300 billion on the line, can you blame them?
    Can we challenge metrics providers to innovate? Absolutely. Should we bag on inaccurate data providers? Yes. Can we complain about marketers being shallow? If you want.
    But let’s not forget the source of these information problems: without transparent publisher or transaction data, these other folks are just doing what they can to be smarter. And, with all that in mind, I believe CPM remains a useful metric.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s