Lies, Damned Lies & Media Metrics

In the coming years, as clients focus on effectiveness more than ever, marketing metrics will be revolutionised.  Taking traditional media metrics online hasn't been a great success.  The only information it seems to provide about consumers is how uninterested they are in banner ads.  But the moment we step outside of the 'impressions' box, the scope for innovation becomes enormous.  Just look at Google PageRank.  The simple idea of measuring influence by noting who is pointing at what content was enough to fire up a whole industry.  Then by adding another simple media metric innovation in the shape of AdWords (ie only pay us when it works) Google solved the '50 per cent' question and created an irresistible commercial offer.  Which also had the bonus effect of making alternatives such as BARB panels and ABC figures look prehistoric.  It could be argued that these two metric innovations alone, and how they have moved the mindset of the market, is where Google's dominance lies.  So what next?  Well, at a Microsoft mega media blitz I attended in Paris last year, the company started talking about 'last click' being a poor way to measure anything because it doesn't offer any insights into what drove people to that point.  Was it a TV ad or a quick dip into a product forum that led to them searching and sealing the deal online?  In response MS was offering a vast joined-up platform - MDAS - which would effectively follow the consumer in their commercial journey and map what they paid attention to.  Over the last year, Microsoft has established the Atlas Instititute to promote this thinking, which now, mercifully, goes by the ad-friendly moniker of Engagement Mapping.  In part it's a strategic stick with which to beat Google.  However, there is a common sense element to it that advertisers may find appealing.  As all media is dragged onto a single platform (aka the internet) in some way shape or form and people's journeys through different media become visible, the ability to track these routes to product could be very valuable to global marketing clients.  But why stop there?  The splendid Purple Motes writes of the possibilities for marketing when its metrics go beyond media to embrace other aspects of the 4Ps: "Considering other factors relevant to user response highlights the meaninglessness of managerial allocation of credit for ad clicks.  An attractive, well designed product helps to induce users to click on ads.  Should the contribution of product design be ignored in allocating credit for last-ad clicks?  Should the contribution of product reliability be ignored in allocating credit for last-ad clicks?   What about product feature set?  Why should Engagement Mapping just allocate credit to advertising expenditure?"  Which I think points at extending the social side of marketing metrics.  He goes onto say that, "In practice, allocating credit is mainly valuable as a tool for personal and bureaucratic justification," which I think is economist-speak for Creative Awards!  The big challenge here is, as ever, execution.  Of course, technology will get there in the end, but whether Big Media is able to crawl out of its siloes and start sharing their precious data is another matter entirely.

Comments

[Charles Frith]

A thought provoking post covering a lot of ground. Alas using a mobile to comment is restrictive but briefly, I find too many unanswered questions in the digital metrics business. For example I'm interested to know more about depth of immersion/content engagement contexts before purchase.

Maybe its poorly thought through but an ad free internet that only serves up a selling message when sufficient criteria have been met strikes me as a model that respects rather than visually spams till hitting jackpot.

I was about to go on but a proper keyboard is a kinder tool to ramble a bit.

[Charles Frith]

A thought provoking post covering a lot of ground. Alas using a mobile to comment is restrictive but briefly, I find too many unanswered questions in the digital metrics business. For example I'm interested to know more about depth of immersion/content engagement contexts before purchase.

Maybe its poorly thought through but an ad free internet that only serves up a selling message when sufficient criteria have been met strikes me as a model that respects rather than visually spams till hitting jackpot.

I was about to go on but a proper keyboard is a kinder tool to ramble a bit.

[James Cherkoff]

Thanks for making the mobile effort Charles! ISTM that the web's promise of 'total data' took along time coming but then arrived like a tsunami and blew everything away. Now we are looking through the tatters and trying to work out what goes with what. That stage is going to take a while and only the organisations with mega-platforms will have sufficient juice. Step forward MS, Google and indeed (still) Yahoo. An ad-free network is a very interesting concept...

[Simon]

"That stage is going to take a while and only the organisations with mega-platforms will have sufficient juice"

I definitely agree. A back of a fag packet calculation working out the amount of data produced via behavioural tracking astounded me! Not only will the winner have a large network, but they will also need a large cloud/farm

[James Cherkoff]

Hey Simon, I'd be interested in seeing that fag packet...

[Simon]

If you make
- an assumption on the number of interactions per impression (ads witnessed/information read/clickthroughs etc)
- and combined that with the metadata involved on each site (depending on what you wanted to measure - time of day, category etc)
- for a top site (say Comscore top 500)

You would be looking at gigabytes per day. If you took a single 4 week ad campaign, you would still need to hold every single piece of data in order to pull out the campaign. That is a lot of data to store and parse!

[James Cherkoff]

Thanks Simon, that's really interesting and I suppose points to some new mega-architecture for ad serving - probably considerably more powerful than what we have now. Thus the Google + DoubleClick and MS + Atlas tie-ups I suppose.