Media Research 2009: a complex consensus

James Aitchison
WARC Online

On a day when the papers were all reporting the IMF’s prediction of a paltry 0.5% annual growth in the world economy, it was fitting to open Media Research 2009 with a keynote address from a major global advertiser.  

What advertisers need from media research …

The company in question was Anheuser-Busch InBev, whose Director of Commercial Spend Effectiveness & Media, Johan Landmark, set out three key themes impacting upon the brewing giant’s advertising and media investment decisions.

Two of these took us down a well-trodden path: increasing pressure to prove the ROI of marketing spend in an era of fundamental and complex change to media and its consumption. But it was the third theme that was most revealing: data analysis.

Landmark explained that A-B InBev used three different econometric models of increasing focus “to connect cost and effect”:

  • Industry model: What drives market growth/decline? By how much will the total beer market grow?

  • Brand Allocation model: What drives brand market share? How much budget do we allocate to each country and brand?

  • Lever Allocation model: How much budget do we allocate to commercial levers?

It’s the latter Lever Allocation model which houses the specific media channels, as well as variables such as price and distribution, which together enable ROI optimisation, scenario building, forecasting and tracking.

It was interesting - or, for one later questioner, “refreshing” - to hear Landmark identify price and distribution as the levers with the most immediate impact on sales and share, and openly adopt the longstanding agency-world mantra that advertising was about long-term investment in brand equity.

Still, despite the econometric data to support his position, I suspect it’s a view that needs continual defence in A-B InBev's corridors of power. “At times like these,” he said, “it’s a real challenge to work out how what we spend now will translate over the next 2 or 3 years.”

…and what the world is doing to get there

That’s the story of how one advertiser is making its media decisions. But to what extent is the agency world helping the corporate giants in their quest for demonstrable ROI and accountability?

Cue Robert Dreblow from the World Federation of Advertisers who introduced the trade body’s Blueprint initiative. Based on research among WFA’s worldwide membership (which accounts for some 90% of global adspend), the project investigated what advertisers really want from media audience research:

  • Quantitative, passive measurement systems
  • Multi-media audience research, including digital and in-store
  • Large sample sizes
  • Better target group descriptions
  • Flexible and adaptive approaches
  • Single-source, with a link between multi-media use and purchase behaviour 

Since its publication, the WFA has been calling for the industry-wide development of consumer-centric approaches to media audience measurement, which Dreblow articulated in a recent WARC Online-exclusive paper: Everyone gains from evolving media measurement.

Despite the termination of the flagship Project Apollo in the US, he said that there are some 17 initiatives around the world compatible with the Blueprint agenda - the leading light being the IPA’s Touchpoints programme in the UK.

Measuring mobile as part of an integrated strategy

For Oliver Roxburgh, MD of Bluestar Mobile, the key strength of the handset medium is its ability to “extend the reach of brands and content”. It was a glowing presentation on the merits of mobile, including its ability to build loyalty through SMS and reach and appeal to niche young audiences.

But, aware of his audience, he also offered up some detail on some of the data available for when it comes to proving effectiveness:

Moreover, he gave an example of how integrating mobile and internet-use data enabled a national newspaper to optimise traffic to its mobile website. Research identified the comparative peaks and troughs in visits to its mobile and main websites through a typical day.

By identifying the morning and evening commuting periods as times of peak mobile use, the newspaper was able to put in place promotional activities to increase traffic to its mobile site during these key points of the day:

From here, Eden State’s Anna Sampson picked up the baton to provide a more methodological overview of how mobile can be integrated into the communications mix. Her main point was that a hitherto divide in the media research community was now outmoded.

It wasn’t helpful, she said, to see traditional media as strategic, long-term and brand-building and newer media as tactical, short-term and sales-generating. She pointed out that campaigns are now employing mobile for branding and TV for sales tactics, so it’s an essentially false divide. The ultimate aim, she said, was to avoid silos and “integrate sales data and brand perceptions to deliver an integrated insight”.

Digital: The ABC of behavioural targeting

In the beginning, there was the byte. Then came the kilobyte, followed shortly by the megabyte. Today, we’re all comfortable with the gigabyte and, perhaps, some of us are even familiar with the concept of the terabyte. But the petabyte?

Well, according to Andrew Newman, Yahoo! Europe’s Director of Strategic Data Solutions, Yahoo!'s web usage data constitutes four of them.

Analysis of this info-mountain enables the online giant to go beyond the traditional norms of online ad targeting which, in increasing levels of detail, can be triggered by:

Run of network
Time of day

Yahoo!’s focus is now below the dotted line, in the realms of behavioural and bespoke targeting, generated by the creation of user-behaviour models from a regression analysis of page views, ads clicked, search queries and search clicks.

These models drive the ads that people get served, which can also be tempered by sector-specific purchase cycle information - for example, florist ads won’t last longer than 24 hours.

But Yahoo!’s analysis goes further: “We are building profit and loss data at the level of an individual cookie. We know if you are making or losing money for Yahoo!.” Such information helps prioritise services, such as call-centre support, for which profitable users can be put to the front of the queue.

Word of mouth, reach and online ad effectiveness

Next up in the digital session came Matt Dodd from Nielsen Online, who introduced Net View: Nielsen’s metered panel of 36,000 individuals, which is used to track all online activities and, ultimately, enable Nielsen to “visualise every click”.

Focusing on data for a client, Kirker holidays, he showed how one potential customer’s journey started with the googling of a Florence-based hotel, a visit to the Kirker site, then on to, then Google maps and finally easyJet.

This minor example confirms what we’ve all long known: that the AIDA (attention, interest, desire, action) model of advertising is a busted flush - and, what’s more, there's no suitable model to replace it.

“No customer journey is the same,” said Dodd, who said that even for the same product such as a holiday, initial interest to conversion could vary from one to eight weeks. His advice was to take the conversion as a starting point and work backwards to see the links between variables such as search, website aggregators, consumer-generated content and your advertising.

Neurological testing

Advances in neuroscience over the last decade or so have stimulated its appliance to a number of industries seeking to understand the mystical secrets of the brain. There was a neuro-economist speaking at the World Economic Forum in Davos this week, presumably to help explain what mainstream economists currently can't.

And, likewise, there was a neuro-marketer at Media Research 2009 preaching a persuasively simple message: that the brain can tell marketers all that they need to know.

"Neuroscience," proclaimed Dr A.K. Pradeep, CEO of California-based Neurofocus, "unlocks the universal language of the brain. It avoids catching the bits people didn't mean to say but did, and does catch the bits they meant to say but didn't."

He uses the humble baseball cap to collect data, albeit a baseball cap containing 64 sensors that, 2000 times per second, measure attention, emotional engagement and memory retention. Combinations of these three variables determine three effectiveness metrics:

  • Persuasion: emotion and memory
  • Awareness: emotion and attention
  • Novelty: attention and memory

This methodology, said Pradeep, can be applied to a wide range of marketing situations. In combination with eye-tracking technology, the readings from the baseball cap can track brain responses to commercials, identifying the strong and weak elements of an ad, and help optimise creative content.

For TV ads, he continued, it can predict wear-in and wear-out levels, which can feed into media decisions (high wear-in warrants boosted frequency; high wear-out warrants boosted reach). Other applications include retail environments (for example, walking subjects through a store), product testing (monitoring the consumption process); and packaging testing.

Pradeep's enthusiasm for neuro-scientific methods was infectious and few would disagree with his observations on the limitations of traditional advertising and media research. In that context, it can make a welcome addition to the toolbox. But the quasi-lab-rat nature of neuro-scientific research is hardly capturing the consumer in a typical, day-to-day environment, which is why it can't - at least yet - be a replacement.

IPA Touchpoints

An initiative that's perhaps better at delivering real-life media data is the IPA Touchpoints database, already lauded earlier in the day as the current gold-standard of integrated media planning tools.

It was Kate Cox, MPG's strategy director, who provided a guide to what Touchpoints had to offer - not as a stooge, but as a generally impressed fan, who summed it up as a "great decision-support tool".

Key findings from the most recent release of data in 2008 (based on self-completion questionnaires and PDA diaries), found:

  • TV is still dominant among 15-24s
  • Web and TV are growing, whilst radio and magazines declining
  • A rise of multi-tasking

Other topline trends are detailed in the 2008 WARC Online exclusive paper: Results from IPA TouchPoints 2.

Transactional impacts

Despite all the resources devoted to understanding media choice, "it still remains a bit of a gamble". This was the view presented by Lawrence Janes, Head of Media Solutions & Development at dunnhumby - the company responsible for analysis of all Tesco's transactional data, which encompasses the shopping activities of some 14m people in the UK.

Such analysis, he says, enables the production of hard data of marketing's impact on shopping behaviour across a surprisingly elongated fmcg tail, in which only 10.5% of brands have more than 5% penetration, and nearly 80% have a penetration of less than 2%.

He illustrated how:

  • Exploratory testing and econometric analysis could identify the effect of specific media campaigns on sales - even down to the level of a specific print title.

  • A confectionery client discovered that half of its £11m budget was not working and (cue Holy Grail) which half it was.

  • A dairy client was able to conclude that 70% of its sales were base, and 30% from marketing. Within that, half its marketing sales were generated by an over-relaince on margin and equity-eroding BOGOF promotions, which promped more investment in media.

Media Research in 2009: a complex consensus

"What pleases me about today is that at conferences like this a few years ago, we'd be trying to build a supercar. Now we are accepting the reality in which we have to work with lots of different data sources." These comments, from afternoon panellist and Initiative planning director, Tony Regan, sum up well an emerging consensus - that the quest for the silver-bullet, single-metric solution of yester-year is over.