And their numbers shall be legion

Marketers are “glass half full” people by instinct. A natural optimism about the world is what drives them, which is particularly useful when the world has not just drunk half the glass, but knocked the table over and spilled the rest.

Data professionals, especially analysts, tend to want to quantify the size of the glass and the volume of the liquid it contains before offering an opinion about whether to be postive or not. For them, turbulent market conditions are an ideal opportunity to amass data and test models since nothing used previously fits current conditions.

This difference in approach may explain why the work of the latter is still ignored by the majority of the former. As Forrester Research discovered, while marketers know they should be optimising their campaigns using customer insight, only one in four actually do so (see “Marketing by Numbers”,for more detail).

So is the problem a fundamental one where a numerate culture is in conflict with a visual one? Or could a bit of give and take by either side lead to a more data-driven marketing practice which delivers better outcomes for the business?

Caroline Johnson, data and insights director at Serendipity2, says it is easy for a bunker mentality to develop on either side. “In my old role at Experian, I was managing a team of 24 analysts and had to manage their time and deliver to the sales people. It was about trying to draw those two camps together and away from a ‘them and us’ situation,” she recalls.

Time pressures are one of the primary causes of this problem. The sales team selling-in analytics to marketing clients rarely has time to fully debrief the business context and background involved. What that risks is a surprise when the outcome of the analytical process is delivered because the analysts may not have been thinking the same way as the marketers about the issue.

Often that is exactly what marketing needs. “Pushing yourself out of the comfort zone can be very rewarding,” says Johnson. Her company has brought about a change at Coca-Cola as a result of giving its trade marketing teams desktop access to key planning data in a bespoke marketing system.

“It is a big company that you would think is clued up about data, but this was a massive project for them. Now they are using our software without having to come to us for insights into their data,” she says.

To launch Vitamin Water, the company was able to build its own thematic maps, looking at the underlying demographics of catchment areas and work out distribution plans for themselves. That is what data-driven marketing should be about since it accelerates business decision making and optimises results.

Marketers know that these tools exist and are also under significant pressure to improve the performance of their activity. Many of them work in companies that have in-house customer insight teams or use agencies that virtually force analytics on them. So why is there still a disconnect?

“Laziness” is the one word response of Ian Robinson, chief intelligence officer at insightTMW. His agency surveyed its client base for its use of analytics to identify what propositions and creative were working best. “We found a real lack of it. When we try to push something like email response analysis, clients say they don’t need to worry about it because, whereas direct mail used to cost them 20p an item, email is only 0.5p,” he says.

This plays to a “business as usual” mindset which sees the complexities of matrix testing, response analysis and analytics as too difficult compared to the volume-based marketing which a low cost channel has enabled.

For Robinson, this is doubly frustrating. “I would have thought at long last we have got a medium in which you can have really thorough tracking and measurement that doesn’t break the bank. But it has gone the other way and clients don’t want the effort and time that involves,” he says.

Trying to change this view can actually be dangerous for external service providers. Robinson notes that when responding to one RFI involving email marketing recently, TMW put together a sophisticated analytical approach intended to optimise the campaign. All the client was looking for in terms of analytics was bounceback and clickthrough rates in a simple report.

This touches on one of the dark secrets of the marketing discipline – its obsession with size. Having the largest number, whether in terms of budget or volume of items, is a common benchmark. Just look at the number of league tables produced for every aspect of marketing.

Mark Robinson, managing director of Qbase Direct, says: “I fundamentally believe it is a cultural issue. Everybody talks about being targeted and improving ROI. Very few people are doing it. Most marketing departments are targeted on the revenue they generate from each campaign.”

This metric sits in direct opposition to the analytically-optimised approach driven by data. Robinson demonstrates with a campaign performance chart created for a catalogue company. It showed that mailing just 36 per cent of the actual volume used would have doubled profitability. However, it would also have yielded 30 per cent less overall revenue, even though those unmailed segments offered a negative ROI.

This creates an instant block, even though the optimisation would involve cutting out loss-making customers. “No matter how much you tell them it costs £150,000 to get that extra £100,000 in revenue, there is a cultural barrier around how management pushes targets down to the marketing department,” he says.

In most large organisations, the business plan comes first, setting out targets for growth and absolute figures for the revenue expected. These are then translated into a marketing plan which calculates what level of budget and type of activity would deliver against those targets.

Only at the third stage, when the marketing plan has to be operationalised, does any data or analysis come into play. “There is often a disconnect with the data team – marketing is not as close to them as they should be. In a lot of organisations, data is part of IT. Organisations with hybrid teams where analysts work day-in, day-out alongside marketing are not the norm,” says Qbase Direct’s Robinson.

This lack of closeness is one of the main reasons for the failure to understand each other’s needs. Analysts may get frustrated that their insights and models are not being adopted by marketing, but the marketer may equally feel that analysts just do not understand the real world they are working in.

Add in the top-down pressure to hit targets for gross revenue, rather than profitability, and you have a volatile situation. “From the marketing point of view, analysts may not understand the process the marketer is working with,” notes Gareth Mitchell Jones, director of analytics at Experian Integrated Marketing.

“You will always get responses that fall outside of the model. If marketing doesn’t target them, it will miss out on those responses,” he says. This is the fear that haunts marketing and provides a constant stick with which to beat analytics. No model is perfect and there will always be actions that can not be predicted. Some of them even yield revenue for the business.

“Analytics is not an exact science,” acknowledges Jones. “That is why you have error rates and is why marketers come back and say, if we had used your model, we might have missed out tens of thousands of pounds.”

As well as the need to hit revenue targets, marketers also have a personal desire to hold on to their budget. Optimisation usually involves achieving the same outcome at less cost. But unused budget may be clawed back by the board in next year’s settlement. For this reason, Jones says, marketers go for bigger volume solutions. “They know they can mail 200,000 people using conventional print for the same cost as they can 10,000 using digital print, even if there is no increase in response,” he says.

In balance, it is fair to say that analysts have not always made themselves very user-friendly. Colette Wade, marketing director at Webtrends, notes that when she first joined the company it was generating web log analysis files. “I looked at them and thought, what would I do with that?” she recalls.

Only when the business decided to become marketing-centric, rather than IT-centric did the value of reports for marketers start to rise. “I could look at it and see visitor numbers, conversion rates, downloads, although it was still hard to look at,” says Wade.

With the release of Version 9, Webtrends has made its reporting far more visual and interactive, even though the same depth of data collection and mining is taking place underneath. “Marketers are frightened of data and there a still a lot of old-school marketers about who put their finger in the air and say, let’s see what happens,” she says.

If that view represents one end of the data-driven marketing spectrum, then the other end is an over-reliance on data without adding any extra value. “Is there a point at which you end up building a website purely from what analysts have found in the data?” asks Zia Zareem-Slade, head of experience planning at EMC Conchango.

She notes that creative teams will push back at a certain point, usually where a data-driven insight contradicts with their instinct for what will work. And sometimes they are right. “We engaged with a client who took a best practice checkout process and designed their own version, put it out there and found it worked better. They can’t work out why from the data, only that it does,” she says.

This is the space in which analysis and marketing need to meet. Data has limits to its ability to predict behaviour and explain what has happened. Marketing has limits to how much it can change. Both sides have numbers telling them what to do – when they match is where the magic happens.