Emetrics has come to a close after a few hectic days in DC, interspersed with seeing the Dalai Lama in George Town, the solar powered homes exhibition on the mall, the White House, hours chatting in the omni shoreham lobby bar and swimming in the invigorating heated outdoor pool at sunrise.

That asides, what has been going on? Or as my American counterparts would say, what are some of the key takeaways in terms of consumer understanding and behaviour. I’ll do another post about Google Analytics and Microsoft’s Gatineau this weekend.

Jim Novo of Drilling Down fame, spoke about speaking the “exec level” language that CEO/CFOs understand. If we think about our sales pipeline, it is the predictive/future likelihood to happen that execs are interested in when it comes to understanding our online data, sales and consumer behaviour so you can focus your efforts, marketing spend and optimisation efforts where they will have the most impact. Which are your dreck customers, your former best customer, new customers and best customers – map them out on a two dimensional value map with an XY slope.

Use recency, frequency and latency (you can even begin looking at these with Google Analytics) to understand your best and worst customers and grow your best customers. And importantly, build your predictive customer performance pipeline with your CEO/CFO so that they understand it, help you build it – which helps significantly with buy-in. Buy-in let’s face it, can be the biggest obstacle to taking action in any company.

Joseph Carrabis, the web analytics association new anthropologist and neuro-behaviourist on the scene, spoke about really taking advantage of our hard wiring to make our audience do and think what we want them to think or do. As human beings we all apply our own stereotypical and prejudiced frame of reference to everything into which we come into contact. For example in the context of images on a webpage, which image and at what position and angle will trigger what emotions or thoughts at a subconscious level. If an image is positioned at an angle, it implies motion. A photo of a couple, an elderly man, a teenager and early thirties woman will also, all provoke different inferences from one’s audience. To illustrate this, Carrabis engaged 50 of us in a persona exercise where we had to sit down after he namecalled six photos to tell him which one we thought was the Economics professor in Beijing. Interestingly, most of us thought the middle-aged conservative looking white man, was that character – and we were right. The key thing being the inferences that we draw.

In terms of multi-variate testing, the weather channel, used a variety of different images, a couple, then a man and also a woman to see which image was working more successfully in terms of optimising the page for it’s audience and hence having the highest conversion rate. Interestingly, the web page version withe the image of a woman on her own had a much higher conversion rate than other versions tested. This can be linked back to Carrabis point about the power of assocations, inferences and our pysche hard wiring on what we think about images, positioning and sound on a webpage.

Neil Mason, a fellow Londoner, talked about segmenting one consumer segments into tribes (richly developed personas in other words), using datamining to provide statistically robust anomalies, patterns, associations that stand out from a business commercial perspective and use these to identify key drivers for purchase and identify the most valuable consumer segment. For example, with a case study on the Royal Mail, segments included price finders (10%), cottage industrialists (2%) and regular posters (1%) – which were the most valuable segment. They also indentified that visitors who “saved a quote” on their first visit were signifantly more likely to become and continue to buy from the website and be the website’ most commercially valuable segment (worth most money). They used these consumer segments to drive email marketing segmentation and discovered that emails sent 4 to 5 days after their last visit were most likely to convert. Less than 4 days was too soon (the visitor was still thinking about it) and more than 5 days and the conversion rate began to drop. It’s all about the timing – oh – it’s recency again.

 ”Where are most valuable customers, how should I optimise my site?”

Thanks for an interesting emetrics everyone and I look forward to meeting those I met again soon