Fragmentation, Optimization, Integration

Fragmentation, optimization, and integration. Those were the themes that stood out for me at last week’s eMetrics Marketing Optimization Summit in Washington, DC. First of all, I was reminded that “Web analytics” is not what it used to be; it’s an increasingly complex space. Secondly, I was shown some great evidence about what’s really required to be an optimization-orientated organization, and finally we’re beginning to see some case studies from organizations that have made the investment in multichannel data and the benefits they’re accruing from that.

A strong theme from the conference was the search to understand the value of social media. I got the impression that if you were a speaker on track, then you were in trouble if someone else had the words “social media” in the title of their presentation. Social media analytics is now definitely part of the landscape, but it’s hard to define what it really is. Everybody is coming at social media from a somewhat different angle – online brand reputation management, fostering innovation, or customer service – so it’s inevitable that there is a disparity of thinking about how to measure its impact. So it was useful to see people like John Lovett from Web Analytics Demystified talking about the framework that has been put out there with Altimeter. I think Lovett would agree that the framework itself is not rocket science, as it’s essentially about linking metrics back to objectives, but I think it’s exactly what’s needed in the space at the moment to take some of the “mystery” and “fluff” out of social media analytics. As one of my colleagues would say, “At the end of the day, it’s just data.”

Optimization has been another hot topic for a number of years now. Before, we used to be treated to presentations that espoused the benefits of using approaches like A/B and multivariate testing and the kinds of uplifts that could be achieved. What we saw this time from the likes of Dell and eBay was what’s actually required from an organizational and technological perspective to implement an optimization culture in a business. Ed Wu and the team from Dell demonstrated the size and scale of the investment that’s required. They doubled the size of the testing team, bringing on an additional 10 hires in a two-month period. They described the challenge in finding enough of the right kind of talent. In addition, they established a project management office to help facilitate the execution of tests and introduced project management tools to manage the whole process. The Dell story just underlined that analytical success is as much, if not more, about people and processes than it is about technology.

That’s not to say that technology isn’t important and Bob Page from eBay outlined the type of investment that they’ve been making in getting their data architecture right to support the various analyst communities in the business. Initiatives included getting rid of the various datamarts that had sprung up around the company and creating centralized virtual datamarts that were able to flex and meet the shifting demands of different analytical groups across the course of the day. Technology is also used to connect different analytical groups together. A networking site joins together the various analytic communities together where initiatives can be shared, knowledge can be captured, and best practices can be developed.

Finally, we began to see some of the benefits that accrue from data integration strategies. Adam Greco from outlined how Web analytics data and customer relationship management (CRM) data can be used together to improve the effectiveness of the lead generation process. Lead data given to sales people can be enhanced with not just what’s on the contact form but also with the prospects’ previous browsing behavior on the site, highlighting the products and services that the prospect might be interested in and scoring them in terms of their levels of interest and stage in the sales cycle.

This type of data integration requires planning and for data from different sources to be matched together. Allen Crane from USAA showed how they had been doing this integration across multiple channels. Over a two-year period, USAA has been bringing data from the different channel into its centralized warehouse and analytics tool. It created a data schema that focused around the customer and built out around that. In what I think was probably one of the best case studies around multichannel data integration I have seen in a conference, Crane showed how they had developed the notion of a “conversation” to encapsulate the interactions between USAA and a member across multiple channels around a single event like taking out a personal loan. That conversation might start online but end up offline, or indeed end up online again. USAA’s data integration allows it to track these conversations and to understand the true costs associated against these different events. It allowed the company to develop new metrics like “containment rate” (the amount of the conversations that happen online) that enables it to prioritize its site and channel optimization efforts. For me, this is where Web analytics is heading. It’s about taking that data, adding the multichannel perspective, and then deploying the right people and processes to change the way the business does things.

Related reading

Big Data & Travel
Flat design modern vector illustration concept of website analytics search information.