Bidding for Business With Customer Data
A version of this article first appeared in eB-21, published 10 times a year in Europe by TBC Research .Based in London and San Francisco, TBC Research helps senior business professionals make more informed technology decisions through its magazine, research, and events portfolio.
Maintaining a customer database is no mean feat for any company but when the number of customers involved grows from a standing start to over 2.6 million in around three years, bespoke systems struggle under the weight. Keeping information on spreadsheets, with reports from different transaction systems, is no longer feasible when the quantity of data being manipulated is so huge.
That's why QXL Ricardo, one of the biggest auction sites in Europe with a presence in 13 countries, is moving at breakneck speed to improve its customer analytics. It began with an initial six-week deployment of business intelligence tools to consolidate reporting last year. The icing on the cake is expected to be the recent implementation of Business Objects' Set Analyzer, which will provide it with the ability to analyse, segment and differentiate customers.
In the meantime, QXL has developed its own data warehouse which was completed at the beginning of 2001. This now holds a record of nearly all the transactional activity of QXL's customers on its various Web sites and site traffic data is currently being integrated as well. The next step will be to use that data to influence buying behaviour. But even for QXL, that scenario is some way down the line.
One of QXL's early gambits was to auction off unwanted Christmas gifts; today, business is a far more serious proposition. "We wanted a central view of our customers so we could pull together the different strands of information to build up a view of their transaction behaviour on the site," says Neil Mason, chief marketing officer. "We need to look at what they have sold, what they have bid for and how active they are, to work out our customer churn."
QXL developed segments to identify the proportion of customers at each stage of its value chain. It has so far created about four or five core segments, ranging from customers who only visit the site once through to the all-important repeat transacting members. Every member who logs on is flagged according to one of these segments and QXL expects to gain a better understanding of customers by learning what they do on the site. Eventually it will extend the segments into more finely tuned specific groups of 8 or 12 to give it more focus on customers' specific preferences. Concept testing is already underway and pilots have taken place.
Mason says: "Set Analyzer is up and running and being used within the business but the real issue is to integrate it into our operations. We need to have it as part of a closed loop environment so it becomes a natural way in which we do business. We need to pull it together with some of our other technologies."
QXL is also honing in on conversion ratios. It is only at the stage of monitoring activity but the past six months have seen the key performance ratios rolled out to its team of managers across Europe with each manager having the opportunity to monitor a range of different metrics daily. Every night QXL takes an update of the previous day's activities and puts it into the data warehouse. From this QXL can learn what percentage of members are converting into active members and what proportion is active at any one point in time. Over a period of time it expects to see the ratios improve and will measure and re-measure on a regular basis.
"We believe we will be able to get a much tighter focus on what really works in our business" says Mason.