Analysing online behaviour in real time, the PositiveFeedback managed service will serve up highly targeted recommendations and advertisements to customers as they browse retailers' Web sites. Until now such technology has been the sole preserve of large companies with big development budgets.

The KXEN solution will be embedded in PositiveFeedback’s online personalisation service – due to launch soon – which will uniquely be sold pay-as-you-go, making the techniques available to a greater number of Web retailers.

Paul Cook, founder of PositiveFeedback said: "To make recommendations in real time takes two things: accurate data and fast analytical models. Collecting and cleaning profile data in real-time is hard enough but creating hundreds of analytical models and refreshing them maybe once an hour takes something special. We looked at the market, talked to our peers and it was clear that KXEN was the best by a long way."

CEO of KXEN Roger Haddad said the tie up took both companies into an exciting, high-growth area of technology: "Online retailers are constantly striving for a new edge. The PositiveFeedback service will not only enrich customers' experience, it will generate new sales opportunities for Web vendors too. We're very pleased to be a part of it."

Dubbed extreme data mining for its speed and ability to handle large numbers of variables, KXEN's technology has also earned a reputation for being easy to use. Conventional analytics can take a team of expert analysts weeks to create a single model coping with only a few tens of variables. With KXEN one non-specialist can build a similar model in a few hours, apply it to thousands of variables, and achieve results of the same or better accuracy.

Paul Cook adds: “These strengths made KXEN ideal for the PositiveFeedback application. Customer interaction data will go, in real time, to PositiveFeedback's servers for analysis. Custom recommendations, advertisements and other information will immediately flow back to customers' screens.

"We'll be offering retailers the kind of cross-sell/up-sell techniques that have so far only been available to big organisations like Amazon. The market has been waiting for this technology to be available and affordable across the board and we've had a great deal of interest as a result. The recommendations solution will be particularly good at driving sales of lower ticket items such as books, tickets or CDs while the ad. serving product will be particularly attractive to sectors with higher pay-per-click costs like financial services. For their customers the interaction and offer presentation will be entirely transparent. They will just get more relevant offers and more useful recommendations."