Many companies now offer a service, often combined with a product, which records actual user sessions accessing your e-commerce system. Real user experience is obtained, so you see the service as they do. The service typically takes the form of the supplier providing reports based on, for example, carrying out the same set of tasks on your web site from a number of different locations world-wide, at the same time each week.

Although targeted at the performance manager, this produces much useful business information, but often not geared towards performance management. It will quickly highlight what gets used on a web site, what causes a user to leave a web site and other such marketing and web site management data. It produces information that can be used for performance reporting from a user perspective, and the products used usually have a high quality interface to the web logs. Often such services now extend to comparison with your competitors' service, so that you receive competitive benchmarking data. If trended such data provides a profile of the performance level of a certain task over time, so it might assist in identifying a long-term trend towards a service problem.

The negatives are easy to see. The performance manager gets no help in identifying why performance is bad, simply evidence to show if it is. The information provided is out of date by the time it is received, thus not suited to performance investigation and correction of problems. The data provided fails the 'so what?' test for the performance manager. Knowledge that performance was bad last Tuesday morning is nice, but the performance manager needs that vital understanding of why it is bad, and what is the cause. This will only come from specific detailed performance software.