Why is this interesting? One of the site developers at one of my customers approached me and said that “Hey, you know we have a couple of internal users who complain about our website performance. They say its always slow and takes forever for pages to load….”. My response was “Ok, what data do you have to prove this is correct”.. the response was “None”

We took the IIS log files and run the following script

Select
    Top 20 cs-username AS UserName,
    AVG(time-taken) AS AvgTime, 
    Count(*) AS Hits
    INTO AvgTimePerUser.txt
FROM
    logs\iis\ex*.log
WHERE
    cs-username IS NOT NULL
GROUP BY
    cs-username
ORDER BY
    AvgTime
DESC

The script produced a list of users who had the longest average time for pages to load during the period of the IIS log file.

None of the users who complained about performance was in this list of top 20 people who had the longest average time. We run another script to check specific for the one user who complained the most and found out that he had very short average time. A second or so.

Select 
    cs-username AS UserName,
    AVG(time-taken) AS AvgTime, 
    Count(*) AS Hits
    INTO AvgTimeOnSpecificUser.txt
FROM
    logs\iis\ex*.log
WHERE
    cs-username = ‘CONTOSO\User1234’
GROUP BY
    cs-username

These two scripts gave the site developer enough data to have a dialog about site performance with the end user in a more productive way.

//Anders