Heuristics
What a great word! Especially as it pertains to web analytics. The reason I love this word, especially as it relates to web analytics, is because of two of the smartest people in analytics, John Marshall and Dr. Stephen Turner. Yup, the creators of (what I will always argue was) the best ever web analytic tool, ClickTracks .
Back in the day, when we would train people on ClickTracks or web analytics, the opening ‘slide’ was ‘You don’t want to reconcile your online sales numbers (revenue/sales) in your web analytics to your sales numbers (database)’.
The point back then (and maybe somewhat now), is the numbers are directionally correct, but should not be compared 100% to your sales data. The reason – heuristics. Spoiler alert – about to get really geeky now.
So web analytics has always been based on data sources that were unstructured. What does that mean- Well, you can think about structured data as data that fits nicely into a spreadsheet. There are defined columns and values that match expected data formats (text, number, etc.), so calculations and queries are easy.
Delving even geekier – web server log files (pre-historic web analytics) were pretty structured. So to unbelievably oversimplify – you could open them up in Excel, sort by ID (cookie) then time-stamp and page and voila – web analytics. And that was the beginning of applying heuristics to data to get behavior reports.
More to come….
