Many of the tools we’ve seen start with an aggregate view of what’s happening on a site. This makes sense, because with so much happening on a website, you should begin with a view of traffic as a whole.

When viewing aggregate data, however, you want to drill down quickly to individual visits. If you’re seeing a pattern of clicks on a page, you may want to replay mouse movements for that page to better understand what’s happening. If you’re seeing errors, you will want to replay visits to see what visitors did just before arriving at that page. When you notice a performance problem, you will want to analyze individual visits that experienced the problem to see what caused it.

You will also want to drill up. Given an individual visit, you’d like to know what was happening to the rest of your visitors. When someone calls the helpdesk with a web problem, the first question should be, “Are others affected?”

Drilling up is part of the diagnostic process. If you have a visitor who’s experiencing slow page loads, you want to analyze all other visitors who experienced similar conditions to see what they have in common. For example, if 10 percent of all users are having very slow page load times, analyzing all of their visits together can show what they have in common. You might notice that they’re all in the same city, or all have the same browser type, or all bought more than $500 in products.

Drilling up means creating a custom segment based on attributes of an individual visit (for example, “All visits from Boston on Sprint using Firefox”). It may also mean moving across several data sources in the process.