“Path Analysis” is a type of analytics report that monitors the page views leading up to and following the view of a specified page. For instance, a path analysis of the main “Products” page on a website might include the three pages leading to it and the 5 pages following it. The report indicates the most frequent paths.
Most of the time, I find this to be quite useless, especially when looking at more than 3 pages in a path. Why? The internet is a web of pages, not a pathway of pages. So path analysis is really just a list of the many different ways a visitor can arrive at a specified page. Only if you have a gazillion visitors to your site will this information eventually (if ever) become meaningful. Sure, there are some experiences in a website that are linear in nature – filling out a serial form, registration, checkout, etc. – but all of those experiences are carefully monitored by conversion funnels (or should be).
However, there are two VERY good situations in which path analysis is not only useful but necessary.
Situation #1: Anomalous Page Views
When viewing your page view reports, you see that a specific page is getting an unusually high number of visitors. This is unexpected to you and you want to know one or both of the following:
- What brought these visitors to the page?
- Where did the information on this page lead them?
Let’s suppose that the popular page in question is a help page defining a technical term and the number one page leading to it is a set of instructions for performing some worthwhile task. In this case, you have probably used an overly-technical term in the instructions without context. A slight amendment to the text will provide a much greater user experience.
Now, let’s suppose that the popular page in question is a blog post about your amazing collection of Incredible Hulk figurines (it turns out that the big green guy is all the rage – no pun intended – these days and everyone wants your “Hulk Smash” action figure.) In the path analysis you discover that the most popular pages following the blog posts are your search page and your eBay inventory page. It seems pretty likely that adding a link from your blog post directly to the “Hulk Smash” posting on eBay would be the best way to provide a better user experience for the visitors AND ratchet up the auction price on your collectibles.
The scenarios I’ve described here may seem trivial, but imagine that you’ve an e-commerce site with a million visitors a day, and you’ll quickly realize that looking a small percentage of customers to unclear instructions or poor findability is a big deal.
Situation #2: Motivation to Register or Buy
My favorite example of path analysis is much more specific case – namely registration. You’ve probably got conversion funnels set up to track progress through the process itself, but what’s terribly interesting from a path analysis point of view is WHY people are registering. Luckily path analysis can help to answer this question in 2 ways.
First, the pages visited prior to registration may tell you what features of your site are most interesting to a visitor that makes them want to register. But more exciting than that, to me anyway, is where they go immediately AFTER registering. If people complete the registration process and then make a b-line for a specific feature, you should make sure that said feature is promoted like crazy in the non-gated parts of your site. Even if it isn’t the core feature of your site or business.
Will your general usage and page view analytics tell you this? Maybe. But it’s unlikely to tell you what comes FIRST unless you’ve set up specific reports to monitor sequences like that. Path analysis is extremely easy to set up and analyze for situations like this. It is also great if you have a service on your site – once they’ve signed up for (or better yet, purchased) that service, what’s the first thing your customers do with it?
To wrap up…
Path analysis is generally no fun and not much help. Buy there are a couple of really good uses for it – you just have to recognize when it’s a good time to invoke a path analysis report, and whether it’s going to tell you what you need to know about a specific type of visitor behavior.