Fullstory Quick Start Guide - Metric Templates

New

This is a continuation of this guide for recommended tweaks to any new FS org you’ve joined.

 

Now that you’ve got your account up and running, let’s talk through some great go-to metrics that you can build in any account regardless of your site.

Wait! Before you go building a whole mountain of metrics, check if any of your teammates have already built them out. In addition to just skimming your library, a great way to see what’s been built so far (and get in on the action) is to review any alerts that exist within your org. You can find alerts at the bottom of the left sidebar. Once you open the alert page there will be two tabs available, one for Segment Alerts and one for Metric Alerts. If you see one you’d like to be subscribed to, open it up and find the Your Subscriptions section and specify how you’d like to receive alerts (note - the Your Subscription section is in a different spot for metric and segment alerts, but both have the same heading). Currently, you can be notified In-App, through email, in Slack, or in MS Teams

 

Ok, let’s get into some metrics that can be set up in any org.

Common Points of Friction - There’s a set of 6 metrics I always build in any new org I join. These will help you discover common points of friction using some of Fullstory’s native Frustration Signals. This set consists of the following dimensionality metrics:

  1. Dead Clicks by Element - Unique Users
  2. Dead Clicks by Text - Unique Users (Shown)
  3. Rage Clicks by Element - Unique Users (Shown)
  4. Rage Clicks by Text - Unique Users
  5. Error Clicks by Element - Unique Users
  6. Error Clicks by Text - Unique Users

 

 

As you can see, all of these metrics look at user-level data, but it’s worth building similar metrics at the session and/or event level for additional insight into potentially frustrated behavior.

Other groupings can provide interesting insights into how Frustration Signals vary (such as by Page), but I’ll leave those for you to discover 😉.

 

Page Performance Metrics - In addition to performance metrics like Page Load Time and First Contentful Paint time, Fullstory also allows you to measure performance via Core Web Vitals. There are a few metrics available, but the big 3 you’ll want to monitor are Interaction to Next Paint (INP), Largest Contentful Paint (LCP), and Cumulative Layout Shift (CLS). If you’re not already familiar, Core Web Vitals were developed by Google and consistently poor scores can reduce your page’s ranking in Google searches. Even if your pages aren’t on the clearnet, these metrics still provide valuable measurements of how responsive pages will feel to your users. Thankfully, Google provides guidance on what kinds of scores you should be shooting for, and FullStory enables you to build metrics to compare against these benchmarks.

 

Tripping Hazard! Many of the following metrics will rank dimensions (e.g. pages) by a calculated number that does not account for population size. For example, ranking pages by average page load time could result in something like this:

When you see this, be sure to check the traffic that each of these pages received. You can do this by scrolling down to the session playlist and filtering by that page (from the Filter by dimension… drop down). If you find the outlying result only has a handful of visits, you can discount this result as not statistically significant. It’s best practice to do this for any dimensionality metric that includes an operator or some method of aggregation, but it’s most important when you have these extreme outliers.

 

Here are some Core Web Vitals metrics to help keep you up to speed on page performance:

  1. INP Time - 75th Percentile by Page (Shown)
  2. LCP Time - 75th Percentile by Page
  3. CLS Score - 75th Percentile by Page
  4. What % of INP times are over 500ms?
  5. What % of LCP times are over 4 seconds? (Shown)
  6. What % of CLS scores are over 0.25?

 

Still hungry for more? Check out the Fullstory Cookbook for more than 50 more recipes for other metrics to help you dig into your User’s behavior!

 

If you have any go-to metrics that don’t rely on specific pages, elements, or anything else that is org-specific, put them in the comments to help out your fellow Fullstory users!

0

Comments

2 comments

  • Thanks for sharing this Logan Rowley , this is really detailed and a great help to newer users of Fullstory! I haven't ventured to the dark mode side but I feel like it's a very popular choice :-)

    I always recommend users create a metric like this looking for bots and unknown devices, like this:


    Setting the group by option to “User Agent” and clicking on the Dimensionality view is a great way to check your session quota is not being used by traffic you don't want to see in your account.  More info on how to block by user-agent can be found here.

    If your plan has access to Metric Alerts, you can set an alert to keep an eye on what's going on if some of these numbers start creeping up without needing to keep checking back on your metrics and dashboards.

    Sometimes I also look at a custom segment set to exclude users who clicked anything:

    This can help me dig into user's who aren't engaging - where are they coming from? Do I have ads running / a UTM campaign that may not be attracting the right audience? Are these users landing on a particular site on my page? It opens up another avenue to explore how visitors perceive your site.

    Hoping to see other users chime in on this thread and share how they use metrics - Please keep sharing any tips and tricks with us!

    0
  • Hey Gemma Atkins, great addition! For anybody interested in blocking bot traffic, double check your data capture settings and make sure that you've used the option to “Block popular bot UAs”. If you're still getting bots after this, then you can copy the user agent from Gemma's metric and add it in this same settings page with the “Add blocked UA” button (Note - Technically you can copy a substring of the UA and block based off of that, which can help with blocking multiple bots with one rule. However, if you don't know what you're doing here you can accidentally create a rule that will block significant numbers of real people, so proceed with caution.)

     

    For anyone who worries that they're still getting high numbers of bot traffic even though Gemma's metric shows little results, it's possible that you have bot traffic that isn't declaring itself a bot. You can search for these types of sessions by using her same metric but removing the device type refinements so that you're viewing all traffic by UA. There's no guaranteed way of identifying this traffic, but here's some tips I've picked up along the way. 

    • Old versions of Chrome - As of June 25 ‘24 the current version of Chrome is 126. You should expect the majority of your Windows/Android traffic to be within a couple versions of the current one. If you find people who are on versions that are several years old (e.g. anything starting version 7x or 8x) there's a good chance that traffic is automated
    • Linux X11 - X11 is a windowing system commonly used for remote machines. This may be included in the user agents of bot traffic. Note - Google has recently revamped their user agents for Android and many Android devices now include “Linux” in their UA. As a result, I would STRONGLY advise that you do not add in a blanket block for “Linux”, instead using a more narrow definition such as (X11; Linux)
    • Anything resembling a URL - If you see a .com or http in someone's user agent, this is a very strong sign that they're a bot. Don't conflate this with any brand name. You'll occasionally see things like FB or Instagram in user agents when sessions are created within an embedded in-app browser, and these are real people.

    Note - None of the above “red flags” are a guarantee of bot traffic, they only predict it. You should still review a large number of sessions and build a couple metrics to dig further into any User Agents this helps you identify to ensure they're actually bots before you block them.

    0

Please sign in to leave a comment.

Didn't find what you were looking for?

New post