Session Replay: From Nice-to-Have to Mission Critical
By Jake Makler on Oct 8, 2018 11:02:07 AM
I remember where I was the first time I saw session replay - what appeared to be a video of a customer’s actual online session played before my eyes. I was sitting in the cafeteria at The Hartford, an insurance company where I worked as a web analyst at the time, and someone showed me how to pull up any user who had visited our website and watch their journey as if I was standing over their shoulder. I could see every click, every tap and exactly what the website responded with - an error, a success message, or nothing.
Session replay helped put us in the customer’s shoes and understand if issues they encountered were primarily based on technology or, more commonly, usability. The tool was actually pretty limited in what it tracked and how accurately it rendered the experience of the customer. A fair bit of guesswork, inference and imagination was required but, as a first-of-its kind capability, it was really exciting!
My initial job entailed reviewing comments written to the website via a customer feedback tool and correlating that data to the associated behaviors from the session replay tool. Was this person citing a legitimate problem? Or just ranting because they were having a bad day and couldn’t reset their password (we’ve all been there). As time went on, I was introduced to other technologies that took it to the next level – not only capturing and replaying what was clicked but actually tracking mouse movement and mobile touches both at individual (DVR-like videos of customer sessions) and aggregate (trending/reporting) levels. Over the years I’ve watched session replay grow from a niche, nice-to-have tool into a mission critical data source often earning its own spot as a budgeted line item.
Although I watched this space evolve for many years, I found uses for this technology remained limited to ad hoc project-based work and qualitative reviews. Product teams would analyze a particular part of the site, have a team of analysts spend weeks aggregating data and searching for interesting replays to drive recommendations. It turned out no one had the time, let alone the patience, to spend their days sifting through millions of hours of unstructured recording data.
While a session replay looks like a video recording, it’s actually millisecond-level chunks of data – for example, representing the coordinates of a user’s mouse and associated click/tap actions - being streamed and interpreted as visuals. Because a session replay is a compilation of data, then why couldn’t an intelligent algorithm be written to automatically identify the proverbial “needles in the haystack” without expecting a human to watch hours upon hours of replays? A solution built to do this well first would by definition be the most effective way to utilize session replay data to optimize digital experience for users.
This modern Digital Intelligence Platform would go beyond session replay and act as the operations center of the digital age. It would be a centralized hub that empowers product, engineering and analytics teams to act faster - and with higher confidence - in their efforts to improve digital experience.
Having spent six years helping organizations build practices around the evolution of the “digital experience NOC,” I want to share a handful of critical capabilities for anyone considering making an investment in this space:
- Performance. It’s critical to have a way to capture all of your web and native app traffic without negatively impacting performance. Having an ability to analyze all visits without impacting performance helps ensure congruency between this dataset and the dataset from your other analytics tools. I often hear about tools that, because of the potential performance impact associated with capturing such a massive amount of data, are deployed in a limited scope to capture only a portion of traffic. While there may be valid arguments for this, sampling greatly limits the value of this investment.
- Real-Time Dashboards and Alerts. The agility with which you can access and action on data is the difference between a tool that quickly becomes shelf-ware and a solution that measurably impacts revenue and profitability. Table stakes for this requirement are that the solution should also have performance monitoring capabilities built in to “keep it honest” (see above).
- Pixel-Perfect Replay. The atomic unit of digital intelligence is the session replay – and your trust and confidence in the data is critical. With the rise of single page architectures and websites becoming more and more dynamic and personalized, replays must render perfectly out of the box.
- Security. Many tools have come under criticism for treating security as a bolt-on feature or afterthought. It’s essential to provide security and privacy by default, with a mechanism such as public-private key encryption to secure PII (and block PCI) at the client source. The use of these keys should be audited within the tool, so that granular access can be granted and monitored in case of a legitimate business need (customer service, fraud detection, etc.).
The above represents only a handful of requirements you should verify in evaluating a potential solution. Want to learn more about the digital experience NOC? Request a meeting to discuss further.