Learn how to have a deep understanding of the quality of engagement on your voice app by leveraging features such as frequency of uses, cohort, visit/user and session lengths.
Hi, I’m Justin Grover. And this is the third video in our voice Analytics series. So, today I’m gonna pretend to be a product manager for a voice skill for a national grocery chain. And one of the things that I really wanna do is I wanna understand the quality of my engagement with my users. So, there’s a couple different ways I can do this. One, I can look at how often my users use the voice skill. And what the frequency of that use is. So, here I have the number of visits, and the number of unique visitors. I can compare this to see how many visits per visitor there are. I’ll simply select the two columns, create a metric from the selection, and I’ll divide them. What this does, this creates a brand new calculated metric. You can see that I have two visits per visitor, for each of the days. However, I may wanna increase the number of decimal places that are shown so I can come in to the calculated metric and increase this up to three, and get a more granular view of the average number of visits per visitor during a week. Here I can see that during the week of February the 17th, I had 1.7 visits per visitor. That week we ran a promotion promoting the voice skill in our stores. As you can see, it increased the amount of usage for the skill. However, that usage didn’t stick, and slowly began to taper off.
Another thing that I can do is look at how long people go between uses of the app, using the Cohort Table. Here I have an Inclusion criteria, which is what will include them in this analysis. And then I have a Return criteria, which is what signifies them coming back. I’m gonna choose a granularity of week. I’m gonna do, a Retention analysis. This is a Cohort Table, and what it does is it shows us how often users return. So, the first column here shows us the number of users that, that came each week. And then the second column shows us one week later, how many of those users returned. As you can see, it ranges between 30 and 22%, and slowly tapers off after that. Now 30% is good, but it falls short of my goals. One of the things that we can do is we can start to feature our voice skill more prominently in our in store displays, and drive that retention up to above 35%.