If you’ve ever looked at a skill, notice that several of the reviews are on the same date, and thought to yourself, “That’s odd…”, this is the post you’ve been waiting for!
But first, light stats
To get a better understand of the data, we calculated the number of reviews per day (NRD). We got this number from:
(number of reviews a skill has)/[(date of last review in our database*)-(date of first review in our database*)]
The NRD represents the average number of reviews a skill received per day between its first and last review. An NRD of 1 means that the skill received an average of 1 review per day in the time span between its first and last review. The NRD from our dataset ranges from almost 0 to 16.
Looking at the data
Our first analysis looks at the NRD (vertical axis) and the total number of reviews (horizontal axis). The number of reviews is on a logarithmic scale, so the distance between 1 and 2 is represented as larger than the distance between 101 and 102.
Beyond the few outliers (super not suspicious), what we see is that skills tend to receive under 1 review per day until they reach a threshold of 1000 reviews. At this point, something happens and the skill picks up enormous momentum. Of the 55 skills with more than 1000 reviews, only 51 have an NRD above 1. Interestingly, of the next 55 most reviewed skills, only 6 have an NRD above 1. We can’t draw any firm conclusion from this, but whatever the reason, it seems like something beneficial happen to skills that hit that 1000 review mark.
Do you have experience with a skill hitting that 1000 review mark? Can you see a significant jump in engagement after that point? Let us know!
Some notes on the data
The data we used is from mid-November and only includes skills with 10 or more reviews. It also only captures the US marketplace. It’s important to note that while our data set is probably the best out there, it’s not perfect*. The unfortunate truth is that only Amazon really knows what’s going on with Alexa. Hit me up Amazon if you want to share some of that sweet sweet data!
