As brand reputation monitoring becomes ever more popular with tools such as Radian6, NetBase and our very own, GOSO, companies are starting to ask us more and more about quantifying the data and possibly using it as way to measure success or another words, performance amongst their locations. This has always been a challenging topic for us because there are many subjective factors at play and we can’t just arbitrarily determine a score and call it our own. I believe transparency is the key to developing a score, but averaging two data points, average review and count, can be tricky.
For instance, to calculate a cumulative score for any location based on an average score (1-5) and a certain number of reviews, you would have to multiply the average score by the times the number of reviews. This would essentially give you the total number of stars that location had ever accrued.
However, this won’t work in comparing two locations because of the following situation:
Location A has an average of 4 stars but only 2 reviews and Location B has an average of 1 star but has 10 reviews.
If you multiple the average by the number of stores you would get the following:
Location A: 4 star average x 2 reviews = 8 stars
Location B: 1 star average x 10 reviews = 10 stars
As you know, the more stars you have the better, so in this theory Location B should have a better overall review score because in total they have 10 stars where Location A only has 8. This obviously doesn’t work though because a review with 1 star is deemed bad, and if you have more of those, it should only weaken your score, and not make it better. So how do we solve this? In order to explain clearly, I’ve designed a simple chart to that divides a review score into 4 simple to understand categories.
A couple of things to take into consideration:
- Number of reviews can be infinite (use the highest number of reviews for that industry or organization)
- Review score is finite (5 is the standard maximum, if you run across a site that has 10, just divide it by 2)
The two baselines which breaks it down into 4 sections are dependent on a couple of things:
- The industry: a benchmark for lets say an automotive dealer and a restaurant will be very different. A consumer will hold a much higher perceived standard for a restaurant or hotel.
- The organization: a franchise or large corporation with many locations, could ultimately determine their own internal benchmark which they consider acceptable and adjust the formula accordingly.
Posted in Announcements
