Measuring growth in EXTREME seasonality

(Andreas Rolén) #1

Hello folks!

Andreas here from Forza Football - one of the most popular football (soccer) livescore apps in the world.

We have a huge problem in our company with seasonality. I’ve tried to come up with another industry that has similar problems, but its hard. We want to measure our progress by looking how grow week by week, month by month, and year by year, but we simply can’t… What should we do?

So why is this?

Well. One month from another differs a lot, one league could be on break, it could be qualifications for some international tournament like the World Championship (which attract a lot more users) etc. Week by week does not work either, because even though Premier League is played both weeks, the first week two big teams, say Arsenal vs. Chelsea could play, and the next non of the more popular teams face eachother. Year by year kind of works, but last year it was the Euro Championship during July, attracting a lot of users during July (which during other years is the month when every league is on break) which then churn somewhat more during August (which during other years are the month when the most users resurrect).

The only way we found to really be able to compare our data is two years back on the same weekday during the same week (e.g. Saturday week 41), but this kind of doesn’t work either since the product evolved a lot during that period.

How would you do in this situation?