Which statistical indicators do you find MOST useful to analyze your big data in commerce and why?
Chantal Cho
2 replies
Average Revenue Per Paid User (ARPPU)?
Retention rate?
Return On Ad spend (ROAS)?
Hellooo. A number of our clients told us that they find those indicators very useful and how about you guys? 💡
We are based in S. Korea preparing to launch our Commerce Big Data Analysis AI service in the US market as well and we were curious / open to any good ideas of key indicators you ACTUALLY use in your commerce. I always think it is a good idea to collect VOS and apply them as much as possible in actual product/services before the launch. So pls feel free to comment!
-Chantal from Laplace, AI driven analytics for your commerce and retail-
Replies
Daniyar Yeskaliyev@dan_yes
For me, the sigma from the forecasted weekly revenue.
Weekly revenue number is very important because having weekly totals helps avoid intra-week seasonality, and also gives enough time for the business team to respond to any 'radical' shift in the numbers.
52 weekly data points form the annual revenue, and the line chart reflect the seasonality really well.
When comparing data year to year, it's easy to measure the mathematical expectation for any given week - including the expected growth per each year in the forecast.
So, having that math expectation, and historical data points, it's very easy to calculate the standard deviation and sigma.
The KEY here is that according to the 3 sigma law, you can say that with 99.7% confidence your weekly revenue won't be lower or higher than Math Expectation + or - 3 sigma.
It gives a very solid understanding of the normal expected lows and highs.
So, when the real revenue number for any week is beyond that range, it means that something extremely harmful or extremely good happened to the business during that week and that it require the attention of everyone involved immediately.
This can be applied to any other fluctuation number that can be measured on a regular basis and that is important to the business.
Good luck with your launch guys, it's a very interesting niche!
My friend @imanmoaz is actually a great expert in Data Science and maybe he can share what is the most useful stat indicator for him, but he is preparing for his own launch next month so I'm not sure if he has time for the detailed response.
Share
