If I'm a business owner and I see that Shopify knows my business as well as I do, I start to ask questions. What's stopping them from figuring out what the margins are in each of these businesses and starting up a competitive business with a few business that have the highest margins?
A diversification of cash flows streams never hurt anyone right? ( Yeah I'm looking at you Amazon)
These are some really professional tools for measuring churn and retention. I've always been really impressed with Cam's work (and the tools he's chosen like probability models and survival analysis).
> Far too often businesses define churn as no purchases after N days; typically N is a multiple of 7 or 30 days
I'm working for a strategy consulting firm, advising large retail clients.
Even though we have looked at doing this kind of advanced modelling for churn metrics, we always end-up using the same kind of very basic rules of customer not purchasing within the last x days.
The reason is twofold: (1) it’s trivially easy to implement and explain, (2) at the end of the day using the complex method is not worth the effort. Because business-wise, there are many other external factors that are much more impactful than the modelling biais – like seasonality, variation by category of product, etc.
It all depends on the decisions those metrics are used for, but most of the time it’s not worth the pain.
That is really interesting. I am working on this right now. It was an interesting piece but I agree with both of your points. Hopefully you can spend the time you save on more interesting questions like, 'why' and 'how do I affect that number' and 'how do I predict this in advance of spending $ recruiting the customer'. Or the real biggie, 'how can I write this in a way that can be run in 1hr per qtr using only a spreadsheet and costing $10 or less, which seems to be what clients really want.
6 comments
[ 5.7 ms ] story [ 30.3 ms ] threadA diversification of cash flows streams never hurt anyone right? ( Yeah I'm looking at you Amazon)
I'm working for a strategy consulting firm, advising large retail clients. Even though we have looked at doing this kind of advanced modelling for churn metrics, we always end-up using the same kind of very basic rules of customer not purchasing within the last x days.
The reason is twofold: (1) it’s trivially easy to implement and explain, (2) at the end of the day using the complex method is not worth the effort. Because business-wise, there are many other external factors that are much more impactful than the modelling biais – like seasonality, variation by category of product, etc.
It all depends on the decisions those metrics are used for, but most of the time it’s not worth the pain.