How do you get buy-in from your data team?


(Jess Burnham) #1

I’m currently the only person in my company running our online community, our company is roughly 170 people, and our data team has about 8 – there’s obviously a lot of strain on our data team to keep up with demands.

This quarter I’m looking to prove that if our customers sign up for our community, they are 99% (or more) less likely to churn.

Can I get some feedback into how i can phrase this request for our data team to apply the resources to this task? As I said, I’m a team of 1 so I rely on y’all for help in this area :slight_smile:

Thanks in advance!!


(Richard Millington) #2

I think you have two distinct challenges here.

  1. Getting the data team to support the project. Every team is different and I don’t know what they’re motivated by or why they would support this. What’s in it for them?

  2. Proving the impact itself. The structure of the question (reduces churn by 99% is rather biased). At that level, you’re almost certainly not going to get statistically significant results. What you want instead is probably a multiple regression analysis reviewing various customer behaviors to determine which are statistically significant predictors and what influence they have. You will probably find, for example, a range of factors influence churn (price, size of the organization, community participation, opening marketing emails) and each to a different degree. Community will be a small part of that.

This highlights a possible relationship, but not a causal one for that you need to do a controlled test (e.g. one group gets exposed to the community and the other doesn’t) and track the results. The easiest way to do this is simply to close the community to new registrations for a month and track the churn rate of that group.


(Jess Burnham) #3

^ I’d probably get fired :joy:

In regards to the challenges, what’s in it for our data team if they support this project? About the same as whatever they get out of any other project that’s outside of their department, which is probably just a lot of thank you’s and their paycheques.

Your 2nd point is an interesting one. Our team here has done a deep dive into reasons for churn and have pulled out themes, like the ones you mentioned (price / org size / not ready to commit etc.) and have all of that data stored. I could potentially compare the email addresses in any of those categories and see whether or not they signed up for the community before.

This has given me something to work with!! Thanks Richard!