Measuring self-serve in adoption

Josh Rosenthal

In our breakout room today, a question we had was [as posted on LinkedIn]
what metric(s) do you use to measure self-serve in adoption?
This would be akin to ticket deflection in Support.

That is, how do you measure the effectiveness of each asset you use to enable customers to self-serve their adoption along their customer journey?

Ed Powers

Hi @Josh Rosenthal,

I assume you're talking about a zero-touch or light-touch, high-scale engagement model, and I've seen this done in some interesting ways.

Conceptually, as users come up the learning curve, usage frequency and access of specific features will ramp up and stabilize over time, usually tracing a logistic ('S') curve. This is your dependent variable (y). You probably have a characteristic that shows what a successful customer learning curve looks like. If you think of this as statistical control limits, you would then measure the deltas between where a customer is vs. the ideal curve at key points in time and set bands for 'Green,' 'Yellow,' and 'Red' actions. In this manner you can trigger emails or other actions to get learners back on track. 

You should also be collecting data on an individual account's access to self-serve learning materials and be able to associate them with usage (y). These are your independent variables, x1, x2, x3 ... xn. You can then use factor analysis or regression to study the effects of the x's on y, or any other dependent variable you choose, such as volume of inbound, training-related Customer Support tickets. Having this data enables continuous improvement because you can study cause-and-effect relationships. 

This should give you the gist, but I'm happy to discuss further: [email protected]. I'm also teaching a course that helps with analytics, in case that's of interest. 

Ed