We recently did a Podcast Episode on CAC being useless.
And today, we released an episode with Leah Tharin.
Amongst many other things, she is very passionate about why NPS is useless.
So let's dive into why.
See here what NPS is if you don't know.
For context, the average NPS score for B2B SaaS is 31-50.
So why is it useless?
1. It's a non-actionable insight
So someone rates your product a 0 out of 10.
Now what?
Is it the whole product? Just a feature? One use case?
How are you going to use this now?
2. It’s a self-correcting metric
Let’s say you want to be scientific with it and measure growth happiness.
You query 100 customers and get an NPS rating.
Then you come back a year later and ask the same cohort again.
The thing is the people who don’t like you (or any of your new features) have already churned.
So by that logic, your NPS should go up.
3. Since there is no standard, you can’t benchmark
There is no one rule on how to survey NPS (because you can’t ask everyone).
Every company is gaming this number just a little bit.
One company I worked at used a color coding scale (red, yellow, green) to tell people that “6” is actually not good.
Another company stopped sending the survey to people that rated the product below 5.
And yet another one, asked them if they like the product or not, and THEN asked them to fill in the NPS.
It’s another really, really dangerous form of lying to yourself.
So what’s a better answer, Toni?
I’ll say it right out of the gate. I am not the expert here. I just have some common sense to add:
TALK TO PEOPLE. There is nothing better than to pick up the phone and ask your users some questions. I am an expert on RevOps though, and I can tell you that talking to people outside your company will 10x your value. So start with customers.
Exit interviews. Just as you should be asking why someone bought your product, ask them why they left. Many times the answers won’t even be product related.
Place questions about specific features at the point of time when someone uses that feature. E.g. you rolled out a new feature, once someone used it, ask then.
These three items might be a bit more effort. But at least you can do something with those insights.