Two years ago, I launched SQL Habit with only 50 lessons (it’s 300+ now). I was desperate to find the first “real” customer – someone I wouldn’t know and who finds value in learning with SQL Habit.
That someone was Rob – CEO of Feather Insurance. We got to know each other in a perfect moment – Rob wanted to build the first dashboards for the company and didn’t know where to start with SQL and Data Analysis.
I remember we first met with Rob at the Feather HQ at 8 AM (we’d later always meet at this time for our “Data Breakfasts”) and after a short intro Rob personally paid me 200 EUR.
Fast-forward 2 years, here’s a quote from Rob’s recent email:
We’re growing as a company, and it’s amazing to have the course alongside our growth. SQL Habit even guides what is next in terms of priority. Like it is really, really amazing.
After this, I knew we had to catch up and retrospect on Rob’s and Feather’s data journey. Here’s the first part of our 2h long interview.
Rob: Anatoli, you’re selling yourself short.
The funny thing is, I started with SQL Habit because I wanted to learn SQL, but I learned how to build a data-informed company. Why didn’t you tell me that upfront?
A data-driven means that the data is a deciding factor in the company.
A data-driven approach is often in conflict with our intuition. For example, we have to roll out a new feature that we think is hurting UX, but the AB-test shows an uplift in revenue.
For example: a hotel booking website shows more expensive options even if you specified a price range (unfortunately, this is a true example).
This approach got some negative traction lately for obvious reasons.
A data-informed means we’re using data to inform our decisions, to help us understand our intuition. We use data to tell us a story (ideally, a full story), but we in the end make a final call.
Anatoli: Hahaha what a start. You’re right though, I might have chosen the wrong name for the course.
Probably, I was in my engineering mind when I was choosing a name. Now I understand that it might sound like “Yet another SQL course” where in reality I’m offering 80% knowledge on how to run a company with data and 20% is SQL.
It’s still the Habit part that I love about the name. It is the main selling point – you’ll be able to answer any question with data in under a minute.
Anyways, why don’t we start and introduce our readers to Feather Insurance?
Rob: On one side, I can bore you with the details around insurance - how we are re-building insurance the way it was meant to be - but actually for your readers it will be more interesting to know that we’re an early-stage B2C company with several marketing channels and a subscription model.
We were 5 people when we started using SQL Habit. Now we’re 15, and we’re looking to keep growing a lot in the next few years.
Anatoli: Rob, tell me about the beginning of your data journey.
When in Feather’s history data became a critical thing?
Rob: In the first 12 months, data doesn’t really help you – you get zeros everywhere. But once you get something working – data becomes really important. For us, the data journey started after we got things working.
In the first 12 months, data doesn’t really help you – you get zeros everywhere.
We started simple – putting dashboards on the wall to see our fundamental metrics and get motivated.
The key is also to be able to build dashboards yourself. Especially as a founder, how can you get things done otherwise?
Also, speed is really important, this is why you need to be able to do everything yourself.
I remember how in the beginning of the course I was reading lessons and building dashboards in parallel.
Anatoli: I feel you 100% on this one, during the first 12 months you have to break through “the wall” and get the data flowing.
What was “the hammer” that allowed you to break through the wall?
Rob: The hammer is to make a single customer use your product. Until you’ve done it once there’s nothing to measure.
Until you have a real customer using your product there’s nothing to measure.
I remember in the beginning I was thinking that one day we’ll have lots of customers, I didn’t realize how hard it would be to get there. It took a while.
The benchmark for us was when someone went through our product without any assistance – then we started to dig into data. Because in the beginning we were talking to everyone and we literally knew their problems first hand.
Anatoli: Damn, you said it so nicely!
Basically, for the first 12 months you don’t need any data. Your data journey begins after 12 months.
Rob: If you’re lucky! Haha
Anatoli: True story.
If a product fails after 12 months – data will be as good as an autopsy.
Rob: Yeah, maybe we can think about it differently.
Feather is not a typical product, there’s no free version. For other product it could be way faster – the moment you delivered value to a user without your supervision data becomes useful.
Anatoli: Oh yeah, totally! In the end, it’s really hard to say when the data journey begins without a stopwatch, but your definition is on point.
Tell me that story when you were putting a monitor with metrics on the wall.
Rob: Vincent (Feather’s CTO) got us to do this at the beginning.
I thought, why are we putting the numbers on the wall? We know the numbers. You see them all the time, you just feel the numbers. As a company, you are the numbers.
I thought that was crazy, that it was a huge waste of money. We do not need a giant screen to put numbers up – I can have a tab opened somewhere.
But once you have it, you know, you don’t realize how you ever lived without something like this.
Let’s recap: once you have data flowing in, start simple. Step one is just super basic visibility on everything. It might even just be “how many people sign up per day?”.
Then, maybe if you get a little bit into a funnel analysis, see where people drop off. I think actually most of the value came from really similar stuff at the beginning. Simple stuff, like “which questions blocked people in the onboarding”.
Anatoli: I love it, agree 3000%, let’s make it a quote:
Start simple. Knowing your percentages is a great example of 80/20 rule.
I think a lot of people are afraid to be simple, but in reality, looking at percentages (CTRs or funnel step conversions) for different cohorts gives you most of the insights.
Rob’s wife brings us some food, and we realize that our conversation should be a podcast. We immediately give it a name – “The Data Journey”. Drop a comment down below if you want this to happen.
The steps to build a data-informed company so far are:
Step 0. Until you have the first real customer using your product you probably don’t even need data. Talk to your users and hear about their problems first-hand.
Step 1. Start simple: get some data to work with (a production database of your product is already enough) and start building basic dashboards.
In the next part we’ll talk about further steps, Data Analysis, mindset for building a data-informed company and data transparency in a company.
In the meantime, check out open positions at Feather Insurance. I personally recommend it, Feather has an amazing product, company culture and data transparency. I’ve been a happy customer for the past 2 years.
If you want to start building a data-informed company – check out SQL Habit for teams. You’ll learn how to use data from examples of a growing startup company.