Back
Milestone
Posted

🏴‍☠️ AARRR! Which metrics I plan on measuring

In the previous post I explained I want to take a more data-driven approach when it comes to product development. But which metrics should I look at?

Ultimately the goal of a business is to generate a profit. Setting aside costs, we can just focus on revenue for now. In case of WIP, that's our monthly recurring revenue or MRR.

But focusing on just MRR isn't really giving us much insight. What's more useful, is looking at what makes up that MRR. Basically that comes down to this:

MRR = number_of_website_visitors × signup_conversion_rate × retention_rate × average_revenue_per_user

Notice how they are all multipliers. That means that if any of those numbers is particularly bad, it  brings down the whole thing. It also means that something small like increasing the conversion rate from 2% to 4% (not real numbers) has the same effect on MRR as doubling the amount of website traffic. 

As you can see it pays to know which metrics to work on. Knowing these numbers will help us better understand where our weaknesses are and opportunities for growth.

The numbers I mentioned don't tell the full story though. For a more complete picture we want to look at AARRR as pioneered by Dave McClure. Here's what that means for WIP: 
  • Acquisition: new makers who sign up for WIP
  • Activation: get them to actually use the product
  • Retention: make sure they keep getting value and don't cancel 
  • Referral: help them spread the word and refer new members
  • Revenue: optimize pricing based on value
Specifically, here's what I plan on tracking:

Acquisition
  • How many website visitors do we get (not logged in)
  • What percentage vies the join modal
  • What percentage actually signs up
Activation
  • What percentage of users completes at least 10 todos within their first two weeks
Retention
  • What percentage of users is completing X+ todos per week (will need to figure out good number)
  • What percentage cancels their subscription every month (churn rate)
Referral
  • How many invites does the average user receive (currently really low)
  • What percentage of these invites gets given away
  • How many of those invites get redeemed (redemption rate)
  • What percentage of invited users ultimately signs up to a paid membership
Revenue
  • What's the total MRR
  • What's the average revenue per user (ARPU)
  • What is the life-time value of a customer (LTV)
They are mostly percentages/rates as those are the most actionable. Absolute numbers like overall signups don't really tells us much, but know what percentage of website visitors ends up signing up does give us a good indication of how well the website converts (regardless of how much traffic we get at any given time), and whether our changes are having the desired effect.

My next step is to calculate some of these metrics so we can start measuring them. I'll follow up with that later.

Let me know in the comments what you think of these metrics and if there's anything else you think I should be measuring. Also let me know what you're measuring for your product.

P.S.
For more inspiration check out this slideshow by Hiten Shah.


For me the core and center for every retention-centered business is CLV/LTV

You tell me - but in my humble opinion retentions your super-over-goal..

if you get just this one right, but f**ck up acquisition and referral you'll still be fine.

Simple but effective -
1. Calculate average CLV
2. Calculate statistical remaining CLV per user
2a. Your model needs to attribute for social impact / "referral strength" or something alike
3. Build simple model to estimate risk on each user
3. risk * remaining clv gives your priority

Decide every day - what should I work on? remove the highest risk..

Great points. I agree retention is the most important. Without good retention, the other metrics don't matter. I think there's also a lot of opportunity for improving Activation. WIP is really a habit-driven product and if people don't get into the habit of using it early, they tend to churn quickly.