Most companies don’t plateau at $2–10M ARR because of product or marketing. They plateau because they’re building for “everyone who might buy,” instead of a crystal-clear someone. This typically involves developing “personas”, “ideal customer profiles (ICPS)”, “jobs to be done” or something similar. The tricky part here is that most of the traditionally processes here are a waste of time. I’ve seen teams waste months here with very little to show for it. This is the process that I’ve found to be much faster and impactful. What a Persona and Why Does It Matter? The single biggest inflection point in most companies happens when they know these groups, collect the data, user the data correctly and then change their strategy. Let's say you sell a product that digs holes. What kind of hole does your customer want? Well that depends if your customer a is: 5 year old at the beach?  A hobbiest gardner?  A construction worker?  An oil prospector? 5 year old at the beach? A hobbiest gardner? A construction worker? An oil prospector? All of these people want a hole. However, why they want a hole dug and what that hole has to look like is very different. Because of that, the value that they put on a hole is also very different. A 5 year old at the beach might only want a single hold dug and doesn’t have any money to pay you. An oil prospector might pay $100k per hold and might need hundreds of them. Granted these holes need to look different, but that’s kind of the whole point. Problem Centric Personas There are lots of different methodologies to define your users personas. In my experience, the most useful one by far is grouping users by the underlying problem they are coming to your product to solve. This is the best indicator of both their retention potential and willingness to pay. At Codecademy, we had four main personas: Hobbyists: People who find coding interesting but don't need it for anything specific. Students: People learning for school assignments. Career Upskillers: People adding coding to their existing job skills, so a financial analyst who wants to move from excel to python + SQL. Career Switchers: People who want to become full time engineers or data scientists. So this could be a journalist who wants to become a fulls stack engineer. Hobbyists: People who find coding interesting but don't need it for anything specific. Hobbyists Students: People learning for school assignments. Students Career Upskillers: People adding coding to their existing job skills, so a financial analyst who wants to move from excel to python + SQL. Career Upskillers Career Switchers: People who want to become full time engineers or data scientists. So this could be a journalist who wants to become a fulls stack engineer. Career Switchers “Hobbyists” don’t really need to learn to program, it is literally a hobby. They were typically retirees or people who just like learning.  “Students” had a need, but learning to program didn’t unlock any immediate earnings, so they retained well on the free product but didn’t have high willingness to pay. “Hobbyists” don’t really need to learn to program, it is literally a hobby. They were typically retirees or people who just like learning. “Students” had a need, but learning to program didn’t unlock any immediate earnings, so they retained well on the free product but didn’t have high willingness to pay. They typically didn’t buy the paid product. If they did, they didn’t stick around. However, “Career Upskillers” and “Career Switchers” had a real problem we could solve. If they gain new skills, they can make more money relatively quickly, so not surprisingly they had the highest LTV. When we used this segmentation information in our data, we saw something like the below. Our best personas had a meaningfully different retention rate and LTV. How to Build Your Personas Like all good consultants, I came up with a catchy & complex framework for you to solve this: Have your founder guess. Have your founder guess. In my experience most of the long expensive research frameworks here are a waste of time. Founders should have been in your problem space long enough and spoken with enough users that they can probably segment people with about 30 minutes of thinking here. If they don’t, you have other problems. You’re better off forming some rough categories and filling in details as you speak with users in the course of routine user research. As with most things in the startup game, velocity is everything. Using This Data to Make More Money For this data to be useful, you need to do the following things: Collect it  Integrate it across your analytics so you can see the differences  Modify the early experience to maximize activation rates.  Use the overall lessons to drive product development & marketing. Collect it Integrate it across your analytics so you can see the differences analytics Modify the early experience to maximize activation rates. Use the overall lessons to drive product development & marketing. The easiest way to do this is with onboarding surveys. I've never seen any material drop-off in early-stage engagement from onboarding and if you use this data to customize the early-stage experience, which you should, you will likely increase activation. Let's take a look at a few examples. Duolingo does this. they asked for the core motivation up front as they can't infer this from the course data. So What Do You Do With This Information As stated above, the basic playbook looks like this: Map out your personas with your founder. Do it quickly.  Set up something during onboarding that allows you to segment them.  Use this segmentation data in all of your key metrics. Look at it weekly.  Pick your ideal persona  Make sure both marketing and product optimize for this over the others. Map out your personas with your founder. Do it quickly. Set up something during onboarding that allows you to segment them. Use this segmentation data in all of your key metrics. Look at it weekly. Pick your ideal persona Make sure both marketing and product optimize for this over the others. The hardest part of this by far is having the discipline in to focus on the best persona in your marketing and product development. This cannot be over stated. The larger your team and/or the longer you’ve been doing this, the more resistance there will be to stop doing things that kind of work and start trying to find thing that really work.

Churn Isn’t One Problem—It’s Four

Newsletter

Read My Stories

The Persona Process Is Broken—Here’s the Faster One

About Author

Comments

TOPICS

THIS ARTICLE WAS FEATURED IN

Related Stories

Annual Plans as a Churn-reduction Lever

10 Best Datasets for Time Series Analysis

12 Best Pre-Installed R Datasets Commonly Used for Statistical Analysis

Import JSON To Google Sheets - 3 Best Ways To Do It

How To Import External Data Into Google Sheets Without Copy/Paste

253 Stories To Learn About Data Analysis

Annual Plans as a Churn-reduction Lever

10 Best Datasets for Time Series Analysis

12 Best Pre-Installed R Datasets Commonly Used for Statistical Analysis

Import JSON To Google Sheets - 3 Best Ways To Do It

How To Import External Data Into Google Sheets Without Copy/Paste

253 Stories To Learn About Data Analysis

Light-Mode

Classic

Newspaper

Minty

Dark-Mode

Neon Noir

Minty

HN StartUps