Join the next cohort of the PLG Certificate Program. Application closes Feb 1.
Join the next cohort of the PLG Certificate Program. Application closes Feb 1.

Retention Metrics That Matter

Retention often doesn’t get talked about as much as user acquisition. Yet retention can make or break a business.

This article looks at the core retention metrics you’ll want to monitor, along with how to measure and improve them. 

You’ll learn:

Ready? Let’s go. 

What retention really means and how to define it

Retention is a measurement of how long your customers stick around and actively engage with your product or business. 

It can take a lot of different formats and there are a lot of different flavors of retention based on your industry and your company type, but it's about having a base of users that experience value through your product. Because of that, they stick around for some period of time. 

The essential retention metrics you should be measuring (it's not the same for everyone)

Retention can be looked at from two primary lenses:

  1. Core retention
  2. Proxy retention

Let’s briefly look at each of them.

1. Core Retention

The core retention metric is the total number of users who are still active after a certain period of time. 

If you acquired certain users three months ago, what percentage of them are still around in Month Three? 

With core retention, users gain value from a product for a given point of time. At some point, you have a clear indication of whether or not that user is going to stick around. I.e. if the user stays after a certain length of time, they’re probably going to stay long-term.

2. Proxy Retention

Proxy retention metrics are action-based metrics. They are the core actions users need to take to support your model. 

If you're an enterprise SaaS product selling an email platform that does outbound sales better than anybody else, your proxy metric is the actual daily usage of that product.

A great example is Netflix, the world's biggest subscription model. The underlying mechanics of how they understand the quality of users that are coming into their funnel and whether or not those users are going to stick around is a proxy metric. 

For Netflix, that’s 18 or 19 hours of content consumed per month, per user. If a user consumes that much content in a Netflix model, it's highly likely they're not going to churn out. 

The proxy metric is a proxy because it's not pure retention, but rather an action that should lead to retention down the road. By looking at how users are behaving today, you can make accurate predictions on how they're going to behave tomorrow, in a week, and in three months.

Along with proxy and core metrics, you’ll want to ensure you have a product-market fit, which gives you a sense of how market-ready your product is. It’s a measure of revenue against the payback period. 

It shows the percentage of users who find value in your product over a period of time. 

 When you've got product-market fit, you can:

  • Scale your company.
  • Be comfortable spending money on growth.
  • Have confidence in your financial modeling.
  • Have confidence in your retention models. 

Next, you’ll want to measure these metrics with cohorts.

How to Develop Measurable Cohorts

A cohort is a group of users who experience the product value more or less at the same time. 

To do this, look at users for one month. 

The starting month is referred to as Month Zero. If a hundred people experience the product in Month Zero, that's your cohort. The idea is to track the behavior of each key retention metric–core and proxy–for that specific cohort of users.

As a product manager or CEO, you can look at the data per cohort to see which core metrics have improved over time. As you start measuring different cohorts, you can find the value within your product that attracts and keeps users around.

How to build a retention chart

The chart below illustrates an Activation Curve based on cohort data for an arbitrary product. It shows the number of users activated over an interval of time. 

Activation Curve based on January cohort data

In this example, users sign up in the month of January. Over the next six to seven months, the number of users activated decreases substantially. Almost none of them are left by the end of the time period. 

Users fall off doing the core action quickly and the curve never flattens – users just go away. 

This is an example of bad product-market fit and really bad retention. A lot of early-stage companies are actually here. They’ve figured out how to get customers in, but they haven’t nailed the mechanics to keep them around.

As discussed earlier, you want to see these metrics improve over time. 

If we look at the arbitrary product, the curve starts to flatten a bit by March. This indicates users are receiving value out of the product month-over-month. This is when you generally have the concept of product-market fit.

This is likely when you can actually hit the gas a little bit.

Activation curve at PM fit based on March Cohort

How to Build your Own Retention Metrics

Now we’re going to look at how you can implement these metrics at your company. 

1. Have measurement tools in place

To build your retention metrics, you need to have enough tooling in place to track signups on a cohort basis. 

Use sophisticated tools such as Google Analytics, Google Analytics 360, Segment, Mixpanel, or Amplitude–they generate these types of reports for you. But you can also stick to the basics and a database to manually export the data, and then plug the numbers into a spreadsheet.

2. Know your proxy metric

It’s important to know what to track and measure. 

As an owner-operator or product person, you really need to understand what you’re monitoring in each cohort. You need to gather the raw user and signup data. Then you also have to understand what that proxy metric is: what action is the user performing? Is it consuming content? Is it a download? Is it being a daily active or weekly active user?

Most likely, your proxy metric is going to be very similar to your North Star metric. Your North Star metric is the ultimate goal-based, single metric to own them all. For example, this may be monthly active subscribers. 

The four main factors that impact retention and how to improve them

So many different elements can impact retention, making it a difficult task. The first step is nailing down your proxy and core metrics.

There are four main ways you can effectively impact retention, with the first two being the most likely to produce the greatest results:

  1. Focus on new users
  2. Target better users
  3. Kill features that don’t provide enough core value
  4. Change your model (or your pricing)

Let’s explore each of these. 

1. Focus on new users 

The easiest way to improve retention is to focus on early customers and the initial experiences they have with the product. This is important is because once users have experienced your product, it is very hard to change their perception of it.

Here are some tactics you can use to create a great first impression:

  • Run a bunch of experiments on user onboarding to test various methods.
  • Include personalization in your sales process. 
  • Build better hooks into your products users experience its value faster.
  • Incorporate feedback loops with your customers as fast as possible. 

2. Target better users

Not every user is going to be the ideal customer, and not all channels are created equally. 

When it comes to targeting better users, it's less about the actual channel itself and more about looking at that channel as a cohort, and understanding how those specific users are performing. 

For example, let’s say you are targeting users through paid acquisition using Google and Facebook. While users are going to come in through those channels, there’s also going to be a bunch that arrives via organic channels.  

Instead of looking at retention and even activation as a blended set of users, you need to take each of these targets, split them out into their own little mini-cohorts, and measure those over time. 

What you're likely to find is some channels work a lot more effectively to attract your ideal customers. As a result, double down and scale those channels.

3. Kill features that don’t provide enough core value

Overbuilding is probably the number one issue that startups experience universally across the board, regardless of industry. Product managers, engineers, designers, UX people, researchers, strategists–all want to build stuff that makes us feel really good.

In an ideal world, the build does add value to the product. But the reality is that a lot of what we build doesn't connect enough value to a large enough percentage of people to justify the time spent on the build. You need to understand where the real value is coming from. That’s where you look at both qualitative and quantitative data. 

You need to look at how a rollout or a build is impacting retention metrics, and kill features that don’t add any value. 

While you're not going to know what features should be killed for a while (think six months to a year), you will be able to track your proxy metrics against those feature builds. 

How do you do this? 

Run Firebase experiments. Split test. Build feature flags in your product based on how key metrics are being impacted. 

But remember, the product management process is always evolving.

4. Change your model (or your pricing)

In the early days, you can change your pricing and be okay. But now it’s more and more difficult since you’re charging users on a monthly basis.

At Sandbox, we never had the intention of building a subscription business. But the opportunity came about when we realized users would transactionally buy over and over again for a relatively short period of time. 

They trust us, and they value what we're putting out. As a result, we extended the lifetime of the customer by adding more features they could purchase. After three months, users stuck around. 

So we thought about it differently. What if you get X number of units of value for Y amount of a cost? 

And it actually worked. 

We haven’t transitioned 100 percent to this model yet, but we’re close. We did this because we saw the opportunity there…we could potentially increase the lifetime value of our customers by making a subscription offer.  

Keep Retention Top of Mind

If your retention curve doesn't continue to flatten and just slowly falls to the floor, it most likely points to a value problem. 

If you talk to customers after they’ve churned and ask why they're churning after a certain amount of time, they're likely going to give some tough love. You might find your pricing or model is a reason, but it's likely not the only reason.

Take the time to think about retention as a company with your main stakeholders. It’s as important as gaining new customers. Understand your core and proxy metrics. Be mindful of the ways that you can actually impact retention. Read up on it. Make it a part of your daily habit as somebody interested in product management and growth to own it.

Craig Zingerline
Craig Zingerline
Craig Zingerline is a 6-time founder who built Growth University to help founders & startups become amazing at Growth.