Growth models: the missing piece in your product strategy

A growth model is one of the best tools we’ve seen for product management. It helps you plan for exactly how you’ll get to success. It helps you identify the levers you can manipulate to acquire and keep users. It can tell you whether the expectations you’ve set (or have been set for you) are realistic. And they help you prioritize decisions for what to build or optimize next.

We first heard about growth models from Brian Balfour, and they’ve become increasingly common in high growth startups. But we’ve never seen an enterprise product team implement one in our consulting engagements, nor have we seen one when listening to startup pitches at our sister company’s venture fund.

Which is unfortunate. Growth models aren’t just helpful when you’re riding a rocket ship. They can be useful pre-product market fit (and even pre-launch) to validate assumptions and provide rationale for how you’re going to get from here to there.

Here’s how to create one.

We’ve created a sample growth model you can play with. You can get immediate access here (scroll to the bottom).

Identify your north star metric.

The first step in building out a growth model is to identify the key metric that will determine product success.

While there will certainly be other metrics you’ll be monitoring, there is tremendous value in identifying a single metric to focus your team.

A good north star organizes everyone’s thinking. It helps you communicate with your team (or investors, or growth board or steering committee). And in our experience it tends to get teams to move faster since they’re trying to improve one thing.

A good north star metric is typically a retention or engagement metric of some kind. For most products, unless you have solid retention you don’t have a viable long term business.

It’s important to keep in mind the nature of your customer relationship — you might have a desire for your user to engage with you on a daily basis, but that might not map to their desire. My mortgage company might want me to have their mobile app on my phone and for me to log in regularly, but if I never have to deal with my mortgage company again I’m happy.

Some examples of good north star metrics:

• Daily active users: good for collaboration apps that are an essential part off my job (email, slack, CRM, etc) or social sites.
• Monthly active subscribers: good for box model or subscription sites. It sounds like it might be good for SaaS products as well, but it typically is a lagging indicator — ideally there is usage data that you can identify that is predictive of eventual churn if you don’t intervene.
• Marketplace businesses usually have some version of a chicken and egg problem. The answer is usually to prioritize the supply side of the equation, as this will lead to the demand side if successful.

That said, consider the nature of your business. WhatsApp prioritized number of messages sent per day, realizing that they were competing with SMS and had a usage pattern of many times per day if successful.

LinkedIn for a long time prioritized number of profiles created. While total accounts is usually something that savvy investors would discount (they care about active users, not total users), LinkedIn realized recruiters were how they made money, and recruiters cared about having lots of resumes to search on.

Figure out the inputs that lead to that metric.

Let’s say you decide monthly actives are the critical metric. The next step is to figure out how to get there.

The goal is to identify the levers that will get you to your end state. Each of these can be manipulated to increase (or decrease) your north star metric.

Acquisition

For acquisition, ask yourself what marketing channels you plan to use. What do you think are realistic numbers for each in terms of visits? How do you think you’ll get there?

Take SEO for example. Often a team will say organic search will be one of the ways they get customers. But how will they get them?

Well, organic search is a function of the number of keywords you rank for, the search volume for those keywords, and the expected clickthrough rate for ranking in a particular position. It’s also likely that ranking will take time. You’ll most likely be engaging in some content marketing strategy coupled with link building. So how many articles will you create each month? Make a guess on average search position across all keywords. Assume that you won’t show up on page one for a while, and that it will take 12 months at least to reach a top 3 position. What does that look like?

Again, the point of all this granularity is to help you frame expectations for yourself and your team. “We’ll get users from SEO” sounds great. But without clear assumptions for how you’ll get there you’re stumbling in the dark and likely to create disappointment when you can’t hit those numbers.

The same should apply to all your other channels:

• PR: how many articles will you show up in? Can you estimate the average page views for an article on that site? What’s the expected click through rate on an article?
• Paid Acquisition: what channels will you be using? What’s the estimated audience for each? What’s the anticipated cost per click and clickthrough rate?
• Social: how do you anticipate growing your audience? What are reasonable numbers for organic reach? How many posts will you be publishing each month?

Activation

What happens once people land on your site? At DI we talk about activation instead of “user registration” because activation is more predictive of long term engagement, and there’s often a drop off between registration and them engaging in your core experience.

Identify the steps users have to take between visit to registration, and between registration and activation. Create some estimates for conversion rate for each step (if you have data already, plug that in instead. Try not to throw up.) Now you’ve identified the areas of product friction within your product and can work toward reducing them.

Referral

People often get confused when trying to model out referral. But it’s not that difficult — the key again is granularity.

Identify every opportunity to refer that makes sense. There are often more opportunities than simple having an “invite friends” screen. For example, lets say I have a collaborative todo list app. I could have an invite screen during onboarding. I could have another one from the dashboard. And I could have an “assign task” flow whenever a new todo is created. Again, take into consideration the nature of the customer relationship. I’m probably not going to send an email to my contacts suggesting they get a mortgage.

For each, document each step and make assumptions on conversion rate. On the invite screen, the user will send X number of invites on average. X% of users will see the invite and click through. X% of those will create an account, and x% will activate.

Add those new users to your total activated users. What you’ll quickly see is the power of referral. While it’s rare to have an app that’s truly “viral”, most apps can get some lift from referral. Which drives down your CAC.

Retention

You have to plan for churn. And for most products, churn is actually pretty steep. Look for baseline metrics for retention 1 month, 3 months, and 6 months later. And build that into your model in the form of cohorts.

Don’t be surprised when it looks ugly.

Once you’ve documented all this, step back and play with your model. You’re probably going to see that it’s going to take WAY more work than you thought to hit your numbers.

That’s okay. Kind of like getting your financial house in order, the first step is to confront reality. You’re not going to get on Techcrunch and be staggeringly successful. It’s going to be grind.

The growth model helps you understand what it’s going to take. It can help you and your team steel yourself for the slog that’s about to come.

Use it to manage your team

The real power of the growth model is as a management tool. You don’t want to create it once time and then ignore it. As you start executing, plug the results into your model. Adjust your assumptions as you start to capture real data.

Use the model to prioritize decision making. Whenever you’re discussing a new idea for a feature or product iteration, ask yourself which variable in the model is it going to impact and estimate what that impact will be. this can be extremely helpful in seeing which work will have the biggest impact if successful. It can help diffuse conflict around what to focus on (or salespeople who rely on anecdotal data, or a boss that’s typically the loudest person in the room.)

Use it to set realistic goals for your team. If you’ve decided to prioritize SEO, it’s helpful to show your team how much content they’re going to have to crank out and how many links they’re going to have to acquire to hit their numbers. It can help you be realistic while also creating urgency.

Get access to our growth model template.

There are few tools we’ve seen that have been more useful for product teams than a growth model. It sounds tedious, and the model will 100% be inaccurate. But the time it takes will be more than worth it in terms of setting expectations, prioritizing the right things, and measuring impact.

We’ve created a growth model template with detailed assumptions for a fictional product that you can customize to fit your needs. Click here to get immediate access.

Originally published at digintent.com.