How To Use And Optimize A Growth Model

A growth model does for resourcing what a financial model does for budgeting: one sheet showing how each lever feeds revenue, CAC and LTV, so you can see what moves what before you commit the quarter. There is a worked one-month model below that you can copy. Structure it around the sub-metrics you can actually affect, and you can see what a 10% change in one metric does to company growth, and trade efficiency for scale on purpose instead of by accident.

I'd argue that the two most important decisions you can make as a leader are 1) who to hire and 2) where to allocate resources. Having a growth model is key to number 2.

It's interesting how we put significantly less time and energy into our resource allocation decisions relative to our hiring decisions. This is because, well, it's not easy to measure and evaluate the resourcing tradeoffs across an organization. But it can be, using a growth model.

Issues with resourcing decisions today:

One of the most important questions companies should be asking themselves regularly -

What is the most effective allocation of resources to maximize our company's growth?

Most companies determine this every quarter by doing some version of an "executive round-table" debate supplemented by business cases. That approach has a couple important issues with it that are worth considering given how crucial these decisions are:

It is subject to "loudest voice in the room" risks where decisions are influenced by how strong of a pitch someone gives or how much weight in the company a person has. This approach is open to bias and opinions.
It is exacerbated by the fact that the business cases fail to provide a unified way to understand, quantifiably, the tradeoffs between resourcing different areas of the business. They are often created by department heads and subject to bias as it is. And oftentimes in the business cases, estimated impact isn't factored by the amount of resources needed either, so this makes apples to apples comparisons impossible.

A growth model is the answer to those issues.

What that actually looks like when it goes wrong

A third of a quarter's roadmap going to work nobody had ranked. Not bad ideas. Three teams each getting a little of what they wanted.

I watched this at a marketplace recently. Engineering capacity was tight, every team lead had something they needed staffed, and nobody wanted the conflict of saying no outright. So the quarter got sliced. A bit here, a bit there, everyone partially satisfied.

Losing the roadmap space is the cheap part. The expensive part is momentum, because that work runs on different cadences, needs constant cross-functional alignment, and drags the velocity of everything around it.

And the actual problem there was not prioritization. Every team had its own metrics and goals, nothing said which metric outranked which, and nothing said whether one team's goal had been approved for another team to spend its engineers on. When that is unclear, the strongest pitch wins instead of the strongest case. That is a popularity contest with a spreadsheet attached.

Defining a "Growth Model"

A growth model is similar to a financial model, except it is structured around growth levers and metrics which you have the ability to affect.
With how important resourcing decisions are, it is kind of crazy that many companies don't put in the time and diligence to create one in order to understand the relationship across growth levers.

Here are some of the most common growth levers. This differs business to business and B2C vs. B2B:

Common growth levers. The exact set differs business to business and B2C versus B2B.

How do you know what will be the best use of your company’s time and resources? Is there not a more important decision to be made than where you put your resources?

Structuring a growth model

A growth model, structured around your growth levers and metrics which you can affect, will help you understand the relationships and tradeoffs between them. It will also give you a shared framework and language to use and evaluate these tradeoffs. Here is an example of what a growth model structure could look like at a high-level:

Each lever has metrics you can move. Those roll up into company metrics that span every lever: CAC, LTV, and revenue.

A growth model helps zero-in on which sub-metrics you are planning to affect and how that translates to revenue for the company. They can also be used to validate assumptions and stress test baseline and upside scenarios.

The metrics within each growth lever are key metrics which you have the ability to affect.

The output metrics, or company growth metrics, have a volume metric (i.e. # acquired) and a health metric (revenue / # of visitors). Also, you will notice that there truly is a relation across all of these growth levers - Revenue, and Customer Acquisition Cost / Lifetime Value.

A worked example: one month on one sheet

Here is the simplest version of the whole thing, deliberately generic, in the shape a real one takes. One month, four blocks, and every number below the first block is derived rather than typed in.

Block 1. The money and what it buys.

Input	This month
Paid spend	$200,000
Cost per install	$4.00
Installs	50,000

Spend and cost per install are the two you control directly. Everything below here is a rate, not a decision, and that distinction is most of the value of laying it out this way.

Block 2. The rates you can move.

Rate	Out
Install to trial, 12%	6,000
Trial to paid, 35%	2,100

This block is product work, not budget.

Block 3. What a customer is worth.

Input	Value
Revenue per payer	$12 / mo
Monthly churn, 8%	12.5 months
Lifetime value	$150

Churn is doing the work here, not price. At 8% a month the average life is 12.5 months. Halve the churn and the lifetime doubles.

Block 4. What comes out.

CAC $95. LTV $150. LTV:CAC 1.6. Payback 7.9 months.

One number is what the whole sheet exists to produce, and 1.6 is thin.

Now change one input. Take trial to paid from 35% to 45% and you get 2,700 payers instead of 2,100. CAC falls to $74, LTV:CAC goes to 2.0, and you spent nothing extra.

That is the answer the sheet gives you and the round-table does not: the cheapest route from 1.6 to 2.0 runs through block 2, not through more spend. You cannot see that by arguing about it, and you cannot see it from a business case that estimated impact without dividing by the resources needed.

The output of a good growth model

With a growth model, you are able to understand what a 10% increase in a sub-metric means for Revenue and CAC/LTV.

The model makes the health-versus-growth tradeoff explicit. You can trade efficiency for scale on purpose, instead of assuming you can maximize both at once.

So, when you have a growth model that calculates changes in one metric in one growth lever of your business, you will be able to see the overall impact to your company growth level metrics, which span across growth levers and allow you to make these decisions.

One thing I've pointed out in the diagram above is the concept of health and growth metrics which is something that should be called out throughout a growth model. You can trade health for growth in many cases depending on your business priorities. The clearest example of this is paid marketing - where you can trade efficiency for scale. However, if you don't have a growth model in place and concept of health vs. growth metrics, an unrealistic expectation can arise around achieving both simultaneously.

Another benefit of a growth model is that it forces you to talk about which sub-metrics you believe you can change within a growth lever and what your hypotheses are to change them, whereas, with business cases this level of detail is not typically achieved, and there are sweeping assumptions made to guestimate what the net revenue impact will be of an initiative.

But perhaps the most underrated benefit of a growth model is that it provides a shared language and basis for discussion when it comes to growth planning and resourcing which typically tends to be segregated by department rather than unified by a shared company growth model.

What to put around it

The model is the instrument. It still needs something to point at, or you are back in the round-table with better arithmetic.

Three to five Tier 1 metrics for the whole company. Not per team. Company.
Put them in priority order. Everyone skips this, and it is the part that makes a trade-off conversation possible when two of them pull against each other.
Decide in quarterly planning which teams support which metrics. Explicitly, on the record.
Fund secondary metrics out of slack time, roughly 10 to 15%, rather than out of the committed roadmap.

Then every incoming request answers four questions. What metric does this move? Is that metric Tier 1? Has the exec team already aligned on it? And do we want this team carrying it this quarter?

A growth model is what makes the first of those answerable with a number instead of a pitch.

Common questions

What is a growth model?

A model, structured like a financial model, but built around the growth levers and sub-metrics you can actually affect. It shows how a change in one metric, say a 10% lift in conversion, flows through to company-level outcomes: revenue, CAC, and LTV.

Why use a growth model instead of business cases?

Because business cases give you no way to compare resourcing tradeoffs apples-to-apples. They're built department by department, biased toward whoever writes them, and usually don't factor impact by the resources required. A growth model puts every lever in one quantified frame, so the decision turns on the numbers instead of the loudest voice in the room.

What's the difference between health and growth metrics?

Health metrics measure efficiency: revenue per visitor, CAC, LTV. Growth (or volume) metrics measure scale: visitors, users acquired, users retained. You can often trade one for the other, paid marketing being the clearest example. A growth model makes that tradeoff explicit so nobody assumes you can maximize both at the same time.