Optimizing Your Growth Funnel

In software you get used to problems that converge to a single, verifiable answer. Growth is a different beast: there are infinite activities that will all help attract, convert, and retain loyal customers. You’d go insane trying them all–not to mention exhausting whatever runway (and patience) you have to work with.

The trick is to structure growth efforts in a way that transforms a fundamentally unbounded problem into a process that can be measured and optimized. In the simplest formulation, we can do that by introducing the concepts of a customer persona and a funnel: the person discovering and adopting your product, and the journey they take to get there.

Defining a Growth funnel

Your prospective customers don’t know it yet, but you’re about to formalize their buying journey by putting them in the fuzzy-but-recognizable state machine that is your growth funnel.

There’s no need to actually implement this state machine in code (any self-respecting CRM should have done this already), but spending some time documenting the states, key metrics, and the exit criteria percentage between funnel stages will turn customer acquisition into a string of focused optimization problems.

A sensible default

If you’re drawing up your funnel for the first time, the AAARRR framework introduced by Dave McClure is a good place to start. It’s an informal standard for B2C SaaS, and unless you have a strong reason to model your funnel differently you might as well borrow prior art.

Framed in AAARRR terms, your funnel might look something like this:

You likely noticed the two different scales of measurement possible in the funnel. KPIs measure the performance internal to a stage, while conversion rates measure transitions between stages.

1. Use conversion rates to understand high-level performance and direct energy to underperforming stages

2. Use KPIs to track and address issues within a given stage

Recall the exit criteria defined for each stage of the funnel. If a customer makes the leap from “acquired” to “activated” by completing an onboarding wizard inside the application, the conversion rate between stages can be measured by dividing the number of users who complete the wizard by the number of sign-ups. This number is dimensionless; and can be multiplied by the conversion percentages of other funnel stages to determine the efficiency of the funnel as a whole.

Defining a funnel and its key metrics is the single most important step in turning growth into an optimization problem. We now have a basic outline of the journey customers take when deciding to evaluate and adopt the product, plus some bare-bones telemetry to see how they’re doing. Like everything, though, the definition is simply a sensible default; as you learn more about your customers and your business, there’s a very good chance the funnel will evolve as well.

Developing the funnel

With a funnel and framework for experimentation in hand, we’re on to the fun part.

Before we get into it, a word about the superpowers that you, as a software developer, bring to Growth. Just because you can automate any step in the funnel doesn’t mean you should.

Automating the wrong thing will simply slow down the pace of experimentation and learning, a potentially-fatal mistake, particular in the early stages of the startup journey. Make it work, make it right, make it fast—just as in software, there will be time for iteration and automation once you know what you need to do.

With that, try to pretend that you’re approaching growth with no technical tools at your disposal. If you’re working on a startup, this likely reflects near-term reality: whoever will take over the growth function in time may not be as comfortable interacting with parquet files, tweaking TypeScript CLIs, or whatever it is you’re tempted to reach for first. And no-code platforms like Zapier, PowerBI, and even Google AppSheets can save time with turnkey integrations to popular growth-related tooling.

Overall, though, expect the state of automation in growth tooling to be jankier than you’re used to: if you’ve been riding around in the Volkwagen Golf that is your day job (comes with git, a reliable CI suite, and all the trappings!) you’re about to jump into a Trabbi. Buckle up.

But the important thing is that the engine works. The analysis going into your growth funnel requires less integrity or durability than the data in your primary datastore, and as long as we’re learning and the site is up, it works.

Optimizing the funnel

Before putting the entire funnel under the microscope, remember that it’s a model—and only one model—of progress towards headline goals like revenue and user growth. Just as your customer persona is shorthand for many different individuals, the funnel is simply a convenient way to understand and summarize all the activities adding up to traction and growth. When it no longer fits your business, change it.

If the current model fits, however, it follows that improving funnel performance should also be good for the business. But how? Just like in product development, many different activities can create value for the user business. You can buy more ads; optimize their click-through rates; send more emails; write a new whitepaper; reach out to press outlets; create new landing pages; and obviously build new features. The possibilities are endless.

The prospect of funnel-wide optimization is daunting, too. The good news is that a focused, stage-wide approach with more tangible short-term results will also generate business impact over time. If your funnel is a reasonable approximation of your business, it follows that improvements to KPIs within any given stage should improve conversion efficiency across multiple stages and the funnel overall.

1. keep a list of opportunities for improvement
2. (re-)prioritize them regularly
3. test a single hypothesis for improving a high-priority gap in a single stage
4. goto 1

Test one solution to one problem in one stage, repeat. Rather than get bogged down in the breadth of the funnel or all of the things you could do, taking an incremental, stage-wise process will help you stay sane while still delivering cumulative results.

Finding what works

When infinite possibilities collide with limited hours in the day, something has to give. Once again, the solution comes from the familiar realm of product development: prioritize. Save your energy for the highest-return opportunities in the funnel—typically (though not always) the earliest or leakiest stage.

You can determine “leakiness” by calculating the conversion percentage between funnel stages. For example, for every “acquired” customer, how many were “activated”? If the conversion rate is low, understanding why customers are signing up but not actually onboarding inside the product would be a very valuable activity.

What you don’t do enables what you do. By understanding your funnel and putting energy where it will have the most impact, you’ll save time for everything else going on in the business.

Ultimately, growth comes down to finding what works. If a product aimed at business customers shows signs of traction with consumers instead, put in some energy and see how far it will go. The biggest factor is to try something, then try something else, then do it over and over and over until you find what sticks.

And don’t stop here! The Guide to Growth includes tactics for optimizing every stage of your funnel. It’s important stuff, totally free, and I hope you’ll check it out.