You're doing "Activation" wrong

We managed to mess up a great, simple metric for good.

I was afraid that my knowledge about product-led growth would become commoditized and useless because AI would just fix the most egregious mistakes in a typical SaaS business automatically.

Luckily (for me at least), I was wrong; if anything, it got even more complex to get a PLG business under control.

At Fyxer as a CPGO, I’m spending dozens, if not hundreds, of hours curating specific high-level documentation from Strategy to ICP, use cases, etc., so my trusty AI sidekick always has the most accurate thing to go off on, and it still is not that easy to actually correctly identify where the problems are.

I’ve also said many times that distribution (finding enough good users and activating them) is the hard part of the product-market fit equation, not just having a good product.

But as it turns out, activating the right users is not (just) a growth problem. It’s a company problem, and AI is really good at generalizing for some users and bringing more of them, not the right ones.

I’m not sure if this is all caused only by AI, but our understanding of what good looks like in Product-led Growth has changed dramatically.

“Activation”

The most prominent concept about product-led growth is probably around your activation metric, and we used it for quite some time now to get all of it under control:

In short, we’re introducing Aha/activation milestones to a company to focus more on customer success and to have a conversion rate to focus on:

The most common definition that you will still find out there is:

“activation rate = [users who hit your activation milestone] / [users who completed your signup flow].”

The activation rate is at least in principle so powerful because it informs you about the most important thing in a SaaS business. Your efficiency. If the activation rate is high, your business will do fine. If you’re efficient, you can grow.

That’s technically correct, but the reality is that most of my clients, and we at Fyxer, struggle to push it up and are not able to move it, and it masks short-term efficiency for long-term efficiency.

If you think about the simplified journey of a customer, then it’s this:

• A user has an underserved need about something that is important to them

• A promise is made somewhere that makes them think “this could be solving my issue”. They have not yet experienced the product, but they form a mental model of what it could be. (Passive: Marketing, Sales pitches, Word of Mouth)

• If they are interested enough, they switch to evaluate any solution. They are intent on evaluating the service to see whether a further time/money investment is worth it. They actively visit your website, engage with sales etc.

• The user completes a sign-up flow (typically creating an account)

• They’re now being “onboarded” onto their first experience through an onboarding flow if there is one, or they’re left to their own devices to figure it out.

• This is typically the time where a limited trial is running that allows them to evaluate the solution more in depth, and they are being “activated”. They get to experience the full product on a continuous basis.

• They pay for the solution and keep using it until they churn voluntarily or involuntarily. Typically, the payment happens after the activation.

A typical cross-functional growth team focuses on the onboarding part. Their mission is defined as “connecting users as efficiently as possible to product value”. Historically, we thought that the best, simple metric to do this is the activation rate as defined above and put a growth team on it.

If you don’t have a growth team at all, then yes, this is a good thing to do.

But it masks a couple of issues that have become bigger in the past years:

Issue 1: starting at “users who completed your signup flow”

It does not measure what happens before a user creates an account. It doesn’t question at all whether you should “have” an account before you expose the product. I’ve been writing about this a lot in the context of Interactive demos, which are essentially shifting customer value before any account creation, and the most extreme example I can think of was from Rows, which I called an ungated freemium: no account at all to use the product until you absolutely need it, for a complex Excel competitor.

The biggest shift in the past years, though, is how complex products have become to explain despite the simple jobs they do and how much friction it takes for them to be successfully integrated into their life.

Wispr and their high-friction signup

Wispr (Voice dictation) is a simple voice dictation hotkey in your system that you can call anywhere, and despite its high friction to get it running, it's extremely successful. They force you to download a 300mb application, install it, and then also learn a hotkey to use it before anything else.

No product for you until you download the application

It takes quite a lot to get going, and I’m pretty sure they have one of the best activation metrics that you can have currently. But not despite their high friction, but because of it.

What’s happening before the users start their onboarding is enabling their growth team to have a first-class onboarding and product experience.

The way that I’m getting around this is to include the signup flow in the activation metric and start when a customer clicks on the first button that requires them to do something (typically after the homepage/pricing page or when entering an interactive demo):

“Leah’s activation rate = [users who hit your activation milestone] / [users who started interacting (before signup)].”

This also gives your teams the freedom to rearrange “where” and how signup happens (a huge topic for us at Fyxer where we require emails and calendars to be connected before someone can use the product).

While this is, in my experience, a good solution to deal with that problem, it starts to hint at an even bigger problem: this is not a Growth team metric anymore.

Issue 2: Activation rate is a company job masking as a Growth team metric.

Yes, thinking about the most optimal way to activate someone is still the right thing to do, and it’s worth it to have great people on your onboarding journeys, including the sign-up steps.

It stems from the romantic notion that “the customer is always right” and we make it right for them no matter the cost. But should we?

Onboarding, or “what” and how users experience your product during a trial, is the most obvious lever to optimize your activation rate.

But it’s not the biggest lever. The biggest lever has become how close to the promise you delivered is to the actual product experience and how long you can keep that promise. And that starts way before your users land in onboarding.

It’s the marketing message, the brand, the word of mouth (good and bad), the things a user will ingest before they even land on your website.

Think about it: you can optimize your experience in the best way possible to serve fine dining food with no objective flaws, but if you keep bringing customers in that want fast food, they simply won’t activate.

You have two choices: you restructure to serve fast food, which optimizes the activation metrics, but completely wrecks the amount of money you make per account, in which case you adjusted the onboarding experience to the marketing message.

Are you now efficient now? Not really; you just shifted the problem downstream. If your kitchen (the core product) was built to serve fine dining food and your restaurant is now full of fast food clients, they will still churn further down the line.

This is exactly what’s happening with AI: AI is really good at making a great first impression, but longer term your users are disappointed if they are not the right ones:

• Your messages look more polished

• Your onboarding flows have fewer objective interaction design problems

• Your payment and cancellation flows are easier to understand and use

• Your pricing pages are easier to read and understand.

Essentially, your acquisition metrics start to look better and better while your company rots away underneath and everyone points at the market and the product.

This short-term churn after having paid is rampant in a lot of datarooms that I’ve looked into. And it’s happening everywhere in SaaS right now. It’s not that the products are crap; it’s just that Marketing and Growth never talk to Product functions and vice versa. They do not “really” understand their customers the same way.

They build and advertise for different customers.

You can test this for yourself; go check your data to see whether the % of 1st-month churn after paying has increased over time. Most people think it’s the market shifting around them, but I suspect it’s their Marketing vs. Growth going completely off vs what Product is building over time.

Our cross-functional product teams with PM’s, designers, and engineers are not equipped to deal with this. Product Marketing teams are in a bit of a better perspective, but they commonly don’t deal with the root cause of this problem, which is measuring how close the promise (messaging, positioning, tone) of marketing/sales is to the experience of the customer when they experience the product.

How can you measure that? It’s the long-term activation metric without the onboarding improvements over time. And that metric can’t be pushed by a growth team; it’s the rate of correct customers being brought in for a specific amount of money that stay for at least 60 days.

Here’s what I use often:

Marketing / Sales long-term Activation: $ spent per user/account that didn’t churn at month 2.

Growth long-term Activation Rate: Still the same; if you’re big enough, split it by ICPs or use cases: [users who hit your activation milestone] / [users who started interacting (before signup)] grouped by ICP / use case.

These metrics work, but they are now hard to influence fast.

This leads to Issue 3:

Issue 3: “Activation” is not granular enough

If you use activation metrics like this, you’ll notice that they are great to measure performance but become hard to read out fast enough and react, exactly when AI made everything faster, and shipping is cheap.

It’s just not that obvious anymore how long your trial should be, where you should charge, and when you force anyone to sign up.

Thankfully, users are still going through different steps which are predictors of success that are universal, that don’t limit us.

Before a user is long-term “activated,” they go through different states, I use these explicit definitions:

Setup: Starting the signup flow or interacting with your product experience in the form of low-friction demos. (For most companies, this is the user clicking on a CTA on their homepage like “Start Trial” or “Show me more” or “Download Client”, it could even be visiting the pricing page in some cases)

Aha Moment: The first time a user has experienced the product (usually day one). They are now capable of evaluating whether this is what they expected initially (They are not necessarily loving it!)

Activation: Continued use of the product within the trial period (or freemium) but not yet formed a habit.

Habit: Has developed a strong habit that can be observed over at least 8 weeks with the product

Aha Moment as a separation tool

The important bit here is that you’re using the Aha moment to inform users as well as possible. Your goal is to inform them honestly, not to convince them your stuff is amazing.

You want to move them closer to the experience from the promise.

Interview those that never reach it:

• Did they not reach it because onboarding was confusing? Growth team problem

• Did they not reach it because it’s not relevant for them, too expensive? Marketing / Sales messaging problem, wrong customers
  

Activation Moment as a qualifier

Same for the above activation metric: interview those that never reach it but have reached the aha moment:

• Did they never activate because the product wasn't needed that much? Marketing / Sales targeting problem, or onboarded to the wrong part of the product.

• Did they never activate because it was hard to use / cumbersome? Growth team problem.
  

Habit Moment as a qualifier

Here’s where this becomes spicy. Users that have reached your activation moment but never developed a habit at this point are commonly

• Product didn’t stick despite no stability/quality problems: Wrong customers again; they commonly churn when the product doesn’t serve their recurring use case at scale (lack of batch features, unnecessary clicks for high-frequency use cases). At this point, it could be the Market running away from you or the strategy being out of date.

• Stability / Quality problems: Correct customers but the product changes too much or is simply not stable enough. Some products cannot afford to make their customers look bad just once. (A failed recording for Wispr, for instance)
  

Honorable mention: Delight

Some companies also use a “delight” moment, where you try to measure how you can impress users beyond their expectations, but in my experience this can happen in any of the above steps and is a topic for a different day.

Think of it this way: if you can’t stabilize the conversion rates from Setup to Aha to Activation to Habit moments, then overfocussing on magical acquisition moments is just creating more of the same problem:

• Short-term churn that masks as good activation metrics.

Creating delight falls into the same category; it’s only good delight if it leads to long-term activation, and in my experience, delighting customers that have reached habit is by far and away more valuable than wowing customers in their aha-moment.

Summary

To really deal with this problem means to rewrite the Product-led Growth book to some degree. It’s easy to lob it all into the same category by saying, “Well, it’s always been about activating users”, but by doing

As the cost for shipping has decreased and judgment on what is good has become more important, we’re seeing that “activating” users is much more than just moving them between two goal posts.

Tailoring them to your company is going to make the difference between a company that is successful and one that is great at creating a good first impression and then deflates like a hot air balloon.