Activation is the most mispriced lever in your growth model

Most growth models treat acquisition and retention as separate lines on the plan. Marketing is accountable for the first, customer success for the second, and the handoff between them is where the economics quietly break.

The first two weeks of a customer’s life are the highest-leverage, lowest-owned stage of the entire revenue model. Fix activation and you reprice every acquisition dollar you have ever spent. Ignore it and no amount of top-of-funnel efficiency will save the plan.

Early churn is an acquisition cost

Every customer who churns before they pay back their acquisition cost does not disappear from the P&L. Their cost is absorbed by the customers who stay. If a third of a cohort leaves in the first month, the real cost of acquiring each retained customer is roughly 1.5x the CAC on your dashboard. Nobody reports it that way, which is precisely why it persists.

At Lumen5, churn ran as high as 33% in bad months. That meant marketing could hit every acquisition target and the business would still underperform its plan, because a third of what we bought leaked out before it ever paid back. Treating activation as a growth problem rather than a support problem was the turning point.

The retention curve is decided in weeks, not quarters

Almost every subscription business shows the same shape: a steep early decline, then a long, flat tail. Customers who cross into the tail stay for years. Customers who do not, rarely return. The renewal conversation is where churn is recorded, not where it is decided.

Illustrative retention curves

The strategic implication is uncomfortable for most org charts. The moment that determines lifetime value sits between functions: after marketing has handed the customer over, before customer success has engaged, inside a product experience that product management is optimizing for feature adoption rather than retention.

Define activation precisely enough to fund it

“Engagement” cannot be budgeted, staffed or held accountable. A threshold can. The discipline I use follows the activation model popularized by Reforge, separating setup, the first moment of real value, and habit, and then working backwards from retained cohorts to find the behaviour that predicts staying.

The output must take the form X actions within Y days. At Lumen5 the data converged on five video downloads within seven days of purchase, which we named 5D7. Customers who reached it churned 86% less, month over month, than those who did not. More important than the number was what it did to the organization: for the first time, marketing, product and customer success were optimizing against the same, falsifiable target.

Give it one owner and a real budget

Activation fails as a shared responsibility. It needs a single accountable owner with authority to change onboarding, lifecycle messaging and in-product prompts, and an experimentation budget measured against the threshold rather than against activity.

Function	Contribution to activation	What they are measured on
Marketing	Sets expectations pre-sale; owns behaviour-triggered lifecycle	Share of new customers crossing the threshold
Product	Removes setup friction; builds prompts into the experience	Time to first value
Customer success	Intervenes when high-value accounts stall	Threshold attainment in priority accounts
Sales	Qualifies for fit and scopes the first outcome	Activation rate of closed-won cohorts

At Lumen5, the levers that moved 5D7 were not campaigns. They were behaviour-triggered: in-app prompts, reminders and layout tests fired by what a customer had or had not done, followed up by email and promotions, with every stall point fed back to product. Monthly recurring revenue grew from $80K to $455K over my tenure, and that growth rested on retention as much as on acquisition.

Revisit the definition as the business moves upmarket

An activation threshold is a hypothesis about your current customer mix. As the company sells to larger organizations, the behaviour that predicts retention shifts from an individual’s first outcome to a team’s adoption. A metric that was right at one stage becomes misleading at the next. I review the definition at least annually and whenever retention moves in a way the threshold does not explain.

The second-order effects most models miss

Activation does not only lower churn. It changes the shape of the entire revenue model, and that is why I argue it belongs in strategic planning rather than in an onboarding project.

It compresses CAC payback. Payback is calculated on retained revenue. Every point of early churn removed shortens payback for the whole cohort, which frees cash that can be reinvested in acquisition sooner. For a capital-efficient or bootstrapped business, that velocity is often worth more than the headline retention gain.

It raises the ceiling on paid acquisition. The maximum you can afford to pay for a customer is a function of what that customer is worth. Lift retention and the affordable CAC rises with it, which opens channels and audiences that were previously uneconomic. Companies that fix activation frequently discover that the paid channels they abandoned as “too expensive” were only too expensive for a leaky product.

It changes what marketing should promise. When the activation threshold is known, acquisition messaging can be written to attract customers who are likely to reach it and to set expectations about the first week. Targeting and positioning become instruments of retention, not just of volume.

It improves forecast quality. Activated cohorts behave predictably. A board that sees threshold attainment for this month’s new customers can forecast next quarter’s retention with far more confidence than one looking at logo counts.

Failure modes I have seen

Choosing a vanity threshold. Logins and page views are easy to move and weakly predictive. If the metric can be inflated without the customer getting value, it will be.

Mistaking correlation for a lever. Customers who would have succeeded anyway often take the action first. Without a controlled test, a team can spend a year pushing customers toward a behaviour that does not change outcomes.

Building the threshold and then not funding it. A definition without an owner and an experimentation budget becomes a slide. The organizations that benefit treat activation like a product line, with a roadmap, a backlog and a weekly review.

Ignoring setup. Teams obsess over the aha moment and neglect the steps before it. Every field, integration and decision required before value arrives is a point of loss. The highest-return work is often deleting steps, adding sensible defaults and showing value with sample data before the customer’s own data is connected.

How I sequence it in a new role

In the first month, I establish the retention definition with finance and build the cohort comparison. In the second, I identify the candidate threshold and run a controlled test on it. In the third, I assign a single owner, wire behaviour-triggered lifecycle and in-product prompts to the threshold, and add attainment to the weekly business review. From then on it is an operating rhythm: one or two experiments in flight at all times, measured against threshold attainment and against the retention of the cohorts that reach it.

Enterprise versus self-serve

The mechanics differ by motion, and leaders who run both should not use one threshold for each. In self-serve, activation is an individual’s first meaningful outcome, and the levers are product prompts, defaults and lifecycle messaging. In enterprise, activation is organizational: the right team members adopting, the first workflow running in production, the executive sponsor seeing the first result. The levers shift toward implementation scoping, success planning during the sale and executive check-ins in the first weeks. Lumen5 served both, with roughly 65% of the business coming from enterprise accounts and 35% from self-serve, and the work only became tractable once each motion had its own definition, owner and review. A blended activation rate across both would have hidden exactly the stalls that mattered most.

Questions for the board

• What is our activation threshold, and what share of last quarter’s new customers reached it?
• What is our effective CAC once first-month churn is included?
• Who owns activation, and what budget do they control?
• When did we last run a controlled experiment on the first week, and what did it change?

The takeaway

Activation is where acquisition spend either compounds or evaporates. Price early churn into your CAC, define the threshold that predicts survival, give it a single owner, and the growth model you present to the board will finally reflect how the business actually makes money.