A free tier is not a growth strategy. It is a bet that your product can do the persuading a salesperson used to do, and most teams take that bet without building anything that tells them whether the product is winning.
I learned the size of that gap building AdRoom at Pixis, an AI creative product that went from $0 to $10M ARR in five months with a 20-person team. The team that hit that number did not skip the hard parts of product-led growth. It built the machinery a slogan skips over: an activation event that meant something, a sales handoff that respected the signal, and a data platform reliable enough that self-serve marketers could trust it unsupervised.
So that’s today’s post.
Today’s Post
Why the PLG slogan is easy and the PLG operating model is not, in six sections.
- The vanity metric every PLG team reaches for first
- The activation event AdRoom picked instead
- Why a free user is not sales pipeline by default
- The instrumentation bar goes up, not down, under self-serve
- A handoff model, not a handoff email
- What PLG costs before it pays for itself
1. The vanity metric every PLG team reaches for first
Most PLG teams start the same way. A signup counter climbs, and leadership calls it growth.
Signup: an email address and a password. Nothing about a signup tells you a user got value, came back, or would ever pay for the thing. Weekly signup counts feel like progress because they are cheap to chart and easy to defend in a Monday meeting.
The credibility problem shows up the moment someone asks a harder question: which signups matter? If product, sales, and marketing cannot answer that in one sentence they all agree on, the team is running a top-of-funnel motion with a PLG label taped on, not a PLG motion.
2. The activation event AdRoom picked instead
Activation: the moment a user experiences the value you promised, defined narrowly enough that an engineer can instrument it.
At Pixis, every new AdRoom brand went through a two-week onboarding sprint with one hard target: get the brand self-serving live, on-brand ads inside that window. We instrumented the published ad, not the login. Four automotive and consumer-tech brands into that sprint, it had already produced $800K ARR in two months, on the way to $0 to $10M ARR in five months with a 20-person team.
Had the team counted logins instead, it would have declared the sprint a win in week one and learned nothing about whether the product actually replaced the designer ticket it was built to replace.
Grade a launch review that leads with signups a 3/10. It measured attention, not value.
3. Why a free user is not sales pipeline by default
Free users look like pipeline to a sales team carrying a quota. Every signup becomes a lead, and reps start calling before the product has finished making its case.
That interruption has a real cost. A self-serve user who gets an outreach email in week one starts treating the product like a lead form instead of a tool. Reps burn cycles too, chasing accounts that never had budget for an enterprise deal in the first place.
Product-qualified signal: a defined behavior that predicts revenue, something more specific than “created an account.” At Pixis, the event we watched for AdRoom was the same one behind the $800K number: a published on-brand ad, not a signup. A brand that published inside the sprint window was a fundamentally different lead than a brand that signed up and went quiet.
Sales does not disappear under PLG. It changes when it enters, waiting for a signal the product already proved instead of a signup the marketing team just bought.
4. The instrumentation bar goes up, not down, under self-serve
In a sales-led motion, a rep can talk a confused buyer through the gap on a call. In a self-serve motion, the product has to carry that persuasion alone, usually in minutes, usually with no second chance.
My first mandate at Pixis was not the AI layer marketers saw. It was the data platform underneath, with a 12-person pod and one target: cut sync delay from 4 hours to under 15 minutes. We hit that across 1M+ keywords a day for 400+ enterprise customers, at 99.6% end-to-end reliability. None of that demoed well. All of it is why the recommendations riding on top of it were worth acting on.
The same discipline runs the other direction at InMobi, where the system of record under a demand-side platform and exchange measures 200M+ devices’ worth of ad events a day, and the month-end revenue close on that data now runs in 12 hours instead of the 5 days it used to take. A team that cannot close its own books on time will not convince a self-serve buyer the product is trustworthy either.
5. A handoff model, not a handoff email
The mature version of PLG does not delete sales. It sequences it.
Self-serve is the default path, because it is the fastest way for the right user to learn the truth about the product without waiting on a rep’s calendar. Sales-assist enters only after a product-qualified signal fires, the kind of signal a team like AdRoom’s watched for instead of a signup. Enterprise sales enters last, when procurement, security review, or integration complexity make a self-serve checkout impractical on its own.
Get the order wrong and you get the worst of both failure modes at once: a self-serve experience interrupted by outreach nobody asked for, and a sales team that learns to hunt signups instead of qualifying real demand.
6. What PLG costs before it pays for itself
The popular version of PLG promises a cheaper go-to-market. The honest version costs more up front, in product investment, before it gets cheaper.
Prism, the AI co-pilot I led at Pixis with a 15-person team, only lifted trial-to-paid by 25% and cut onboarding time by 40% after we shipped interactive demo accounts, an ROI savings calculator, and guided walkthroughs, then treated the whole funnel as an experiment. None of that was cheap, and none of it was optional once the trial itself became the pitch a human used to make.
The discipline that made the roadmap trustworthy was saying no. Fifteen-plus voice-of-customer interviews on that launch turned into exactly five roadmap features, not fifty, because declining the other forty was the actual work.
My own rule from this site’s home page applies here too: unwritten means undecided. A PLG motion without a written activation definition is a motion nobody has actually decided to run.
Takeaways
Six panels, one per section. The callout is the bet every PLG motion is actually making, whether the team wrote it down or not.
Final Words
Most teams will keep treating PLG as a pricing-page change: add a free tier, open self-serve signup, and wait. The few that win will write down what activation means before they ship, instrument that event instead of a login, and sequence sales around the signal instead of the signup.
Pick the free tier your team already shipped, write the activation definition it should have had, and check whether anyone can name the number this week.
Cheers, Hoshank