What Product Led Growth Actually Requires
Product led growth means the product itself does the work of acquiring, converting, and expanding customers, rather than a salesperson. It requires four things: value delivered fast without help, self serve onboarding that survives a real user, usage that grows with the value received, and a purchase small enough to make alone.
The definition, and what it excludes
Product led growth means the product performs the functions a sales organisation would otherwise perform. A user finds it, tries it, understands the value, decides to pay, and expands their usage, without a person from your company involved in any of those steps.
That is the whole idea, and it is more demanding than it sounds, because every one of those steps has to work without a human patching over the gaps.
Which is why I am blunt about what does not count. Adding a free tier to a sales led product is not product led growth. It is a discount with extra infrastructure. If the free user still needs an onboarding call to get value, a demo to understand what it does, and a contract to pay you, the sales motion is intact and you have simply given away the first part of it.
The test is not whether there is a free tier. The test is whether a user can get all the way from arriving to paying to expanding on their own. If any step requires a person, that step is your actual growth model, and everything before it is a lead generation mechanism. Which is fine, and it should be called what it is, because the two are measured and staffed completely differently.
The four preconditions
Fast time to value. The user has to experience the thing the product does before they lose interest. Not understand it, experience it. If your product requires importing data, connecting systems, or configuring rules before it does anything, you have a gap that a salesperson currently fills, and removing the salesperson does not remove the gap.
This is the precondition most often missing and the hardest to fix, because it usually means changing the product rather than the funnel.
Self serve onboarding that survives contact with a real user. Not a documented happy path. A flow that works for someone impatient, on a bad connection, with messy data, who did not read the tooltip. Most onboarding is designed by people who already know what the product does, which is exactly the wrong qualification.
A usage metric that tracks value. Something that naturally increases as the customer gets more out of the product: records stored, jobs run, seats active, messages sent. Expansion revenue needs something to attach to, and if the only lever is a tier upgrade decided by a human once a year, expansion is a sales activity again.
A purchase decision one person can make. If the buyer needs three approvals, a security review, and a procurement process, no amount of product polish gets them through it alone. That is not a product failure, it is the shape of the market.
Missing one of these is workable. Missing two means the model is being applied to a product that does not support it, and the result is a lot of activity that never converts.
Where it reliably struggles
Products with high switching costs, where the trial cannot demonstrate value without a migration. Regulated environments where evaluation runs through compliance rather than through use. Anything sold to a committee, since committees do not self serve. And products that are genuinely complex, where the honest onboarding is a person explaining it. Recognizing this early saves a year of trying to fix a funnel that was never the problem.
The hybrid is usually the answer
In practice most companies that describe themselves as product led are running two motions at once, and I think the framing causes more confusion than the reality warrants.
The product acquires, activates, and converts small customers entirely on its own. Then usage data identifies the accounts where many users have appeared inside one organisation, and a person reaches out to that specific account with a specific reason.
That is not a compromise or a failed PLG implementation. It is the model working, because the product has done the expensive part, qualifying interest at scale, and the human is applied only where the deal size justifies it.
The mistake I see is teams treating any sales involvement as evidence the strategy failed, then abandoning a motion that was generating exactly the pipeline they wanted. The measure is not the absence of salespeople. It is whether the product is doing the qualification work, so the people are spending time only on accounts that already demonstrated value.
If you want one practical starting point: instrument the moment a user first gets real value, then look at what percentage of signups reach it and how long it takes. Almost every product led problem I have looked at was visible in that single number, and almost nobody had it.
Frequently asked questions
- What does product led growth mean?
- The product itself does the work a sales organisation would otherwise do: a user finds it, tries it, understands the value, decides to pay, and expands usage without a person from the company involved in any step. It is a distribution model rather than a pricing or packaging decision.
- Is a free tier the same as product led growth?
- No. Adding a free tier to a sales led product is a discount with extra infrastructure. If the free user still needs an onboarding call to get value, a demo to understand the product, and a contract to pay, the sales motion is intact and you have given away its first part.
- When does product led growth not work?
- Where switching costs are high enough that a trial cannot show value without migration, in regulated environments where evaluation runs through compliance rather than use, where the buyer is a committee, and where the product is genuinely complex enough that honest onboarding requires a person explaining it.
- What should I measure first?
- The moment a user first receives real value, then what percentage of signups reach it and how long it takes. Almost every product led problem is visible in that single number, and most teams do not have it instrumented, which makes every downstream funnel debate speculative.