PLG vs Sales-Led in 2025: The Honest Trade-Offs

I want to start with a confession: for most of 2019–2022, I told almost every founder I worked with to go product-led. Free trial, freemium, self-serve onboarding, viral loops. It looked like infinite leverage. Customers signed themselves up, paid themselves, and expanded themselves. What's not to love?

Then 2023 happened. Capital tightened. ACVs came under pressure. Buyers got more skeptical. And a lot of the PLG playbooks that worked in the ZIRP era stopped working — quietly at first, then loudly. By 2025, the conversation has matured: PLG isn't dead, but "PLG by default" is.

Here's the framework I now use, and the trade-offs I wish I'd been clearer about three years ago.

When PLG still wins decisively

PLG still beats sales-led when all four of these are true:

• The product solves a problem an individual user can experience and value within ten minutes, alone, without needing approval or integration help.
• The buying decision involves one or two people, not a procurement committee.
• The data the user shares to get value isn't sensitive enough to require a security review.
• The annual contract value is below roughly $5,000.

Most of the canonical PLG winners — Notion, Figma in its early years, Linear, Loom — fit this pattern almost perfectly. If your product fits all four, going PLG is still the right call.

When sales-led quietly wins

Sales-led wins when any of those four conditions invert. In particular:

• Multi-stakeholder buying. The moment three or more humans need to agree, self-serve checkout becomes a friction generator, not a friction remover.
• Data sensitivity. If the user has to share PII, financials, or anything regulated to get value, you're going to need a security questionnaire on the first call. PLG can't carry that.
• High ACV. Above roughly $25K, the math on self-serve breaks down. The customer expects a human, and the unit economics support one.
• Long time-to-value. If the user has to integrate, configure, or invite a team before they feel value, the PLG funnel will leak before they get there.

The hybrid is the answer more often than people admit

The model I see working most often in 2025 isn't pure PLG or pure sales-led — it's sales-assisted PLG. Self-serve sign-up is the front door. Inside the product, an AI or a human picks up the highest-intent accounts and gives them a white-glove path to expansion.

A pattern that consistently works:

1. Free or low-friction trial for the individual user.
2. Product instrumentation that scores accounts based on usage, team size, and integration depth.
3. A small AE team that only touches the top 5–10% of accounts, focused on multi-seat and enterprise expansion.
4. A self-serve checkout that handles the long tail without human involvement.

Done well, this gives you the CAC efficiency of PLG and the ACV expansion of sales-led without doubling your go-to-market cost.

The mistake we see most often

It's almost always the same one: forcing a model that doesn't match the buyer. Founders pick PLG because it's culturally fashionable in their network, even though their buyer is a CFO. Or they pick sales-led because their last company did, even though their new product can deliver value in five minutes to a single user.

Match the motion to the product, not to fashion. And revisit the choice every twelve months — because the product, the market, and the buyer all change faster than your GTM model does.