Before and after product-market fit

Until fit is real, the company has one job: learn what to build for whom. After the threshold, the job flips to keeping up with demand without wrecking unit economics. Doing the second job during the first one is premature scaling.

Fit is also a spectrum and a moving target. "We have PMF" is a hypothesis you keep testing.

The big idea

Before fit, money and calendar time are scarce. Every hire multiplies coordination cost. Heavy paid acquisition papers over a leaky bucket. Fancy offices and premature departments are company theater.

After fit, you earn the right to invest in repeatable go-to-market, specialists, infrastructure, and efficiency metrics. You do not earn the right to stop discovering.

The analogy of a gate that swings one way breaks: you can have fit in a beachhead and lose it chasing a larger, colder market. A competitor, a platform policy change, or a new technology can erase yesterday's fit.

Walking through it

Pre-fit priorities

• Talk to users constantly (see user research)
• Ship small, measure retention, iterate
• Keep the team tiny
• Avoid heavy paid acquisition
• Avoid company theater

Pre-fit, the best engineering work often looks unimpressive: instrumentation, rapid experiments, deleting features, sitting on support. Optimize for cycle time to learning, not architectural purity. You are still in search even if LinkedIn says Series A.

After the threshold, invest in

• Repeatable GTM — channels and a sales process that work without founders on every call (go-to-market)
• Hiring — specialists, without importing big-company process too early (hiring)
• Infrastructure and margin — the product must stay up; unit economics must work at volume
• Efficiency tracking — burn multiple, magic number, rule of 40 (efficiency metrics)

A practical post-fit test: can a newly hired AE or a self-serve funnel produce pipeline without the CEO rewriting every email? If not, you may have founder-market fit more than product-market fit.

Keep a lightweight Ellis pulse or cohort review when you expand ICP, raise prices, or ship a second product.

Common mistakes

• Hiring a sales army before founders can sell repeatedly. That is premature scaling.
• Declaring search over. Expansion into a new product line restarts search.
• Importing big-company process the week after the inflection. Specialists yes; theater no.
• Chasing a colder, larger market and losing the beachhead that actually pulled.

Check yourself

Retention is flat, Ellis is 22%, and the plan is a national ad campaign. What should happen first?

Stay small, talk to users, and find a segment that pulls. Ads will paper over a leaky bucket.

What is a cheap test that you have a machine, not just founder-market fit?

A new AE or a self-serve funnel produces pipeline without the CEO on every email.

Key takeaways

• Pre-fit: learn, stay small, measure retention. Do not scale spend.
• Post-fit: repeatable GTM, careful hiring, infrastructure, efficiency.
• Fit is segment-specific, non-binary, and perishable. Keep measuring.
• Engineering flips from learning speed to leverage and margin.

Further learning

The only thing that matters (Marc Andreessen)An archive of the best articles from Marc Andreessen’s now defunct blogpmarchive.com

How companies die by scaling too early: why startups fail.