Product-Market Fit: What Design Can Do to Get There

What product-market fit is, how to measure it with Superhuman's 'very disappointed' question, and what design can do to move closer to that point.

info Quick Definition

Product-market fit is the point at which a product clearly satisfies a market that needs it. Design does not decide it on its own, but it can move a product closer to or further from that point, especially in how users discover and keep getting value from it.

What is it, and where does the idea come from?

The most cited definition comes from Marc Andreessen, in a June 2007 essay: “Product/market fit means being in a good market with a product that can satisfy that market.” In the same essay, Andreessen credits Andy Rachleff with helping him crystallize the formulation.

The idea is simple. Before fit, the team pushes the product to users. After fit, users pull the product toward other users, and growth stops depending on each individual sale.

Why does this matter for a junior designer?

Because the early stages of a product feel like a race for fit, and that changes which design work gets priority. In that phase, polishing the interface matters less than learning which segment needs it and what makes people come back. A junior who can tell “the user uses the product” apart from “the user needs it” can propose more useful experiments.

Also, many teams mistake a lack of fit for a screen problem. Redesigning a flow nobody wants to complete does not change the fact that the product does not solve a problem important enough.

How do you measure it?

Superhuman published how it searched for product-market fit. Its method rests on one question asked of active users: “How would you feel if you could no longer use Superhuman?” The possible answers are very disappointed, somewhat disappointed, or not disappointed.

Some details of the Superhuman case, as they published them:

User filter. They only surveyed people who had used the product at least twice in the two weeks before. At the time that meant 100 to 200 users. Without that filter, the question says nothing useful.
Response segmentation. They analyzed the “very disappointed” group by persona (founders, managers, executives, business development) to understand who benefits most and what they love. From the “somewhat disappointed” group they only paid attention to people who valued speed. The “not disappointed” group they deliberately set aside.
Score over time. According to the source, the share of “very disappointed” respondents started at 22% in summer 2017, rose to 33% once the team focused on the right segment, and reached 58% within three quarters of product work.

The reference number, 40% answering “very disappointed”, appears in the same case. Superhuman attributes it to Sean Ellis, who reportedly got it by benchmarking nearly a hundred startups: those with traction almost always exceeded that threshold. It is a heuristic, not proof.

What can design do?

Help find the segment. “Very disappointed” users tend to share something. Design can ask, observe, and group to understand what they have in common.
Shorten the time to first value. A user who does not quickly reach what they care about does not become “very disappointed”. Onboarding and the first screen matter more than anything else.
Reinforce what they value most. Once you know what your segment loves, you can protect that feature and make it more visible.
Design the reason to return. Notifications, history, or progress only help if they reflect something the user already values.
Measure the experience by the same criterion. Track time to first value and retention by segment, not just visits.

The article on product-led growth explains how product use can replace sales in some models. If you want to connect these metrics with team decisions, the article on KPIs helps you choose which ones to watch.

When should you NOT pursue fit?

Without active users. The “very disappointed” question needs people who use the product. Without use, the survey measures imagination.
When the product has no recurring use yet. The question measures little when nobody comes back. In that case, first understand why people do not return.
As an excuse not to talk to customers. The metric says whether fit exists, not why. User interviews are still needed for that.
When fit is measured in the wrong segment. A good score among users who are not the target can lead to false conclusions.

Example

Illustrative example: a note-taking app for small teams with 300 active users.

Survey: users who used the app twice in the last two weeks are asked the question. 18% answer “very disappointed”.
Segmentation: the “very disappointed” group are design teams that share research notes. They love tag-based search.
Design action: tag-based search moves to the main screen, and onboarding shows an example with research notes.
Measurement: the next quarter, the share of “very disappointed” respondents rises, and 30-day retention improves in that segment.

The change does not come from one metric. It comes from understanding who is disappointed and why, then designing for that group.

If your team uses AI to analyze those open answers, the article on AI product metrics covers what to watch when a metric relies on models.

Common mistakes

Treating product-market fit as a one-time event that is reached and forgotten.
Surveying users who are not in the target segment.
Taking 40% as proof rather than as a heuristic.
Using the question without filtering for recent use.
Redesigning the interface to solve a value proposition problem.
Acting on the “not disappointed” group as if they were the majority.

Sources

Superhuman, How Superhuman built an engine to find product-market fit
Marc Andreessen, The only thing that matters (2007)

Frequently asked questions

What is product-market fit?

It is the point at which a product clearly satisfies a market that needs it. The most cited definition is Marc Andreessen’s (2007): being in a good market with a product that can satisfy that market. He credits the formulation to Andy Rachleff.

How is product-market fit measured?

A known practice is to ask active users how disappointed they would be if they could no longer use the product. Superhuman published that it used a 40% ‘very disappointed’ threshold as a benchmark, and attributes it to Sean Ellis. It is a heuristic, not proof.

What can design do to achieve product-market fit?

Help find the segment that values the product most, shorten the time to first value, and understand what ‘very disappointed’ users appreciate most, so the team can reinforce it.

Is product-market fit achieved once and for all?

No. It is a state that can be maintained or lost. A product can fit one segment and stop fitting when the market changes or the product grows toward other users.

Can design achieve product-market fit on its own?

No. Design contributes to the experience, but fit also depends on the proposition, price, distribution, and the market. Confusing it with an interface problem often leads to redesigning what is not the problem.