Product Research
Product research is the work of finding out what a market will buy before budget is committed to build, launch, and scale it. Research designed to falsify gives a weak concept the chance to fail early, when failure costs a test budget instead of a product line. Most research programs are designed to confirm a decision already made: focus groups approve, surveys report intent, and the concept proceeds to launch, where the market evaluates it under real conditions.
How it actually works
In the standard sequence, research arrives late. The concept is developed, the budget committed, and a focus group, an intent survey, or a generously read soft launch serves as a checkpoint near the end. At that stage the organization has a large investment in a positive result, and stated-intent methods rarely return a negative one, since agreeing costs a respondent nothing.
Falsifying research tests the purchase instead of the opinion. Landing pages for a product that does not exist yet measure whether people act on the offer. Small paid campaigns test which framing of the problem draws a click from a cold audience. Preorders and waitlists measure commitment. Each test is small and fast, and a weak concept can lose it. Concepts that pass receive the next increment of investment; concepts that fail cost days instead of quarters. The economics are an option purchase: the test buys the right to abandon a weak concept before the build budget is committed, and it is worth running whenever the test cost is below the build cost multiplied by the probability of failure. That option value rises with the size of the planned investment and the spread of possible outcomes, so the larger the launch, the stronger the case for the staged tests in front of it.
What gets tested is wider than the product. A launch can fail on three separate fits: product-market fit, whether the segment wants the thing; offer-market fit, whether the price, packaging, and positioning make it buyable; and delivery-market fit, whether the way it reaches customers matches how they buy. A strong product with the wrong offer tests as a failure, so the research has to separate the three. The fix for each is different.
In practice
MrCool grew from $3 million to $100 million across our engagement. Throughout that period, nothing reached scripts or ad creative unless it was backed by data from within the target market, and concepts were proven at small scale before production budget followed them. The larger investments went to ideas that had already sold in small tests. Expansion without that research has a measured cost. A brand that held the number one pre-workout on the market expanded into a biography, a women's supplement line, spicy supplements, and a mobile game while the proven product went underfunded; all their customers wanted was more flavors of the same tried-and-true product. Their competitor doubled down on its own competing pre-workout and took it into GNC, where it did $36 million the following year, while the diversifying brand dropped 25%. Each expansion had a business case on its own; they were not all equally grounded in customer research ([reading the curves](/insights/s-curve-of-growth#reading-the-curves)).
Where we come in
We run product research as staged falsification: the riskiest assumption is tested first, purchases are measured instead of opinions, and the three fits are read separately so a failed test identifies what failed. We test small and prove results before investing, the sequencing we applied at MrCool, Potbelly, and Old School Labs.
Start a Revenue Health Pre-Assessment →See it in action
Related terms
- Falsification
- Designing a test so the concept can fail. A test that cannot return a negative result functions as confirmation.
- The three fits
- Product-market, offer-market, and delivery-market fit. A launch needs all three, and a test that only reads the first misses two of the ways it can fail.
- Smoke test
- A live offer for a product that does not exist yet, measuring real responses, clicks, signups, preorders, instead of stated intent.
- Stated intent
- What respondents report they would buy. It runs well above actual purchase behavior, because agreeing in a survey costs nothing.
- Concept testing
- Comparing framings of a product or message against each other with a cold audience, before creative or production budget follows the winner.
- Minimum viable test
- The smallest, cheapest experiment that can produce a decisive read on the riskiest assumption. Sequencing these keeps the cost of being wrong low.