User testing
User testing, also called usability testing, is a research method in which real users complete specific tasks on a website or in an app while a researcher observes their behavior, problems, and reactions. The goal is to uncover usability barriers before launch, or to find the spots on a live product where users get lost, drop off, or fail to complete an action. Unlike surveys or analytics, user testing shows not just what is happening but, above all, why.
User testing is one of the investments with the highest proven return in the entire digital product development process:
- Just 5 users can uncover 85% of problems. Jakob Nielsen and Tom Landauer built the model on 83 studies, and Laura Faulkner validated it with 60 users (Nielsen, 2000; Faulkner, 2003)
- Every dollar invested in usability comes back a hundredfold. Forrester: 9,900% ROI. NNGroup (42 redesigns): investing 10% of the budget improves key metrics by 83% (NNGroup, 2018)
- 70% of shopping carts are abandoned. Baymard Institute (meta-analysis of 49 studies, 4,400+ tests over 14 years): improving the checkout can raise conversions by 35% (Baymard, 2025)
- Late fixes cost up to 100x more. The IBM Systems Sciences Institute documented that the cost of fixing a defect grows from 1x in the design phase to 60–100x after deployment (IBM SSI)
FAQ
How many users do I need for testing?
For qualitative user testing, 5 participants per iteration are enough. On average, they uncover 85% of usability problems. This was confirmed both by Nielsen and Landauer's mathematical model (1993) and by Laura Faulkner's independent validation with 60 users (2003). The key, though, is to test repeatedly: 3 rounds of 5 users beat a single round with 15 participants. Each round brings new insights from the revised design (NNGroup, How Many Test Users in a Usability Study?).
When is the right time for user testing?
As early and as often as possible. You can test paper sketches, clickable prototypes, and finished websites alike. The earlier you catch a problem, the cheaper it is to fix: according to IBM data, a fix in the design phase is up to 100x cheaper than one after launch. Nielsen Norman Group recommends allocating 10% of a project's budget to usability testing, because the average redesign after testing improves key metrics by 83% (NNGroup, Return on Investment for Usability).
What's the difference between user testing and A/B testing?
User testing is a qualitative method: you watch how specific people use a product and find out why they run into problems. A/B testing is quantitative: you compare two variants across a large sample of visitors and measure which one performs better. The two methods complement each other. User testing uncovers problems and suggests solutions, and A/B testing then verifies whether those solutions actually work at scale.
Sources:
- Nielsen Norman Group: Why You Only Need to Test with 5 Users (2000): mathematical model based on 83 studies, the basis of the 5-user rule
- Faulkner, L.: Beyond the five-user assumption (2003): validation with 60 users, published in Behavior Research Methods
- Nielsen Norman Group: Usability ROI Declining, But Still Strong (2018): survey of 42 redesigns, average KPI improvement of 83%
- Baymard Institute: Checkout Usability Research (2025): 4,400+ user tests, 14 years of qualitative research
- Baymard Institute: Cart Abandonment Rate Statistics (2025): meta-analysis of 49 studies, average abandonment rate of 70.19%
%201.webp)