5. Measure success

Define both quantitative metrics and qualitative metrics to measure success. Early in the product development cycle, the number of users may be small or if you are modifying an existing product for a large number of users, the effect may not be measurable at first. Qualitative research techniques, such as user interviews and usability tests, allow you to measure the impact of your work with a small sample size. However, do not leave quantitative metrics to the end. Since you want to measure a change, you need to see before and after numbers. Ideally, you should look for numbers that increase or decrease weekly or monthly over the course of the development cycle as the software improves.

Always remember to test your metrics with qualitative analysis, just because people visit your FAQ doesn’t mean it was helpful to them. Be careful of surveys. It’s better to actually look at behavior rather than asking people what they think they do — well-intentioned people often self-report incorrectly for a variety of reasons. Involve experts in user research who are experienced in methods for evaluating responses.

Success metrics may include business outcomes such as reducing infrastructure cost or human resources required to provide a service.