Selection bias
Error introduced by how people came to be in your sample, when that process is related to what you are measuring.
Selection bias is upstream of everything else. A link posted to your own newsletter reaches people who already like you. A survey advertised as being about cycling gets answered by cyclists. A comparison between people who used a feature and people who did not measures the choosing as much as the feature.
It is not solved by weighting, sample size or a careful questionnaire, because the problem is in who had the chance to answer. It is reduced by choosing a sampling frame that covers the population you want to describe, and by assigning people to groups at random when you are comparing anything.
Describe recruitment in your report, for example: “Respondents were recruited from our mailing list.” This helps readers assess who the findings apply to.
Where it goes wrong
Comparing self-selected groups and reporting the difference as an effect. Randomise the assignment, or describe the result as an association.