minimum-viable-product, beta-testing, testing, prototyping, iterative-developmentSitting a user down in front of a prototype or early iteration is an ideal way to test the assumptions behind a product/service and drive the direction of an minimum viable product (MVP) public release. The whole point of this early stage testing is to test assumptions and try to identify where the MVP is.
For this kind of early stage testing, how many test subjects would someone need to validate their assumptions?
What would be an ideal scenario? What would be a reasonable scenario where the startup has very limited resources (including time)?
See this related question, answer, and follow-up discussion:
https://startups.stackexchange.com/questions/1349/how-many-customer-interviews-do-you-need/3175#3175
In a nutshell, iterate with groups of 5-6 users. In each step, you’ll encounter the pieces of information that stick out, and you’ll already begin to find duplicate points. Address them, and then restart the process with a different set of users (if possible… if not, find 15-20 users and split them in 3 groups).
Important point: you need to conduct synchronous, qualitative information gathering to get meaningful data. This means actually interacting with the users in a reasonably live environment – as in speech or interactive chat. You want the raw, unprocessed, undigested information.
If the information gathering is asynchronous, as would be the case with email, forums, forms, and so forth, you run the risk of missing out on very salient data because users will rephrase what they mean to make it shorter or draw their own conclusions.
Also, keep your eye on the actual requirements. Per Henry Ford: “If I had asked people what they wanted, they would have said faster horses.”
All content is licensed under CC BY-SA 3.0.