4 ms·
Isn't this basically the same as A/B-testing, which I assume is better and safer because real people are involved in testing new features/changes?
by thatwasunusual 4y ago
Isn't this basically the same as A/B-testing, which I assume is better and safer because real people are involved in testing new features/changes?
- Belalmostafa 4y agoThanks for your response, and yes, you are very right. However, A/B testing has many vital issues, such as it can take really long time and take lots of resources, that is not something that everyone can afford. In addition, A/B testing is needed for every variable under test, But in the case of an AI tool simulating digital users on a scale, it can minimize the time and resources, in addition, it can actually test for multiple variables at the same time. . I would love to know more if you have other questions, or concerns? what would encourage you to actually test or use it?
- thatwasunusual 4y ago> However, A/B testing has many vital issues, such as it can take really long time and take lots of resources I no longer understand your product, when you explain it like this. Are you saying that with your product, product developers no longer have to develop features to test out?
- Belalmostafa 4y agoNo, I mean you no longer need to test the feature on actual users to know how they might react to it.
- thatwasunusual 4y agoIMO, the product is useless then. The cost of developing the feature is developing the actual feature, not testing it. A/B-testing is a no-cost these days; you create two branches, and then deploy it with a 5/95 (or whatever) balancing.
- Belalmostafa 4y agoIn addition, A/B testing means you actually invested in launching the feature and testing it on some real users as you mentioned, however, in our case, you get to get feedback before actually doing that without risking - at least this is our promise - the validity of the conclusions.