Excellent observation. A/B testing has no implicit morality. Indeed, driven to its logical conclusion it could result in your business engaging in spam, fraud, or encouragement of addictive behavior.
A/B testing is best used as sandpaper, to round off sharp corners, if you just let evolution guide your product with a free hand you are just as likely to end up with a giraffe as with a black mamba.
> there is a point between pandering to people and providing a service.
Good insight. People cry foul when they perceive a service to have gone too far in the direction of pandering. And I feel like products that tend toward that end up seeming empty and derivative. There's a certain level of non-imagination that things that were extensively focus group tested seems to elicit. You can almost perceive how safe they are trying to be.
And the terrible thing is that they succeed at it.
It's not a popular opinion, but I'd be more interested in seeing products that were not as reliant on A/B testing. I'd prefer to see more iterative approaches taken. Users are saddled with a certain mental model about how things work, and may not be able to think clearly enough about the problem to realize there's a better way to do things...if we forget about all of we think we should know.
On this site we love (ohmygoddowloveit), AB Testing, metrics tracking and all that goes along with it.
"Games getting more addictive" is very similar to "We used Abingo and this is what got the most responses."
I'm not saying this is bad so much as it doesn't represent everything and that there is a point between pandering to people and providing a service.