5 ms·
Apart from the current answers let me give another reason; product managers (I'm also one). We are excellent at finding niche corners and performing AB testing
by ilayn 5y ago
Apart from the current answers let me give another reason; product managers (I'm also one). We are excellent at finding niche corners and performing AB testing and this and that and then making a business case to increase whatever metric whatever much. This is not much to do with the tooling but the current climate of doing business. On their own right, indeed metrics go up and profits again demonstrably increase.
However, this is (100-eps)% of the cases, demonstrably, worsens every product or at least causes its initial innovative aspects. The reason is that everybody thinks locally and makes the assumption that the function is monotonically increasing and when hit on a plateu or decline, make extreme changes to the product (and calling it disruption in the meantime). And much to my regret, management is one of the worst accountable professions similar to Human Resources in our era. If you listen to PM courses, cringe worthy practices are sold as success stories. On top of this, we have much too power with very little accountability hence following the entropy principle, much higher probability of doing the wrong thing.
Writing this while I desperately wait for Jira to render the page.
- mohanmcgeek 5y ago> We are excellent at finding niche corners and performing AB testing and this and that and then making a business case to increase whatever metric whatever much. I've always had my reservations against this approach but at this point I am wondering if data driven product design produces positive results at all. Clearly YouTube PMs did run experiments before deciding to kill the dislike button but didn't pause to think what the real purpose of that button was supposed to be..
- ilayn 5y agoTrue but you might surprised what that dislike button is actually correlated to.I don't know but I can guess. Often "content creator experience" is way less important than preventing some companies having their campaigns downvoted to death. So there might be a strong commercial incentive which also youtubers don't pause and think who is benefiting from this.
- mohanmcgeek 5y agoThey already had the option to let the creator hide dislike counts. YouTube is less useful now that I don't know what's clickbaity scam. All these PMs and not one of them thought.. "hmm.. would we remove product ratings if we were Amazon"
- dredmorbius 5y agoW. Edwards Demming is strongly recommended as an antidote to the local optimisation / global Schlimbesserung phenomenon.
- mindslight 5y agoI'd like to point out that "performing AB testing" is itself exactly the problem described by OP. Change some functionality out from under some users, to find out how they adapt. It's essentially performing ad-hoc psychological experiments with your regular userbase.
- ilayn 5y agoYes but I don't see your objection. That's basically what google, apple and alike are by definition. The marginal benefits and the techniques used to reach them are very sound. The cost function they choose is not.
- mindslight 5y agoI don't understand the distinction you're trying to make. The past alternative was to hire UX testers, do focus groups, and have alpha/beta testers. It's obviously advantageous for companies to instead push this work onto the users, and the cost function they choose of course doesn't take into account what are now externalities.
- ilayn 5y agoThere are way more details to the AB testing over. UX is just a percentage of the total product. Users are being subjects because their behavior is what you make money out of. Is it moral that's a different discussion.
- mindslight 5y agoThe UX part is mutually beneficial, so it's at least possible to make an argument that companies doing so are benefitting users more than traditional UX design processes would. I still think it's creating externalities for users and would appreciate informed consent (something along the lines of bringing back proper versions where nobody would ever be forced to upgrade). Optimizing to the detriment of the user (eg for engagement, conversion, whatever) is straight malevolent and I've yet to run across a moral justification for it apart from market nihilism.