Hacker News Reader

Comment by vasco | Hacker News Reader

If heads of product were forced to use their own products for a full afternoon every week the quality of the web would increase a lot.

The amount of high level product people that can't demo their own product and cannot find features they hype up in slide decks is depressing.

View on HN

In my experience it's not just about using a product but understanding how a product functions under the hood. Yelling "fix it, fix it, fix it, fix it" in a loop at random teams tends to not actually get things fixed. Instead you'll get a lot of blame shifting, a lot of random "projects" to improve things and in the end no improvement (or often an even worse experience). In other words you'll get corporate politics. Very few executives understand a modern search stack in anything resembling a useful level of detail much less how their teams map to it.

I've worked on products with ALL of these exact issues. The answer is usually that the data isn't annotated well enough, or that there are not enough developer resources to make special case fixes for things 99% of people don't notice or care about (and so are not losing us money, they just annoy us). At the companies I worked for, fixing these problems was not within the remit of the head of product, they were cross-departmental issues and so were in that pernicious category of problems that are highly resistant to change. I assume the reason you think product people are responsible for this is because you aren't a product person, but in my experience those are the people who most desperately want this type of shit to be fixed, because it makes them look like clowns.

The responsibility of a product person is that the product works well. This includes working with design and engineering partners as well as any data annotation departments. If the search experience is shit and the head of product doesn't think it's up to them, they aren't being responsible for the product.

The reason the responsibility for this is on heads of products is not because they're directly responsible for annotating, it's because they're responsible for knowing what can and will get annotated so the feature they asked for actually works.

The work of product person shouldn't stop at the idea phase.

Not disagreeing, but this example seems like a data/tagging problem which is probably some other department.

Which would be flagged in the weekly review on the first week?

Data tagging makes Walmart throw a huge banner ad for pancakes into your search for lentils?

I mean if someone tags pancakes as lentils... that said we had great success by using gpt4-v for product tagging tbh. it really is a game changer.

Lentil pancakes are actually a thing to be fair.

The money from showing a pancake ad make them do that. It's not a search problem. They know it's not a good experience but they also know it makes them more money than not showing it. The goal is money. Your experience is merely a means to that end.

The profit margins on selling products are abysmal for e-commerce. The profit margins on ads are amazing.

Ideally, manufacturers and sellers would provide more, accurate information about products -- but I generally like how Google solved this within Google Maps via crowdsourcing by prompting users to answer questions about places (e.g. does this place have a wheelchair ramp?). I wish there was a better process to collect and store product information so it could be more easily shared.

And in any sane company, the minute a product person says "oh this egregious error is a problem for some other department" they'd be fired immediately, so it still sounds like a useful exercise.

I'll be the first admit I don't have a ton of respect for the "product owner" role which has morphed largely into Tom Smykowski from Office Space[0] but the places I've seen them be actually effective and net positives for an organization are where they have the reach and authority to demand things like this be fixed, regardless of what department is "in charge." IMO a product owner who doesn't do this is refusing to do one of the few useful product owner roles.

[0] https://www.imdb.com/title/tt0151804/characters/nm0726223

I’ve had good results with brute forcing tagging even when it’s not my job.

If I use my product and hate it, I make it better. Knowing they some other department sucks for years is no good.

I think all this boils down to companies not really giving a flip.

The product is bad. Shifting the blame around doesn't magically make the product better.

Reply on news.ycombinator.com