Live data from Hacker News

Website search hurts my feelings

openmymind.net

121–130 of 201 posts

Re: Website search hurts my feelings

#121

On officedepot.com when searching for a produect I have found no way to filter by "in stock at my chosen store" you have to search, then click through each individual item. The one and only reason I am considering office depot as a place to spend my money is because they're local and I can go get whatever cable or thing I need, right now.

This drives me crazy. Local stock search is different on each store and most don't work correctly.

Try searching on Walmart to find local stock. You have to click "In-Store", then change the Fulfillment Speed to today and tomorrow to find things that are in stock. Except now they also show results from other stores, like Advance Auto Parts, and there's no way to filter that garbage out.

I tried searching Academy Sports for local stock. The list of items says it's in stock but when you go to the individual item it says it's OOS.

Lowe's and Home Depot local stock does seem to work reasonably well, for now.

Re: Website search hurts my feelings

#122
post #92

Earlier quoted context omitted.

> Searching "rice" for example will probably get you parboiled white rice, instant red beans and rice, chicken and rice canned soup, and rice pudding— many of those facets make perfect sense. Well, no, if I ask for rice, the odds are overwhelming that I'm looking for rice, in which case chicken and rice canned soup is a totally inappropriate response. It would make as much sense as returning alcohol-free wine when as…

But chicken and rice canned soup contains rice and the word is part of its logical title. You should be able to unearth the soup with a partial match. What makes the alcohol/alcohol-free problem is that alcohol free is one token that can't be subdivided.

That is not a consistent position. If "alcohol free beer" doesn't contain the word "alcohol", then "chicken and rice soup" doesn't contain the word "rice".

Moving the argument into the realm of things that matter, chicken and rice soup is not a product that can be called "rice", and therefore isn't an appropriate response to a search for "rice". The title is irrelevant; if you're stocking rice with a fancy name and its title doesn't include the word "rice", that product still should show up in a search for "rice".

Re: Website search hurts my feelings

#123
post #92

Earlier quoted context omitted.

> Searching "rice" for example will probably get you parboiled white rice, instant red beans and rice, chicken and rice canned soup, and rice pudding— many of those facets make perfect sense. Well, no, if I ask for rice, the odds are overwhelming that I'm looking for rice, in which case chicken and rice canned soup is a totally inappropriate response. It would make as much sense as returning alcohol-free wine when as…

But chicken and rice canned soup contains rice and the word is part of its logical title. You should be able to unearth the soup with a partial match. What makes the alcohol/alcohol-free problem is that alcohol free is one token that can't be subdivided.

You’re describing the problem from a computer scientist POV. OP is describing the problem from a user’s POV. Nobody cares if there’s a subset of their query in the title. That item is not “rice”. No one is searching for “rice” and wants soup. They would search for “soup” then. Tokens aren’t all of equal value in a title, and this is a big part of why ecom search is horribly bad (I have spent years working on ecom search and partnering with vendors trying to improve it. It’s a hard problem.)

Re: Website search hurts my feelings

#124

Love this article. Worked as a search relevance consultant for years and these are exactly the problems we work on. Most of the time fixing search is a thankless job. People expect it to work and when it does they take it for granted, but when it doesn’t it damages your reputation and brand. If you want to get started fixing website search relevance I recommend the books “Relevant Search” by Turnbull and Berryman, an…

Great recommendation on Relevant Search. I worked on an ecom site trying to improve search relevance for years, and it’s an incredibly difficult and challenging problem. Especially because it’s difficult to measure success, and execs would constantly come up with “hey why does this query not return exactly the results I expected” which led to a lot of resources spent optimizing a handful of queries based on executive input instead of based on query frequency and query result conversion metrics.

Re: Website search hurts my feelings

#125
post #117
post #106

Earlier quoted context omitted.

analytics have confirmed that surfacing inventory status is bad for the customer because it reduces conversion

Bad for the store you mean? Because the customer doesn’t convert (buy) something if it says “hurry only 1 left in stock” because everyone knows that store inventory is inaccurate and they ultimately don’t have 1 in stock, they have 0 in stock?

The way I interpret the comment, it says this:

A. [explicit claim; humorous] Analytics confirm that, if you show people whether products are in stock locally, they won't buy those products on your website.

B. [by implication] This is bad, because it "lowers" your sales. So you can't display this information.

C. [by implication] The retailers are making a mistake - the reason people don't buy products on the website when availability information is displayed is that they go to the store and buy the product there.

D. [summing up] If you measure something you don't care about, you will end up doing things that make you look like an idiot.

Re: Website search hurts my feelings

#126
post #30

Earlier quoted context omitted.

Do not want to miss a chance to empty your pockets

That doesn't make a ton of sense in the Netflix case since presumably it's actually more expensive for every additional movie you watch since your subscription price is fixed and each additional movie you watch costs bandwidth and compute?

Netflix wants you to justify the cost in your mind. The bandwidth is free. Netflix is such a high volume of internet traffic that the CDNs actually have edge content caches within the DCs of most major ISPs, so when you stream you’re actually streaming from a box at your ISP (or a cross connect to Netflix) and not costing Netflix any internet-facing bandwidth besides API calls. This setup is cheaper for everyone because the ISP doesn’t have to subscribe as large of an internet pipe because the high bandwidth streaming is siphoned off into private networking and never leaves the DC.

Re: Website search hurts my feelings

#127
post #26

On the other end of the spectrum, I've been impressed by Netflix search. I would search a name and they do not have that movie, but would suggest a mix of similar movies, movies with the same actors or director. It's pretty clever, I just wish they were more explicit about it, as in showing the reasoning behind choosing a result.

Searching google's movie service for older movies is annoying because a lot of times you just get a remake. Even if they have the old version! The only way I found to get to the original is to search for some actor in it and scroll...

Re: Website search hurts my feelings

#128

Fixing this is the promise of enterprise deployments of LLMs, correct? That you can take the list of initial result texts from a shitty elasticsearch implementation, pass it into an LLM with a prompt "you are a user looking for , please use common sense to comb through these results and find a few good ones" and get better intramural search quality. The only other enterprise application I am aware of is applying the…

Implementing search correctly would solve the problem too, without requiring a square kilometer of GPUs.

If you have >100k items, “implementing search correctly” is an intractable problem if you don’t have a hundred person team of search relevance optimizers and taggers. Search is terrible until you gather tons of metadata about each item. Search is not nearly as simple as text matching within product titles. An LLM by comparison is peanuts. You could run it overnight and just have it build tags and metadata for all your items and it would probably be 75% accurate and a hell of a lot cheaper. Things can then be fine tuned by people later.

Re: Website search hurts my feelings

#129
post #71

Earlier quoted context omitted.

I don’t understand. How is it useful for Netflix to show me movies that they don’t have? And how hard is it to implement search over their catalog when they only seem to have about 100 movies?

Just because they don't have the rights to stream The Office doesn't mean they don't know what shows and movies they have are enjoyed by people who also enjoy The Office. This is literally the entire point of search - show what you're searching for if it exists, and show related things after it.

> This is literally the entire point of search - show what you're searching for if it exists, and show related things after it.

That might make sense in an abstract way, but no, that's not what people expect. If I'm searching for something at Netflix, I don't want the answer to "does it exist?", I want the answer to "do you have available for viewing?".

Re: Website search hurts my feelings

#130

The point of website search (at least on commerce sites) is not to help the users find what they're searching for, but to put other, possibly related, products in front of their eyes. On most sites drilling down through limiters falls apart after about the third or fourth restriction is chosen, often because the sites seem designed to not give three or four clear choices (which is what the user wants to see), but ten…

It's the digital equivalent of putting the cans of nacho cheese dip on a little rack in front of the shelves of tortilla chips.
Post reply on HN