If I'm searching for Scooter, it may be good that it returns also mopeds and motorcycles.
But if I'm searching for Nikon d850, and get first 10 entries for Nikon d750, Nikon d610, canon 5dMkii, Sony rx100, and then finally a d850, then its really really annoying; and that's 80 percent of my experience.
For the other 20 percent, When I do want to search for something generic like bicycle or synthesizer, I frequently want to see new entries since the last time I searched - I don't want to see same synthesizer (or heat pump dryer:) day after day after day. Instead, Facebook's fuzzy algorithm fill cheerfully show me what it thinks is the best answer day after day and week after week, even though its a crappy casio that shouldn't even be listed as a synthesizer and it's been there for 7 weeks and is actually sold; it will refuse to show me chronological list so I can see what's new and added. It will then cheerfully ignore geographic distance and get me excited about something 4 hrs away.
So for me at least it's a lose lose scenario - our experiences clearly differ radically, and it may be simply down to how specific of a thing we want to. I typically want something specific, and if I want generic, then I'd like a category please, something which Facebook also sucks at as others have pointed out :).
(and I lied in my example to give FB a chance... In reality shen I search for "scooter" , I get motor scooters, electric scooters, the Segway wannabes, and six other types of things I don't want, so even then fuzzy search fails me :)
why on earth doesn't occur to anyone that heterogeneous search methods producing a range of results with a identifiable classifiable process or at least perceptual sensibility that users can get some understanding of, might provide not only consumer choice but profit?
The only thing that I can imagine to be happening is the overriding objectives must be econometric optimization and blow every other consideration. Or maybe it's possible with the adherent mantra of "internet speed" development that without constant investment in maintenance "development" nothing will work if left alone while bringing in alternative abilities.
> What absolutely confounds my comprehension of marketplace sites and any online business driven by user directed discovery, is the unbroken and never deviated from practice in every industry and field of making the method and techniques used for search as unique inviolable individual apotheoses of a imaginary USP.
I don't get marketplace sites. Specifically, it seems like every company decides they have a unique selling point, and they should build all search features around emphasizing that selling point at all costs, instead of building it around showing the user the thing they were looking for.
> why on earth doesn't occur to anyone that heterogeneous search methods producing a range of results with a identifiable classifiable process or at least perceptual sensibility that users can get some understanding of, might provide not only consumer choice but profit?
Why don't they provide a search interface that will be familiar to their users? Wouldn't that be both more user-friendly and also more profitable?
> The only thing that I can imagine to be happening is the overriding objectives must be econometric optimization and blow every other consideration. Or maybe it's possible with the adherent mantra of "internet speed" development that without constant investment in maintenance "development" nothing will work if left alone while bringing in alternative abilities.
Maybe the companies are Goodharting themselves to death? Or maybe they value developing new features over making sure that their core infrastructure actually works.
They're trying as hard to sound smart and going too far so it's just incomprehensible.
"the unbroken and never deviated from practice in every industry and field of making the method and techniques used for search as unique inviolable individual apotheoses of a imaginary USP"
Who the hell speaks or even writes like this?
EDIT: looks like their whole posting history is like this. Someone fed postgrad literature to GPT-3 and created a bot from it.
If I search for Nikon D850, and some of the search results are (for example), Nikon D750s at a significantly lower price[1], then at least for me personally I'm happy to be shown that result.
This isn't a hypothetical - I bought a Canon 1000D after searching for 1200Ds and 1300Ds and finding it for a crazy-low price locally.
I understand not everyone is like this, though, and they should at least properly support the option for less fuzzy or no-fuzzy searches.
[1] I don't know cameras well, but I've assumed the 750 is the previous year's or slightly lower specced model of the 850.
The D850 is a seminal state of the art camera that internally is identical to the contemporary flagship the D5 in every way except for major improvements like high resolution and exceptional sensor which is arguably not surpassed almost six years later. Whilst the D750 is a cost engineering marvel. The D750 is a very fine camera, but nothing in the league of the D850.
If you are actually considering purchasing a D750 I'd strongly recommend that you investigate the D780 which only appears to be much the same, but includes the majority of the successor mirrorless technology developments including excellent video giving live view operating far better than any other DSLRS from Nikon including the flagships.
Nikon managed some exemplary product differentiation in the last range of DSLR cameras. Unfortunately for everyone Nikon can't seem to be anything like as clear and focussed with the Z series mirrorless range.
Edit: the D850 compared with the contemporary flagship D5, is capable of only slower rates of shooting, has a smaller battery, no extended grip and dual orientation controls and the sensor can't -quite- perform in low light like a D5. In every non sports shooting environment the D850 would be regarded as a flagship, but for historical reasons and the attachment rate of D3/4/5/6 users buying proportionately many more elite lenses, it's the lower resolution action and low light big pro cameras that get the flagship accolade. Single digit pro cameras also optimize photo processing in camera for JPEG output as is big agency wont and urgency the need for finished product out of the box without any processing. In case you're looking into this, JPEG output from the D5/6 is incredibly good. "Merely" 20MP sensors in these cameras are far better than any specification can indicate, being developed to produce the cleanest pixel level signal, which really is visible in the output. A hardly used very good condition D5 can be had for $1500 lenses dumped en masse for Z upgrades are bargains. If you can deal with the bulk of these cameras, you'll very unlikely i've dissatisfied. And de facto unbeaten autofocus with a major feature omitted from new designs Group Priority AF which glues the focus to your selected nearest subject without any deviation that happens in the Z bodies because of the omission of this professional feature. Novices are apparently buying the Z9 mirrorless flagship because it's not got AF compromises to frustrating degree. But you can have a complete kit for the same money in the DSLR line.
But even then, Marketplace will heavily bias you towards D850 parts if you type in "D850". It's not like it autocorrects your search term to "camera" - it instead just inserts a few generic listinga into your results.
Random people listing stuff on local classified sites don't always use the correct keywords, and it's in everyone's best interest for these listings to actually get a good amount of views.
I suspect this "fuzzier" algorithm accounts for the far lower amount of spam too. A search for "heat pump dryer" on Gumtree will return 25 listings in a row from the same shop in Oakleigh, whereas Facebook will show different products from different sellers (some businesses, but mostly randos) - not to mention the people selling their heat pump dryer without including those magical two words in the title.
It's uncanny how good this algorithm is at returning what the buyer is actually looking for.