Does Makerworld play favorites in Searches

Ok, how does Makerworld create the order of models when you put a search term in and hit go?

I ask this because from my position it looks like they play favorites (or blackball?) some designers. Whenever I look up any tag that I put on one of my models the model list appears and my models (all) are at the end. Some are small lists so that is ok, but on the big lists it really is effectively banishing me from having any looks. This has been checked multiple ways with the same result.

Now I have checked and models above (even first 12) mine are older, are hueforge art, etc. Pretty much the same but I am at the bottom.

If anyone wants to check - search tag = papercraft
account = mclanesmemories

I have had a number of other people search and I have always been last… Am I being secretly punished??

It’s not like you’ve whinged about Makerworld at all in the last month…

Hang on, I thought my lack of prominence in the MW system was my range of models being of little interest to many people.

Are you suggesting there is another reason? I have also been complaining a lot.

I think it could be both.

:crazy_face:

The search results show based on the selected filter.

If you search “articulated fidget” number 93 shows the first of my models.

There is clearly a ranking system based on multiple factors unless you select things from the filter.

I am different from most, I don’t care if people like what I produce as I just create things I wish to design. It is nice if people like and download them, but, at my stage of life, that is just a bonus.

1 Like

Actually, I asked this question back on March 22 but didn’t get many replies then (I was very clear in the post). That was before Makerworld started the wholesale chopping of the point system - hence my “whinged” started after.

But I wonder, if I blindly worship Makerworld and didn’t try to let them know the effects of their changes - would I get better search results?

Example using search - marvel bookmark

the first is the oldest file, one has no downloads, the second of mine has more downloads that most… but every search ends up like this.

I doubt it.

It’s easy to play the victim and feel like the world is out to get you.

1 Like

ok, please do the search I did above and show me what you get. Am I playing victim or is there something more… no photoshopping please :wink:

We have the same results.

Same here too :+1:

The difference I see is that the hueforge models that rank higher are labeled as such in the title.

857e3e3c9a0f7fa4d3df40fa55226b01e6b90174_2_690x446

So has pretty much everyone tbh

I feel the need to complain about that!

4 Likes

IMO Printables is 10x worse. I have a model with nearly 2.3k downloads and it’s nearly last in search, behind models with 0-5 downloads and likes. 700 likes, 2200 downloads, 25 makes, 4.5 star (bad bed adhesion is NOT an excuse for a low rating :expressionless:). Sort by popularity and boom it’s second. Be glad MW is better than that.

The problem I see with that screenshot, is there’s only 3 actual bookmarks that came up. At least they’re towards the top.

If I didn’t have any context, and there wasn’t any wording, I wouldn’t have guessed someone even searched for bookmarks based on the way this looks.

4 Likes

Interesting. I do have some with hueforge in the title but that was before they cut the title length in half.

Ok, I found one that has it in the title and it is still at the bottom of the list. (searched Pokémon bookmark). I did find another person who appears to be suppressed as well - TeraBite. his bookmarks are mixed in with mine and a few just before mine at the end.

I tested this.

Searching for “Minecraft”, one of my kit card that was featured in the past appears as twelfth item in the search results, with the first 11 items having lower downloads than my Minecraft models.

(Sorry I tried embedding screenshots but forum won’t allow me)

Searching for “Minecraft kit” however, first nine (9) items were all my kit cards, with the first three (3) being the ones that were featured in the past.

My 2 cents is that, it depends with the the search string entered. I guess the more specific strings that match your model’s title, description and tags will make you models appear at the top of the search results.

2 Likes

Search terms appearing in the title are favored significantly. And titles with shorter length that fit the search term rank even higher. This is also true with hits on keywords.

I think they shifted to that system because titles and keywords were being spammed with a lot of combinations that folks used to get their model to show up everywhere. So now you need to be careful on your title and keywords to use the minimum amount required to make your part searchable. Do you want to show up at the top of a search? Make your title and keyword list short. You will show up very high for that one search. But you won’t show up on anything else. Or you can spam the title and keyword and show up middle ground everywhere. Choice is yours.

There are other factors clearly at play but the above is a significant part of the search ranking.

It has a big Facebook feel to it, but I reckon its like the majority of modern life that we’re fed what they deem to be worthy to us. It’s unfortunate that its come all at once and there has been many changes that things like this can be slipped in. End of the day we have nothing keeping us here, so are free to leave any time. Personally even with no rewards I’m content and happy to stay :slight_smile:

1 Like

Perhaps there is opportunity for MW SEO professionals :joy:

Personally I would prefer the main entry to discovering models be a search box and a For You page, with consistent but not released listing algorithms. This gives a wider discovery window as long as multiple factors are taken into the algos.

Listings by Downloads and Likes counts leads to concentration, which is fine but doesn’t cast the discovery area wide.

Featured is where I’m not in favor of. And Featured while in a contest is doubly questionable imho.

:grin:

Another weird thing to me is when you print a model through the handy app this doesn’t count as a download.
That’s not the case when you print it via Bambustudio. This is a DL + a print (if successful of course).

But that means that the sorting per downloads is not accurate. I don’t understand why they are doing it this way, that makes no sense to me :man_shrugging:

If it’s just a question of limiting the reward points they should just count it as a DL and a successful print should equal to 1 DL not 2.