I'm a new contributor to the site and I've upload only a few files to test the waters.

As I started the use the site more, the results are not good at all. The search engine is a mess and contributors are spamming like crazy. If you do a search for "Hospital icon" on SS, the default search results are good and there isn't a lot of repetition unless you change to the relevant tab. Do a search for the same term on Fotolia, and the results are really lousy.

Go a few pages deep and you see the same images in about 30 variations. Fotolia's search engine is not smart enough to know that, but at least tweak it to prevent repetition of the same keyword set from getting such high rankings.

At it's current state, it's a bad environment for new contributors. I like that SS is more strict when it comes to quality. Maybe Fotolia is starting to become more strict, but something has to be done about the spam on the site that is hurting not only new contributors, but everyone else as well.

Are you sure, that SS is better?

http://www.shutterstock.com/portfolio/search.mhtml?gallery_landing=1&gallery_id=1256674&page=1&safesearch=1&sort_method=newest

Yeah, but those images are not high on the search results.

It seems like the majority of the contributors at Fotolia are keyword spamming and the search results favor the legacy keyword spammers.

At Zazzle there is the same problem.

Are you sure, that SS is better?

http://www.shutterstock.com/portfolio/search.mhtml?gallery_landing=1&gallery_id=1256674&page=1&safesearch=1&sort_method=newest

Wow. That's alot of weed. 

At least 100 pages of it!

Quote from: Dumc on July 03, 2015, 11:47

Are you sure, that SS is better?

http://www.shutterstock.com/portfolio/search.mhtml?gallery_landing=1&gallery_id=1256674&page=1&safesearch=1&sort_method=newest

That has nothing to do with keyword spamming.

To give Doug Shutter credit, I think his keywords are quite accurate.
 

Now it's a bit too late to implement it, but from the beginning search results should have considered number of sales according to number of files for each portfolio. Very big ports with few sales would be thrown back at the search. That would hurt spammers and overall we would have much less ugly similars, and more exposure for new images.

I would say 90% of new images in my "niche" are pure spam.
That's why we are seeing millions of new files per month.

Welcome to crowd sourced microstock Padawan.  You're not revealing breaking news here.

Now it's a bit too late to implement it, but from the beginning search results should have considered number of sales according to number of files for each portfolio. Very big ports with few sales would be thrown back at the search. That would hurt spammers and overall we would have much less ugly similars, and more exposure for new images.

I can't speak about fotolia, but that's the way Alamy works and it isn't always for the good.

A while back (years) there was a particularly egregious best match algorithm on iS with all sorts of poor results. Turned out that one was indeed favouring contributors with many sales. But that had two problems: some of the people who'd been there for years and had accumulated many sales had many spammed files (either deliberately or because they'd chosen not to disambiguate when the system was introduced); also (from my own field of interest and I'm sure it was replicated elsewhere) for example some high ranking studio photographers went on safari, uploaded lots of pics and somehow didn't seem to know what they'd been looking at and had misidentified a lot of the animals (why they didn't listen to their guide, I can't imagine). So a lot of their totally wrongly-labelled files were trumping correctly-identified files from lower ranking contributors. Like I said, I have no reason to think that wouldn't have happened in other subject areas too.

Welcome to crowd sourced microstock Padawan.  You're not revealing breaking news here.

Exactly!

There is no consistency and everyone has their own idea of how to keyword images. Add images that are keyworded by services or programs that generate bad results, and non-english speaking people relying on a translation service or a guess , or even worse simply copying and pasting what they see in other images and you get some pretty odd sets of keywords. The agencies don�t help and it�s too late for any of them to change the rules. If they do, they get nothing but complaints. It�s a free for all. Here are some actual examples (I�m not making these up).

Some put a keyword in for every little thing in an image or make up qualities an image doesn�t have. If a man is holding a laptop they also list words like hair, teeth, fresh, energetic (he�s just standing and holding a laptop), clothing (just in case someone is not looking for a naked man holding a laptop), mood, healthy, happiness.

One red apple get you these keywords - vitamin, set, lifestyle, composition, macro, meal, circle, object. Are any of those needed?

Some miss the important words. If they took a look at the databases they would see that there are already �enough� photos of certain subject but they still shoot and upload. You may argue that there is still room for �better� shots of any subject but how do you get those in front of the buyer who is entering 3 words in a search and then weeding through thousands of images? Not by the keywords.

Different sites weigh keywords differently. Some say that the first ones in the list are the most important, some allow phrases, some sort each word alphabetically. Reviewers may take a cursory look at the original keywords but most only review quality. I imagine many reviewers are not english-speaking.

Some sites allow keywords to be changed at any time. Some do not. One agency went through and removed the words breasts and nude from tons of images because people know that those are common search words. That is what you really call spam. Agencies rank images differently. Each search query interacts with the site�s ranking factors (sales, views, portfolio size, etc) to determine what images will show up on the first page and those algorithms are tweaked from time to time. No one knows how each sites ranks an image and each site does it differently so you can�t game the system if you upload to multiple agencies.

No search algorithm is effective if it includes keywords that are incorrect. If images of gorillas contains the keyword monkey (gorillas are not monkeys) those image show up when a serious buyer searches for a spider monkey because monkey is in the keywords. The database doesn�t know that the image is not of a monkey (most people do not know that a gorilla is not a monkey either). The buyer gets mad when gorillas show up in the results but there is nothing that will prevent this other than removing the word monkey from the keywords.

Intelligent keyword search analysis is evolving and improving search results when there are tons of data but it has a long way to go. As much as we hate google they have pushed everyone to come up with better SEO. Today, the importance of each keyword is weighed differently for each image and helps buyers by displaying the most relevant images first. For example, if a buyer searches for �radish� and photos of single radishes are most often purchased, more weight will be given to those single radish images than the countless food shots of prepared dishes which also contain radishes as an ingredient. Too many keywords in any image dilutes the results no matter how much a person argues that they should be allowed to add unlimited keywords. The world of search technology is evolving but no matter the code behind the system the axiom �garbage in, garbage out� still applies.

Spam or just plain ignorance. There is a fine line.