Thanks, that’s an useful and needed feature

I’ve noticed that most all of my new images aren’t selling under this never search system. It has been about a week since it went live and I have noticed it has had a negative effect on my sales so far. SS has always been the one agent that I needed to upload a constant supply of images to or my sales would drop off quickly – the new content made up a large portion of my sales but now those newer images aren’t selling in the same fashion as they used to. My keywording team are highly trained and do a great job of tagging my images so I’m not concerned my lack of sales are down to keywording issues and I’m confident with the image content and quality.

This system is obviously in it’s infancy so perhaps my port isn’t fully up and running yet. quote:It’s designed to get better over time. We will be refining it as well, based on what we learn. Results will vary from day to day, and may vary from user to user as we make adjustments: end quote………………I hope I personally see an improvement.

Anyone else seen a similar pattern?

Dean

Definitely.
Same here. New ones don’t sell as they used to since those changes took affect.
Going to do some changes, but for me it doesn’t work so far.

Definitely.
Same here. New ones don’t sell as they used to since those changes took effect.
Going to do some changes, but for me it doesn’t work so far.

I’ve noticed the same effect. And it’s negative, definitely. It seems that newly uploaded images are dropping into nowhere or selling, but very weakly. All these started right after those big changes. However I can’t say I experience rather noticeable dropping of sales but anyway I expect further improvement of these innovations in search engine.

if i understand ‘relevant’ searching, that may mean that keywords must be heirarchical in relation to describing the images..
what i mean is.. if it is an image of a dog… if ‘dog’ is the first keyword, then it will show up higher in the search than if that keyword were placed as the 50th keyword. that would explain why images might be buried deep in the search results..
does this mean that we’ll have to edit every one of our images and re-keyword to ensure that the keywords are in their relevant order ?
this is important to know

As you know, once image has been uploaded to Shutterstock, it’s keywords automatically sorted in alphabetical order so I don’t think we need to re-keyword any of our images. Don’t forget that this is actual for all authors. I only can suppose that new sorting order is much more advanced algorithm based on the number of different factors, it looks like another variation of iStock’s “Best Match”.

I came here to write exactly what the others already wrote…
New images used to boost sales, but now they either get no downloads at all, or they sell 2-3 times on the first day and then downloads for these images stop completely. I guess that (also?) has to do with the fact that after a few days, images only show up on the keywords they previously were downloaded on. I.e. if an image with 30 relevant keywords was downloaded twice on 1 keyword, it will only show up in a search for that particular keyword, but not at all on any of the other 29 keywords. And needless to say… if a potential buyer does not see it (s)he will not be able to buy it.

I know my keywords are fully accurate (we all know how strict IS is and I’ve never had a keyword rejection there) but most of my images are nowhere to be found, even if they were downloaded a lot on their key keywords. For instance my bottle of champagne image. It’s in the first row of the most popular search and was mainly downloaded on the word champagne. Still, it does not at all show up when I search for ‘champagne’ with the relevancy search.

In Darkroom, Image Gallery Stats, we can see Downloads per Keyword. This implies Shutterstock “know” the keywords what was used for every single sale. If the image has the sale with some keyword (for example “woman”), it means the image was for buyer relevant and new algorithm should prefer this image with sale in next searching with exactly the same keyword (example “woman”). I think this is mainstay of relevant algorithm.

I tried using this new tool and it doesn’t seem to be working. I entered the word baby and didn’t get a response and I should have for that word.

When comparing how high images show up using “Popular” and “Relevant” searches, it seems to me that the number of keywords assigned to the image by the contributor is what makes the difference. If you have a picture of an XY, and “XY” is one of just five keywords, it shows up much higher in the relevant search than an XY picture with say twenty keywords, provided that both score more or less the same in popularity.

It would be interesting to hear SS’s comment on this.

My problem is that I have often used a lot of keywords, and although I think they are all relevant, the “relevant” search downgrades them compared to another equally-well selling shot, whose contributor has simply been lazy and inserted just a few keywords. So fine-tuning your keywords has turned from an asset into a liability. Keep it simple from now on, and use only the keywords that you believe the buyer will most likely use.

For the same reason, the new “relevant” search probably gives better results to the lazy searcher who doesn’t want to refine his/her search very much. But a very refined search will suffer. Is this a positive development?