Quote from: BelieveInStock on October 28, 2024, 12:09

I just wanted to confirm a sudden drop in sales for long term high ranked contributors. In last two weeks I fell from about 150th place weekly to around 250th. Sales down from 200 a day to 140 a day.
Constant uploads, mix of real and AI. I have been at Adobe (Fotolia) since about 2008.

Very discouraging, must have been some change in algorithm.

Is it already confirmed that Adobe has changed the algorithm? There was perhaps some larger discussion about this on the Adobe Stock discord? Since the end of August I've been seeing declines in my weekly earnings and weekly position, and the last 2 weeks have been downright dramatic! Since September 2023, my weekly position has regularly fluctuated between 200 and 400 and my earnings have oscillated between $500 and $700, averaging $560 per week during that period.

The last 2 weeks have seen a drop to almost 800 position and earnings of only 350 and 325. I can already see this week ain't gonna be any better.  I completely don't understand this, because until now I have never had such large swings, and Adobe Stock is a full-time job for me. I will mention that the daily amount of stock I upload has been constant for a long time

Ditto. It must be ann algorithm change?
If it was competition from AI, then the decline in sales was gradual, not a sudden change for several contributors in the last few months - AI is not new, it�s been around for the last 2 years with lots of contributors. (This spring I was on the best sellers list and never was beyond 4,000 now I�m drastically sliding down as well to 16,000. Last autumn was the best season for me. This time barely anything got sold)

One thing that I can guess is the algorithm favors contributors who submit a lot of content and it doesn�t have to be a quality or even a selling content. That�s why I think AI  contributors get more spotlight than traditional photography or illustration.

I would be very curious to see how Adobe will deal with thousands of AI ports that upload a lot compare to contributors who upload seldom in small amounts. Maybe spamming is the future?

Don't think that spamming is the future.
On average lightning, composition and colors of AI images is better than classic microstock images. 
Before AI most of the images have been just good enough for sale. With AI lots of the images are at the level of best sellers.
There is no change in algos, it has been always like this. Images are replaced in search results with better performing images. 
Spamming is just more low quality of the same.

There must have been an algo change. It's not possible that for over a year I've been in the 200-400 range EVERY SINGLE week, and now just in 2 week times I got kicked into positions 700-800. Last year September and October were my all-time best months. This year October earnings were 40% lower than October year before.

Its the season for Xmas and New Years images.
These images are quite easy to copy with AI. So i think that there is much more competition because of AI nowadays.

 

If the change is abrupt and at the same time for several contributors with very different content then it is probably an algo change.

ai competition comes slowly and gradually.

and spammers can never effect the current season because whatever they copy for christmas now will go live in the next year.

I am steady, slightly down.

However it is a lot less than I expected.

Currently forecast to get 6% more than November last year.

But I was hoping it would double like all the other months.

However October also started depressing but ended at 97% more yoy.

What I like is that I have less bestsellers than last year, sales are spread very widely across the port.

So even if  somebody copies my files, they will only get a small volume of sales and then hopefully move elsewhere.

I had one day with 70 sales. That was also the day many people images went live. But it wasn�t people that was selling.

So, I am steady, but this is bad for me because 60% of my sales are from the winter season.

It takes 2-3 months for content to be approved. So if it gets created in September it might be going live in these weeks. I don't think it would be fast enough to affect sales so drastically.

And the biggest problem is that the new ai people are not fully aware of what sells, they just sort by downloads then duplicate. And duplicate.

Then files need to be picked up, lightboxed, projects realised, most of my files start selling after 12-18 months. So the effect of competition is very delayed.

Is Flux better than midjourney?

I tried it on nightcafe but especially for hands I found it just as bad as midjourney and for faces I find midjourney more authentic.

Perhaps I am using it wrong, though.

eta

I do my standard test searches every day, including browsing the new content coming in. The advantage of ai is that it has better lighting, colors and composition than the amateur snapshots that we were getting before.

But it is obvious that over 90% will never sell, because no research was done for the genre and they are just duplicates of duplicates.

And I don't think this will change so quickly. Research takes a lot of time and ai cannot do it for you.

customers can search only ai or no ai or everything together.

it is one of many parameters they can choose, just like with or without people, transparency etc�

If you do daily test searches, you notice algo changes immediatly, not just for your own files but also for the other files that usually surround them.

Sometimes it is just an exchange, you still have files in good positions, but they are different files. But at other times your files suddenly disappear from pages 1-3 entirely and that is when the drastic drops happen.

Only a very diversified port can give you some protection.

If you rely on individual bestsellers that is very risky.

Just wanted to add that customer interactions of course also strongly change algos. At certain dates or seasons customer interests change, that also changes results.

At the moment I am stuck around pos 1420.

I have the exact same sales volume as last year November.

But this means that unless so,ething changes in the last two weeks, November will be the first month with no growth yoy. Every other month was 100% more.

Not good, half my sales are from the winter season.

I have a good and very diverse mix of sales, everyday files get their first sales, people images are beginning to pick up as well.

But the volume is half of what I am expecting.

In my test searches, I have good positions but with other files.

Maybe they are not as attractive in leading eyes to my port, I don�t know.

With uploads I am focussing on spring summer. Especially with the long queue I think it is important to start much earlier.

I do get some files accepted very quickly, which helps a lot with visibility.

It's rather strange... I believe I have more images than you (from what you've said in the past), but seem to have a lot less daily image sales at the moment... I do have the seasonal items, unique content, useful content, etc... so - not sure... But I did notice for some items that I had that did well in the past - it seems like now there are "literally" 1000+ copies/variations, etc... a bit annoying I do admit. However, hopefully my daily image sales pick up too.

The first two weeks of November have been quite slow for me with all my agencies. In 2023, the same period was also sluggish, but things picked up in mid-November, and I reached my best monthly earnings.

I've also noticed a significant drop in downloads on Artlist over the past two weeks. This is probably due to the U.S. elections and November 11 being a public holiday in many countries, leading to decreased buyer activity.

Today, however, I had eight good sales on Shutterstock. I'm hopeful that the next two weeks will make a significant difference!

There must have been an algo change. It's not possible that for over a year I've been in the 200-400 range EVERY SINGLE week, and now just in 2 week times I got kicked into positions 700-800. Last year September and October were my all-time best months. This year October earnings were 40% lower than October year before.

nice to know someone knows what MUST be happening !  might there possibly be other explanations?

if it's an algo change where are the folk who benefit?  why do we only hear from those who think the marketplace is designed to thwart them!  a drop in earnings could mean there's more competition

Well, maybe checking file positions on keyword searches is more useful to commenting in threads you don't like.

But to each his own.

Maybe next week? Congrats!

Quote from: Infinity on November 09, 2024, 03:28

There must have been an algo change. It's not possible that for over a year I've been in the 200-400 range EVERY SINGLE week, and now just in 2 week times I got kicked into positions 700-800. Last year September and October were my all-time best months. This year October earnings were 40% lower than October year before.

nice to know someone knows what MUST be happening !  might there possibly be other explanations?

if it's an algo change where are the folk who benefit?  why do we only hear from those who think the marketplace is designed to thwart them!  a drop in earnings could mean there's more competition

I always benefit, because they like me best and favor anything I do.

actually i really encourage folk to do daily searches to see their place in a basically meaningless 'statistic'  - the more time they spend there means less time creating images to compete with mine

Actually it's just for ego or watching, like watching a car wreck, won't change anything, it's still going to happen just the same..

Here's a top for the "watchers", make a database and sort it. soon you will know how many DLs = what position, within a few numbers. Let me put that another way... DL = position. The only way you will go up or down is when someone else has more or less downloads in the same time period, than they did, in the previous period.

It's all relative, and changes like the waves on an ocean, but sea level remains the same.