pancakes

MicrostockGroup Sponsors


Author Topic: how istock search algoritm favors exclusives  (Read 5297 times)

0 Members and 1 Guest are viewing this topic.

« on: December 02, 2008, 18:12 »
0
I want to give an example of how the search algorithm on IS favors exclusives and thus how important such an algorithm is to get sales.

Let's examenate this file of me.

It was uploaded 8/2007 and had between 30 and 50 downloads until 01/2008  Than downloads stopped and fell back to <10 / month

At 18/01/2008 exclusive member liliboas uploaded this file.  It has 100 DL / month

Both file are almost similar.  If we run a search on IS with the keywords red + pink + rose + isolated and sort it by most downloads, those 2 files are n1 and 2.

But if we sort it by best match, liliboas file is still first, mine drops to place 50 of 466 files. some other observations :

- only 2 non-exclusives in the top 50
- many files with even 0 downloads, all exclusives
- even some old files with 10 downloads or less, all exclusives
- my file has a DL / view ratio of 294/3550  lilliboas file : 821/17013, which is much less. 

What do I want to say with this ?

It's obvious to me why my downloads stopped drastically in January for this file.  To non-exclusives this mean that you need to have a very unique file if you want to float to the top.  With average files you'll be burried under photos of exclusives

I'm glad IS dropped to 4 place for income, they deserved it :o) :o)
« Last Edit: December 02, 2008, 18:14 by Perrush »


CofkoCof

« Reply #1 on: December 02, 2008, 18:22 »
0
First of all let me say that I agree with you: best match favours exclusives and even more so since the last change. However it's really hard to prove anything like this. Why? You don't know how exactly the best match works:
- maybe the recent downloads are more relevant
- many people confirmed that images with a good dl/view ratio get pushed towards the back (even the ones from exclusives)

Also it's hard to compare yourself with someone who has quite a lot more downloads than you. Buyers bookmark contributors and someone with much greater exposure than you is (probably) bookmarked more than you are. And get's more downloads from people directly searching from his/hers portfolio.

« Reply #2 on: December 02, 2008, 18:56 »
0
I think that it is pretty obvious and it has been shown over and over again that the best match favors non-independents.

That being said, I would like to be honest with you.

If I had to choose between the two thumbnails, I would have to say that I actually like the image from Liliboas better because the colors are more "colorful" and saturated.  The problem is that if you would have submitted that image for review, it would probably have been rejected for oversaturation.  ;D

On closer inspection of the images (at 100%), I find that your image does not contain real flowers, which is another reason that buyers might prefer the one from Liliboas.

« Reply #3 on: December 02, 2008, 20:09 »
0
I think Stockmaniac is right about the artifical flowers.  But curious how the image is similar,,,

Anyway, I don't think is ok to priviledge exclusives just because of that, and in any search we see this happening - pages and pages of exclusive material, with excellent non-exclusive material buried in further pages.  Quite frankly, I don't see how it can be good for IS - relevance (keywords) and quality (perhaps expressed in dld/mo) should be the most important things so buyers find what they need.  I accept that IS wants to priviledge their exclusive - it makes all the sense - but I think they are giving too much weight on exclusivity.

Regards,
Adelaide

« Reply #4 on: December 02, 2008, 20:24 »
0
As an exclusive with a small portfolio consisting of mainly old photos I have to say that I noticed a difference. The graph below exaggerates the effect of the best match change because I uploaded new images right before the change that account for 40% of my downloads for the last 3 months (more so in first part of September when new uploads were at the front of the best match), but even without the new uploads there would have been a noticeable increase in sales due to the new best match. BIG CAVEAT: I have a small portfolio that can't be used independently to draw any conclusions about the site as a whole.


AVAVA

« Reply #5 on: December 02, 2008, 21:15 »
0
Hey Perrish,

 I thought your presentation was well thought out and showed you took the time to try your best to show your point in a sensible fashion. I was especially impressed by the detail you went into. You will not convince everyone but I think you made a darn good effort at trying to explain your point. Well done and very helpful.

Best,
AVAVA

« Reply #6 on: December 03, 2008, 04:26 »
0
hi all,

first of all let me say that I'm not telling that my image is better then the other.  It's not in many ways.  This is not the point of my comparison. 

I'm only showing / wondering why I'm so far in the best match search even if I narrow the keywords so that they match my image almost perfectly.

If I leave the 'pink' out and search for 'red rose isolated', I even can't find my image  ;D

That said, newbies need to be aware that this is one reason why they don't get many downloads.   I'm lucky I've already 300 DL on that image so it show up as one of the first of the non-exclusives. 

---

Quote
Also it's hard to compare yourself with someone who has quite a lot more downloads than you.

first thing you need to ask yourself is why did she get that amount of downloads ?  Is it a coincidence that my downloads/month dropped from 40 to <10 just at the time she uploaded that file ?

Just look at lilliboas portfolio.  She had more downloads the past 300 days than Yuri.  This makes here the top contributor on IS !!  Whit that portfolio ????  come on !!

Quote
I would have to say that I actually like the image from Liliboas better because the colors are more "colorful" and saturated.

you mean oversaturated and blown out in some channels ??  ;)  I still wonder why buyer keep preferring such images.  turning up saturation is easy, turning it down without spoiling the image is a lot more difficult.

Quote
But curious how the image is similar,,,

well, there are not so many top view, isolated rose for sale.  just go out and find a view I would say  :D
« Last Edit: December 03, 2008, 04:36 by Perrush »

« Reply #7 on: December 03, 2008, 04:39 »
0
No doubt, that exclusive files are favoured, but in this case the download/view ratio plays a big part I believe. iStock seems to favor with this best match a file which has a lot of views compared to downloads. Liliboas's file has ~20 views/dl while your file has just 12/dl. That also explains why an old file with almost no dl the last year has a prominent place in the best match in my portfolio. It was once offered as a free file and has ~425 views/dl.

« Reply #8 on: December 03, 2008, 04:54 »
0
@ freezing :

20 views/dl is worse than 12 views/dl

it means that for each 20 people who looked, one DL'ed the file.  Hence found it a good one.  While in my case for every 12 people who looked bought it.  So my image is seen as a better one by the buyers because as a % more of them bought it.

This is an easy way for 'killing' spammers.  Their images will have many views (because they show up in many searches), but few DL (because no one like their image).  So the algorithm pushes them back.

I'm almost sure DT uses this methode.  And it makes sense. What IS does, doesn't make sense and in the long run it will hurt them.  I'm sure for that. 

Why ? Because many microstock images are so easy to reproduce.  It's just a matter of time before all the good selling images of exclusive on IS are copied and can be found on other microsite too.

« Reply #9 on: December 03, 2008, 06:03 »
0
Perrush, I agree with you, although I think you cannot make best match position dependent on a dl to view ratio at all.

« Reply #10 on: December 03, 2008, 09:22 »
0
Perrush, I agree with you, although I think you cannot make best match position dependent on a dl to view ratio at all.

not alone, but it's one of the factors that counts

« Reply #11 on: December 03, 2008, 10:32 »
0
It should not count at all, because it does not say anything about the image.
1. There are interesting photos which are nice to look at, but do not sell at all = a lot of views and not much downloads
2. There are interesting photos, nice to look at, and people love to buy them = a lot of views and a lot of downloads
3. There are not so nice images to look at, but they sell great = low views high downloads
4. There are not so nice images to look at and they do not sell= low views, low downloads
5. There are featured images which lots and lots of views. Compared to that downloads are always low even if the image has thousands of downloads

So does the view to download ratio help in any way to find out which are the best images?

Not at all!

« Reply #12 on: December 03, 2008, 11:06 »
0
freezing :

Quote
This is an easy way for 'killing' spammers.  Their images will have many views (because they show up in many searches), but few DL (because no one like their image)

« Reply #13 on: December 03, 2008, 11:29 »
0
Maybe, but that images must be at least interesting to look at, because if a buyer searches and finds an image which is absolutely not relevant, why should he look at it (click on the thumnail which counts as a view)?

My guess is that top contributors can generate more views/image/dl, because of all their admiring fans and those who look for inspiration. So they automaticly are coming better up in searches. That might be an explanation why top exclusives seem to do better than others.

bittersweet

« Reply #14 on: December 03, 2008, 11:46 »
0
Maybe it is a plot to get the porn to the front of all the searches.  ;D

lagereek

« Reply #15 on: December 03, 2008, 11:48 »
0
Quite true!

I know for a fact, Ive got many regular buyers and Design-studios buying for a Corporate end-product.
Many of them have mailed me and on private mail.
This is nothing new. In the Trad-RM world it happens even more so.

Picture agencies whatever policy and feature can give great exposure.

shank_ali

« Reply #16 on: December 03, 2008, 13:30 »
0
I have a series of images of block paving.It use to be on page one of best match and sold 20+ in a month.Then best match changed and the image went bye bye.It shall return and it's swings and roundabouts.
If you want the benifits of being exclusive close ALL your other accounts and apply,if you don't but still want the same benifits ( a fair shake  in best match) as an exclusive contributor.TUFF S*** you aint going to get it( full stop)

« Reply #17 on: December 03, 2008, 15:35 »
0
Perhaps istock has decided to add a Christmas bonus for its loyal contributors (long-term exclusives) to the other perks. They must be doing pretty well this month.

« Reply #18 on: December 03, 2008, 18:50 »
0
For me its gone over the edge.

Not so long ago I had more than 10 sales daily on week days and about 5 at weekends.

I had one or two sales in last 10 days , both bellow 0,30 , haven't reach payment last month first time in few years, and if that continues that way I wonder if I will reach it in Decembers
cause those 3$ that I miss look huge in such short time.

Featurepics is earning me at least hundred times more than IS at the moment , I will wait a month or two and if things don't change , I will stop loosing my time there.






Uncle Pete

  • Evidence please...

« Reply #19 on: December 03, 2008, 19:40 »
0
Perrush, I agree with you, although I think you cannot make best match position dependent on a dl to view ratio at all.

not alone, but it's one of the factors that counts

Sounds reasonable that one of the factors is based on the photos download number, then views/downloads. Also someone could be raised in the best match by their totoal views and total sales, which have nothing to do with an individual photo.

To explain better. Someone who has higher sales of photos of secretary on the phone, happy business people shaking hands and person perplexed looking at a computer, will get a better rank for their photo of Red Rose, because of the other photos. Photographer rank being the additional factor that aids them.

This is why some of the Microstock Factories that produce highly researched and popular photos are going to be at the top of many of the searches for whatever they produce.

It's not just individual sales and individual photos.

There was a quote in an article on stock, that iStock and DT had both limited submissions to prevent a flood of photos from the Photo Factories, (my term, not theirs) big professional microstock producers. This was their way to make it fairer for the little people who may only produce 10 or 15 images a week. So they do care about preventing their photo choices overwhelmed by the large production companies. Nice touch.

We don't know, but a single photos views and sales aren't the only factors, and I'm suggesting that photographer's views and DL ranking, and whether they are exclusive or not, can also be playing a role here. Age of the photo could also be factored in, dropping lower in the search if it hasn't been pruchased in 90 days, which makes room for more new material, closer to the top.

« Reply #20 on: December 03, 2008, 20:58 »
0
Racephoto, true those are factors that taken into account too.  And you forgot the acceptance ratio  ;)

« Reply #21 on: December 05, 2008, 19:26 »
0
I told my thinking about iStock (which they name in my language is very close symbolism to Cattle, and with letter "i" in front means internet Cattle).
Thinking hmmm like Clint Eastwood says ...... about thinking or oppinion.
Last 2 years I dont expect anything positive news from them as nonexclusive contributor and with more and more uploads my stats are going down (on other sites are in positive growing trend).
I just dont like to sign my name like exclusive Cattle members on shoes how somebody will be able to recognize me by reading my name from my shoes while I am in the * of somebody which name means Cattle!!!
My best Christmas wishes for iStock (iCattle) is to drop down or bankrupt fastest with this kind of they "nepotism business"
Look at the bright side of life...........MP not at the cattle side of life.
 ;D


« Reply #22 on: December 06, 2008, 10:20 »
0
First let me say I am a small timer at istock with about 350 pics and I am not exclusive. That said I don't understand all of the agnst about Istock favoring their exclusives. These people agree to work with them only and provide exclusive content. Why wouldnt a company take care of it's in-house people first? It is also an incentive for others to sign on. I don't intend to go exlcusive and my sales have dropped but I am not offended by them taking care of the exclusives.

Uncle Pete

  • Evidence please...

« Reply #23 on: December 06, 2008, 16:19 »
0
First let me say I am a small timer at istock with about 350 pics and I am not exclusive. That said I don't understand all of the agnst about Istock favoring their exclusives. These people agree to work with them only and provide exclusive content. Why wouldnt a company take care of it's in-house people first? It is also an incentive for others to sign on. I don't intend to go exlcusive and my sales have dropped but I am not offended by them taking care of the exclusives.

I'd agree with you, no matter if I was exclusive or not. The people who put all their efforts into iStock should get some kind of favoritism for that loyalty. It's also makes sense that iStock wants more exclusives because generally all the same people and photos are on every other site, and from a marketing standpoint, IS can sell based on having unique images.

You need to read the other thread that claims the new search is favoring non-exclusives.  ;D

« Reply #24 on: December 10, 2008, 14:29 »
0
First, I think that the second image is much better than the first one. So it actually proves the point that better image will be selling better.
Second, why IS should NOT favour exclusives in best match? I mean - it pretty much obvious that they do but why is it wrong? Why everybody's eager to prove it like it's some kind of bad deed?

PhotoDuneMicrostock Insider

 

Related Topics

  Subject / Started by Replies Last post
9 Replies
2618 Views
Last post November 25, 2006, 10:26
by kosmikkreeper
62 Replies
9568 Views
Last post February 17, 2008, 04:55
by rjmiz
35 Replies
6641 Views
Last post November 23, 2008, 12:11
by hali
16 Replies
3278 Views
Last post December 01, 2009, 21:48
by Uncle Pete
54 Replies
7844 Views
Last post April 07, 2011, 17:22
by cathyslife

Sponsors

Microstock Poll Results

Sponsors