Quote from: ShadySue on September 20, 2013, 19:45

Also, I don't understand their almost total abandonment of checking for accurate keywording. Almost every search you sort on 'New' is full of the most awful irrelevances, mistagging and deliberate spam.

: From their point of view (and therefore ultimately for us too) it probably doesn't matter provided that over time keyword relevancy can ultimately sort the wheat from the chaff. The images which get clicked and bought gaining relevancy over the years.

OK, I accept that that's your opinion and we'll have to agree to differ.

If that's what their policy now is, why haven't they told us? They did announce that their acceptance criteria were going to be relaxed.

Anecdotally, I uploaded a photo earlier in the week and by the time it hit the database it was blow position 1000 on its most important keyword, which was in the first position. It will hardly have any chance to rise. Of course there are modifying keywords, but they have said that most buyers don't modify their original search.

The system of 'rewarding' and 'punishing' keywords seems to be extreme. I'm not the only one to have noticed that any file uploaded in the past so many months seems to have punished in the best match when they get any number of sales under ten. I discovered something yesterday. I posted already that I had a mystery sale which came through with the GI sales as PP, although I have no images in the PP (and have taken out a ticket to query it). Anyway, that pic had had over 300 views but only one sale. The night the sale was announced, I looked at it, and noticed a 'surprising' collection of similars, which I think was a result of a spammed series by one person. I noticed that what I would consider the 'most important' keyword phrase was in position 1. The next morning, I noticed that the 'similars' were totally different - and found that that same keyword phrase was now in position 10 of 16. As it happens, the top two keywords are still relevant, but it seems like a huge punishment on just one sale.

Of course, it can be argued that with a better-selling file, that sort of thing will be evened out, but still, it seems like a sledgehammer to crack a nut, and of course any file so affected will sink far lower in best match.

Of course, it may be that someone who searches on apple will find that they actually wanted a photo of a spanner (imaginary example, but based on some equally bizarre examples), but it's more likely that someone making a search by 'new' or even 'popular' will just be hacked off by irrelevance.

That's just my opinion of course. If they want contributors to throw in some random group of keywords with any image, they should tell us. Not that I'd do it anyway.

So what are you saying ?

Their system is broken [1], has been for virtually a year, and they show no inclination to fix it, meaning there is no incentive to spend money, time or effort on producing new imagery, other than newbies who don't know any better.

It's a very odd scenario, but I'm prepared to accept that it could be part of some Plan. Though what use many thousands of irrelevantly-keyworded files will be to them, I couldn't begin to imagine.

[1] i.e. how they say it's supposed to work, with proper keywording and best match.

If it was a small boutique collection with a few hundred contributors then it would make sense to build it around a much more intensive and detailed inspection process. But iStock is much more like Flickr with thousands of people uploading huge quantities of just about anything. Lots of it probably irrelevant. As at Alamy I don't believe it can make sense for the inspection process to be labor-intensive.

It gives them a pool of content all of which can be quickly on sale and from which they can further choose content to promote. It's not stupid.

If it was a small boutique collection with a few hundred contributors then it would make sense to build it around a much more intensive and detailed inspection process. But iStock is much more like Flickr with thousands of people uploading huge quantities of just about anything. Lots of it probably irrelevant. As at Alamy I don't believe it can make sense for the inspection process to be labor-intensive.

It gives them a pool of content all of which can be quickly on sale and from which they can further choose content to promote. It's not stupid.

If they are about to drop all their prices, then saving money on almost all admin functions makes some sort of sense.

The comparison with Flickr is irrelevant IMO. I upload pics to Flickr I'd never expect to be on sale anywhere, and I'm sure thousands of others do too. Most relevant people know the difference between a photo sharing site and a stock agency.

Keywordzilla must be crying him/herself to sleep every night. They correctly corrected my stupid tick on Photograph (art and craft) instead of Photography (image) as a generic keyword. I have no issue with that.

I doubt they care even slightly. People learn from a thing, learn what works and does not, earn a wage and move on. What they probably learned from trying to fix the keywords was that it was a never ending uphill task. It was presumably useful even if only perhaps to show that it was not the best way of doing things.

Trying to manually fix the keywording spam in millions of images, with thousands more arriving every week, is maybe stupid and pointless. Perhaps it was the wrong solution.

Perhaps that is why they now seem to have gone with a solution which seems to be about keyword relevancy in which the buyers decide over time and contribute to the pool of data with their clicks and buys. I hope so - because that makes sense. And that seems to be the message.

There is no suggestion that keyword spamming gives anyone an advantage and every reason to assume that accurate keywording is to everyone's best advantage still.

the New search being so incredibly poor in many cases means that buyers won't use it.

You have no evidence to support this statement. Perhaps we can assume it will be used less but without stats we can only guess. Also - in many cases it will not be so incredibly poor. Granted it may be more erratic - but it will also be fresher. And it may sometimes provide an element of randomness which some searchers appreciate. We cannot make assumptions and we have no data.

�Captain, we�ve hit an iceberg, we�re sinking and don�t have anything like enough lifeboats, what are we going to do?�
�How about we rename the ship �Titan� and put a big dot at the end�

Love it!