JeffGreenberg

workflow: how to determine one's weakest pseudo?

Recommended Posts

Posted (edited)

If one has multiple pseudos, say 6 or 10, & all appear in several searches,

one would expect them to always appear in returns in similar order,

but I don't find that to be true -- do others???!!

So how to determine one's weakest pseudo amongst many pseudos?

Ranking by CTR doesn't match search returns either...?

Clearing out all fields & insert multiple nonsense words, same order,

in one image per pseudo & then search each nonsense word?

 

So how?  HOW?!!!

Edited by JeffGreenberg

Share this post


Link to post
Share on other sites

I have at least one pseudo that has never had a single zoom, but in my better performers, the rank of the individual images counts too. Each photo on Alamy carries its own history that effects its placement in searches and I believe that this is what you're seeing. Clearing out fields doesn't delete the history that Alamy stores with it.

Share this post


Link to post
Share on other sites
Posted (edited)
4 hours ago, JeffGreenberg said:

If one has multiple pseudos, say 6 or 10, & all appear in several searches,

one would expect them to always appear in returns in similar order,

but I don't find that to be true -- do others???!!

So how to determine one's weakest pseudo amongst many pseudos?

Ranking by CTR doesn't match search returns either...?

Clearing out all fields & insert multiple nonsense words, same order,

in one image per pseudo & then search each nonsense word?

 

So how?  HOW?!!!

 

Here's my suggestion.

 

Upload a number (=no of pseudos you have) of new images in a single submission, with all images taken on the same date.

Once they pass QC, make a note of their Alamy refs, hopefully they will have first 3 or 4 letters the same to help find them in the search results.

Allocate identical keywords to each one including a reasonably popular term (so that ~1,000 images will be returned in an Alamy search).

Allocate one image to each of your pseudos.

Wait 24 hours for search database to update (don't wait too long to  minimise chance of views or zooms occurring)

Carry out an Alamy search using the keyword.

Set page size to 100 images per page and use <Cntrl -> to get 10 images per line (helps count image position)  

Now locate each of your images amongst the ~1,000 images returned. (Use <Cntrl  F> to look for first 3 or 4 letters of Alamy ref on each page).

 

Why 1,000 images? It's a tradeoff between ensuring that there are plenty of other other contributors images for your to be scattered amongst and the ease of finding your own images. If you're happy to spend longer finding your images, then pick a more popular search term/subject.

 

Mark

 

 

Edited by M.Chapman

Share this post


Link to post
Share on other sites
Posted (edited)
5 hours ago, JeffGreenberg said:

If one has multiple pseudos, say 6 or 10, & all appear in several searches,

one would expect them to always appear in returns in similar order,

but I don't find that to be true -- do others???!!

So how to determine one's weakest pseudo amongst many pseudos?

Ranking by CTR doesn't match search returns either...?

Clearing out all fields & insert multiple nonsense words, same order,

in one image per pseudo & then search each nonsense word?

 

So how?  HOW?!!!

 

 

Why do you expect that to be the case?

 

If the images were allocated randomly to these pseudos - simply divided up with no distinction between them - then they may well all be approximately the same strength in terms of Alamy rank and so search order position is also random. Each time it could be much like rolling dice because each pseudo is equally balanced. Something has to turn up first but there is no reason to expect consistency.

 

I have two main pseudos deliberately created to be different. 'geogphotos' has all images that have ever sold, been zoomed, or selected by independent external editors. 'geogphoto' has all the rest. In every search I do 'geogphotos' comes first. In the other words - my 'dice' are loaded. :)

 

I suspect that when people suggest that pseudos don't count any more it is because their pseudos are much the same in terms of Alamy Rank and so that  inherent randomness is displayed in their search results.

 

I am not suggesting that this is your situation because I don't know. But to return to my original question - why do expect to see a similar order?

Edited by geogphotos

Share this post


Link to post
Share on other sites
2 hours ago, M.Chapman said:

 

Here's my suggestion.

 

Upload a number (=no of pseudos you have) of new images in a single submission, with all images taken on the same date.

Once they pass QC, make a note of their Alamy refs, hopefully they will have first 3 or 4 letters the same to help find them in the search results.

Allocate identical keywords to each one including a reasonably popular term (so that ~1,000 images will be returned in an Alamy search).

Allocate one image to each of your pseudos.

Wait 24 hours for search database to update (don't wait too long to  minimise chance of views or zooms occurring)

Carry out an Alamy search using the keyword.

Set page size to 100 images per page and use <Cntrl -> to get 10 images per line (helps count image position)  

Now locate each of your images amongst the ~1,000 images returned. (Use <Cntrl  F> to look for first 3 or 4 letters of Alamy ref on each page).

 

Why 1,000 images? It's a tradeoff between ensuring that there are plenty of other other contributors images for your to be scattered amongst and the ease of finding your own images. If you're happy to spend longer finding your images, then pick a more popular search term/subject.

 

Mark

 

Looks like a good  setup.

Maybe take all images in Miami including a number. Something colorful; big numbers. ABC would be another possible subject.

My guess is though that this will just replicate your BHZ result.

 

wim

 

wim

 

 

Share this post


Link to post
Share on other sites
5 hours ago, geogphotos said:

I have two main pseudos deliberately created to be different. 'geogphotos' has all images that have ever sold, been zoomed, or selected by independent external editors. 'geogphoto' has all the rest. In every search I do 'geogphotos' comes first. In the other words - my 'dice' are loaded.

If an image is zoomed in geogphoto and you transfer it to geogphotos,  the image will still retain it's geogphoto pseudo zoom tag

Share this post


Link to post
Share on other sites

why do expect to see a similar order?

 

Note:  I asked about WEAKEST pseudo(s).

My STRONGEST few pseudos are apparent, always appear before my other pseudos,

despite my images being distributed randomly amongst pseudos.

Its determining the weakest of my weak that I seek,

to possibly replace them with FRESH neutral strength pseudos...?

(or if ranking strength is now associated ONLY with images, not pseudos, then...)

 

My guess is though that this will just replicate your BHZ result

 

I wonder if BHZ ranking would more accurately show one's pseudo ranking

strength if one had only BHZ as supertag & had NO OTHER supertags or

tags or caption?  Right now I have BHZ as the only supertag, but have true tags

& true captions for my BHZ images as I still want some chance of them licensing...

 

 

 

 

Share this post


Link to post
Share on other sites
2 hours ago, LawrensonPhoto said:

If an image is zoomed in geogphoto and you transfer it to geogphotos,  the image will still retain it's geogphoto pseudo zoom tag

 

But if it has been zoomed it must, by definition, be an image that is of interest to at least one accredited buyer and therefore I place it in my strongest pseudo.

Share this post


Link to post
Share on other sites
2 hours ago, JeffGreenberg said:

why do expect to see a similar order?

 

Note:  I asked about WEAKEST pseudo(s).

My STRONGEST few pseudos are apparent, always appear before my other pseudos,

despite my images being distributed randomly amongst pseudos.

Its determining the weakest of my weak that I seek,

to possibly replace them with FRESH neutral strength pseudos...?

(or if ranking strength is now associated ONLY with images, not pseudos, then...)

 

My guess is though that this will just replicate your BHZ result

 

I wonder if BHZ ranking would more accurately show one's pseudo ranking

strength if one had only BHZ as supertag & had NO OTHER supertags or

tags or caption?  Right now I have BHZ as the only supertag, but have true tags

& true captions for my BHZ images as I still want some chance of them licensing...

 

 

 

 

 

 

Jeff, 

 

Please excuse me offering a few thoughts in the form of questions:

 

What is your objective in terms of having pseudos?

Why/how did you decide on the number that you have?

What is wrong with having a 'weakest' pseudo?

Isn't it inevitable that you are always going to have a weakest pseudo?

Why is it a concern for you to identify the weakest pseudo/pseudos - if you were to elevate them in searches would they not then compete with you and damage your higher pseudos?

 

All offered with the best of intentions and upmost respect.

 

Ian

Share this post


Link to post
Share on other sites
Posted (edited)

a.>What is wrong with having a 'weakest' pseudo?

b.>Isn't it inevitable that you are always going to have a weakest pseudo?

c.>Why is it a concern for you to identify the weakest pseudo/pseudos -

> if you were to elevate them in searches would they not then compete with you and damage your higher pseudos?

=====

Can answer these:

a. nothing

b. yes, unless two or more are tied for weakest

c. my pseudos exist amongst hundreds or thousands of other images;

is it not possible that one's weakest pseudo can be reintroduced as

a NEW neutral strength pseudo without pushing down one's next weakest pseudo???

 

Which analogy better describes what will happen:

 

I am a juggler of, say, 10 pseudo balls;

am I standing on ground & lowest pseudo ball always at same height (as you suggest)

or am I slowly ascending a set of steps leading to a ceiling (top half page 1) & therefore

lowest pseudo ball is always rising until highest pseudo ball is bouncing off

ceiling & I elect to go no higher....?

 

Edited by JeffGreenberg

Share this post


Link to post
Share on other sites
Posted (edited)
1 hour ago, JeffGreenberg said:

 

c. my pseudos exist amongst hundreds or thousands of other images;

is it not possible that one's weakest pseudo can be reintroduced as

a NEW neutral strength pseudo without pushing down one's next weakest pseudo???

 

 

 

 

 

In that case won't they offer more competition to the images in your higher ranking pseudos and, logically, have a negative effect on them and drag them down? 

 

You did not answer my first two questions - about what you are trying to achieve.:P

 

Using your juggling analogy and if it was possible to freeze-frame. I think you have to decide which ones to fix in the air and in what positions you want them. Otherwise as one goes up the other goes down, and nobody can have all their balls at the top at the same time. 

 

Meanwhile there are 70,000 other people standing in a long line juggling balls in the air.:)

 

 

Edited by geogphotos

Share this post


Link to post
Share on other sites
10 hours ago, wiskerke said:

My guess is though that this will just replicate your BHZ result.

 

Only if BHZ is applied to images in each pseudo where all other factors have been equalised. i.e. Images were taken at the same time uploaded in the same batch,  have identical caption, tags and supertags and they haven't had differing number of views, zooms or sales. This effectively means they need to be freshly uploaded images.

 

NB. I don't know for certain that all these factors affect image position, but in discussions with Alamy at the Photography Show at the NEC in March they mentioned that a variety of factors can be taken into account, including the age of the image and the level of interest shown in it by customers. But they also emphasised that different criteria may be applied depending on the search and even the customer. For example, the age of a photo may be taken into account for searches of popular travel locations, but not for searches of food. However, they also said that they frequently tweak the algorithms. So what's true at one time may not be true later. I'm pretty sure I see some evidence for date taken affecting some of my travel shots (but not others).

 

My suggested test therefore tries to exclude as many of these factors as possible from the results to leave only the Pseudo's rank. 

 

Mark

Share this post


Link to post
Share on other sites
17 minutes ago, M.Chapman said:

 

Only if BHZ is applied to images in each pseudo where all other factors have been equalised. i.e. Images were taken at the same time uploaded in the same batch,  have identical caption, tags and supertags and they haven't had differing number of views, zooms or sales. This effectively means they need to be freshly uploaded images.

 

NB. I don't know for certain that all these factors affect image position, but in discussions with Alamy at the Photography Show at the NEC in March they mentioned that a variety of factors can be taken into account, including the age of the image and the level of interest shown in it by customers. But they also emphasised that different criteria may be applied depending on the search and even the customer. For example, the age of a photo may be taken into account for searches of popular travel locations, but not for searches of food. However, they also said that they frequently tweak the algorithms. So what's true at one time may not be true later. I'm pretty sure I see some evidence for date taken affecting some of my travel shots (but not others).

 

My suggested test therefore tries to exclude as many of these factors as possible from the results to leave only the Pseudo's rank. 

 

Mark

 

 

This suggests that there is data 'attached' to each individual image or am I misinterpreting?

Share this post


Link to post
Share on other sites
10 hours ago, geogphotos said:

 

 

This suggests that there is data 'attached' to each individual image or am I misinterpreting?

 

That was my interpretation.

 

"Age of the image" could be upload date and/or shooting date which we know are stored with each image.

"The level of interest shown in it by customers".... that's much harder to know exactly what this means. If interest shown by customers is broadened to mean "interest in images from a contributor" then it could simply mean a pseudo's rank is taken into account, which is based on CTR which could be called "interest"?

 

However, my feeling is that zooms, downloads and sales are probably stored on an image by image basis so could be being used. I'm not sure about views. Like Wim, I also have seen cases were images that have sold may have been demoted. But with all the other (hidden) factors at play it's hard to know for sure. 

 

Mark

Share this post


Link to post
Share on other sites
Posted (edited)
27 minutes ago, M.Chapman said:

 

That was my interpretation.

 

"Age of the image" could be upload date and/or shooting date which we know are stored with each image.

"The level of interest shown in it by customers".... that's much harder to know exactly what this means. If interest shown by customers is broadened to mean "interest in images from a contributor" then it could simply mean a pseudo's rank is taken into account, which is based on CTR which could be called "interest"?

 

However, my feeling is that zooms, downloads and sales are probably stored on an image by image basis so could be being used. I'm not sure about views. Like Wim, I also have seen cases were images that have sold may have been demoted. But with all the other (hidden) factors at play it's hard to know for sure. 

 

Mark

 

So we don't know for sure what happens when an image changes from one pseudo to another - does its 'juice' stay with it at an individual image level or does it get left behind with the pseudo? 

 

Please see related 'Supertag Experiment' for further discussion of this.

Edited by geogphotos

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now