The numbers from the Poll Results... what do they mean / represent?

I certainly would not make any binding decisions on where to place my imagery based on what you see in the very flawed results you see in the poll.

The numbers represent the average income (in dollars) of the number of contributors who report for that agency.

That's definitely not it (viz it's definitely not dollars per month income), though I can't remember what it is. IIRC it's a percentage of 100, but what the 100 originally represented, I either never knew or can't remember.

This is what Leaf said:
http://www.microstockgroup.com/newby-discussion/explanation-of-the-poll/msg412459/#msg412459

I may have heard this in a dream, but I'm sure somebody once said that if an agency has a score of 100, then it means everyone averages $500 a month from them. So for example, iStock is 25 so the average is $125 a month.

Like I say... I may have completely made that up.

I think that's right, though the different earnings bands in the survey make even that dubious.
In favour of the poll, it does seem to me to offer a reasonably accurate indication of how lucrative a site can be at least at the level of telling you that SS is more promising than Alamy, and Alamy does better than Bigstock. However, different sites are into different markets, so your pictures might suit one of them better than another, regardless of ranking.

I may have heard this in a dream, but I'm sure somebody once said that if an agency has a score of 100, then it means everyone averages $500 a month from them. So for example, iStock is 25 so the average is $125 a month.

Like I say... I may have completely made that up.

I don't think you made that up - if you did, then I had the same dream!  I remember $500 figuring in there somehow and it being a percentage relative to a fixed amount.  Or something.

I think the only way to sum it up is that: higher numbers = good; low numbers = not so good; no numbers at the bottom = variable; and specific numbers = meaningless.

All of the above, that looks at the poll has been pretty accurate. I'm not going to pick each one or quibble.

Yes it is roughly 20% of the actual numbers that people have reported. No one made that up and no one is dreaming. 

True there are ranges, but up or down rounding should even out. It's unnecessary to have every exact dollar amount.

The reason Leaf gave for not using actual numbers was just the part that some people will complain about. It's not meant to be what you will earn, or dollars, it's representative of people who took the pole.

People take the poll on a voluntary basis, some months I might, others I might not. True people could lie, but they can on any poll, anywhere.

I have nearly 5,000 files on SS and many fewer on AdobeStock, a couple thousand on Alamy or iStock. Many less on DT. Without all the details, SS and Alamy take Editorial, Adobe takes some of my illustrations that SS doesn't. I'm an audio exclusive on iStock/Getty. My only video at this point is on SS. The same can be true for anyone else, including someone with a very big, strong, model released collection that isn't on iStock at all.

For the same reasoning it's also relative, which is what we should be looking at. Roughly, people might expect to make 10 times more on SS than DT. Assuming that same person has all the same files on DT as SS?

Here's something from way in the past and I remember it from back when we could see earnings on IS. The poll represent the upper 5% of all people in Microstock. When Leaf does the annual pole you will see similar results. When we could see how many photos artists had, the same was true. Some people who take the poll are capped because the poll doesn't allow them to put in their full amount. That's an effort to smooth the results I'm pretty sure. You can't enter higher than $2,500+ as the top. I know a couple of people who make more than that some months on SS.

The cap also holds down some IS exclusives, I'd guess.

Notice the number of people at some of the agencies is nearly identical. We don't know if that's the same people, but it's logical that it is. SS, IS Adobe, 123RF, and DT are in one group. Alamy, DP, BS, CS are the second similar size group of results. Not sure if I should have included Alamy, they are different. And Pond5 is more likely to be video contributors.

All the rest don't have the 50 votes needed to show in the poll, but are ranked in order of earnings. You can't get in ClipartOf, Stocksy is a high end closed shop, Canva dropped many people, self hosted is what it is. Mostphotos ranks higher than Deposit, but only has 25 votes. Worth a look or is there someone who has a huge collection there making big numbers, moving the average way up?

A more complicated poll, and this is not in any way a complaint, would yield more interesting results. For example, simple but something someone could use, just as this one, for evaluation. How many photos do you have on agency 1, how much did you make last moth from agency 1, what is your RPD on agency 1. Now we'd see something based on numbers. But say person A is like me, and I do a shoot and upload 3-5 best photos, person B uploads 80 from his project of sliced vegetables isolated on white, person c has models and sets and uploads 80 individual shots, with four changes of cloths and background changes.

We all do not have the same interests, shots, materials, numbers or styles. Some are video only some are photo only, some people make vector illustrations primarily and some do all of the above. Oh yes at there's audio. Between that, the number of files, the agencies we choose and all kinds of other variables, the simple poll tells us a simple answer without over complicated attempts to over analyze what we will make.

Back to what I think the poll represents, relative potential of each agency, compared to each other agency. Nothing more. Not how much you can expect to earn, because we have no way of knowing who's taking the poll or if they are the same at all of the agencies, or their materials are the same at all. And don't forget that years ago, the numbers showed that the iStock members were the top 5% of the earners on IS.

You might see that as some really good and dedicated artists are here, on the forum and they take the poll. We don't know who or what, but the numbers are just a guide, not something that says "you will make this".