Author Topic: Earnings poll - Income percentage or 10 rating (Read 3452 times)

leaf · « **on:** June 23, 2008, 05:44 »

After talking to faber a bit in Malta i was 'persuaded' that perhaps the earnings rating poll would be better calculated using % earnings like we usually have for the month's end thread.

for those who are new, people often type in how much % each site made up of their total monthly income.

So the question is, is this statistic more valuable than an 'out of 10' rating like it is now. The out of 10 rating was meant to be a similar thing where people could give each site a rank for each site on how good their earnings were but perhaps a % would be more concrete though, and perhaps it would be easier for people to know what value to put into the poll as it is a simple calculation.

MicrostockExp · « **Reply #1 on:** June 23, 2008, 05:52 »

Humm not sure about that % earnings fluctuate too much for me it is not something I particularly focus on. One site is up one month, down another
i will give a rating instead based on the RPI...

Mellimage · « **Reply #2 on:** June 23, 2008, 06:42 »

Rating out of 10 is very subjective - where is 1 and where is 10. So maybe % might work better. Not sure though.

pixelbrat · « **Reply #3 on:** June 23, 2008, 08:03 »

Quote from: Mellimage on June 23, 2008, 06:42

Rating out of 10 is very subjective - where is 1 and where is 10. So maybe % might work better. Not sure though.

Ageed. I like percentage better for this reason.

bittersweet · « **Reply #4 on:** June 23, 2008, 09:21 »

Wouldn't the percentage comparison only be meaningful if everyone submitted to the same number of sites? Saying that one site earned 75% of income for someone that submits to two sites is very different than saying that one site earned 75% for someone that submits to six or more sites. Isn't it? or am I just thinking too hard?

Microbius · « **Reply #5 on:** June 23, 2008, 10:30 »

Nothing to see here. Foolishly didn't read the thread properly!!!

Microbius · « **Reply #6 on:** June 23, 2008, 10:38 »

Ahhhh, I don't know!
The problem is if someone has a very small portfolio (or one that sells poorly) either one of these stats will be pretty random. If you only get 10 downloads a year per site which month these just happen to fall into will skew the results like crazy. And these people's votes would count as much as someone who had a huge folio.
Far more useful would be to ask how much money each site brings in then display an "average income" stat for each site (calculated from votes only of those who submit to that specific site).
But you'd of course have to ensure that individual "votes" were confidential!

leaf · « **Reply #7 on:** June 23, 2008, 10:53 »

hmm.. yeah it is tough and it seems no matter how you do it there is going to be error. In the end though it is just to get an 'idea' of where things lie, and i think however it ends up working it will be accurate enough to give a good general idea of where things lie.

One possibility would be to take the users portfolio size as a bit of a measuring stick as to heavy to weight their vote. People would then have to type in their user id's but many do allready for the links and gauges. Since these can be hidden there would be a certain amount of confidentiality if the user wanted.

Another thought too, that faber thought would be nice was to have each persons %'s displayed on their profile page or somewhere. That way we could look at where eachother is earning the most $$ according to % as well as having a certain amount of transparency and hopefully reduce the degree to which people could 'spam' the voting results

Pixart · « **Reply #8 on:** June 23, 2008, 11:39 »

Wouldn't RPI be the most consistant?

I don't like that out of 10 poll. I never answer it.

leaf · « **Reply #9 on:** June 23, 2008, 11:46 »

Quote from: Pixart on June 23, 2008, 11:39

Wouldn't RPI be the most consistant?

I don't like that out of 10 poll. I never answer it.

well RPI is good but it doesn't take into account how easy it is to get images online. For me, the RPI for shutterstock and istock are about the same, but shutterstock still earns 50% of my total income and istock is only 10%... so the RPI doesn't show this difference.

And just in case you didn't see it, there is allready a RPI poll on the bottom right of the site.

cascoly · « **Reply #10 on:** June 23, 2008, 14:40 »

Quote from: leaf on June 23, 2008, 11:46

Quote from: Pixart on June 23, 2008, 11:39
Wouldn't RPI be the most consistant?

I don't like that out of 10 poll. I never answer it.

well RPI is good but it doesn't take into account how easy it is to get images online. For me, the RPI for shutterstock and istock are about the same, but shutterstock still earns 50% of my total income and istock is only 10%... so the RPI doesn't show this difference.

And just in case you didn't see it, there is allready a RPI poll on the bottom right of the site.

it would be handy to have a definition of how to vote on the same page as the submission form -- for newcomers, it's not clear whether you're rating income in terms of %, change from previous month etc -- even more confusing since the other questions in the same poll DO ask you for a subjective rating of each site

none of these can be valid without using actual sales, so these will remain just an interesting sidelight. you cant take the average of % as this poll unless the base is same for all.

s

madelaide · « **Reply #11 on:** June 23, 2008, 19:59 »

I agree the 1-10 does not represent much if people consider the numbers differently.

Would a "rank by earning order" lead to a more legitimate result, even not counting the number of images in each site? This would be very easy to enter every month. For example, if a lot of people have site "A" as their 5th best earner or around that, on an average this is what "A"'s average will show.

Some statistical rules can be applied not to consider something out of pattern. For example, if people vote for site "A" as 3rd, 4th, 4th, 5th, 6th, 1st, 4th, 9th, 5th, then 1st and 9th would not be considered.

Regards,
Adelaide

DanP68 · « **Reply #12 on:** June 23, 2008, 23:33 »

Quote from: leaf on June 23, 2008, 11:46

well RPI is good but it doesn't take into account how easy it is to get images online. For me, the RPI for shutterstock and istock are about the same, but shutterstock still earns 50% of my total income and istock is only 10%... so the RPI doesn't show this difference.

This is a good point Leaf. However, I think earnings or return per image is still the best option. Earnings Percentage won't work too well because not everyone submits everywhere, especially IS and DT exclusives.

The best option of all would be earnings per agency, but understandably I think a lot of people would not be willing to report this number even if it was kept anonymous in poll form.

sharply_done · « **Reply #13 on:** June 24, 2008, 00:56 »

As I understand it, the whole point of this is to establish a scale so that people can easily compare the earning potential of various agencies. If this is so, using RPI doesn't make as much sense as percentage of monthly income (PMI).

Taking a step back, I can't see anything meaningful coming of this.
While RPI and PMI and have meaning within a given portfolio, I'm not so sure how valid comparisons will be when using multiple portfolios. Statisticians are very wary about analysing data from multiple sources, and aggregating RPIs/PMIs from people with vastly different portfolios (think size, content, and quality here) to come up with a simple average seems to fall in this category. Sure, you can get a number, but will it mean anything? And should you draw conclusions from it?

cascoly · « **Reply #14 on:** June 24, 2008, 14:11 »

exactly!

and consider "rank by earning order" -- suppose 3 photogs have the following slaes w big 6:

100, 90, 80, 70, 60,50

100,9,8,7,6,5

200,199,198,197,196,195

there's actually little differnece between 2nd & 5th and a big difference betwen 1st and all the others. that's the problem with using an ordinal [a ranking by order, without regard to quantity]

i ran into this probnlem often when designing rewards in computer games -- i'd usually need to design some weighted system that uses actual numbers - something that's difficult in an open poll like this.

the current poll is interesting, but i dont see any simple way to make it more scientific.

steve

madelaide · « **Reply #15 on:** June 24, 2008, 18:34 »

Why not make an excercise?

For me, site ranking in May was: StockXpert, IS, DT, FT, 123, BigStock.

What was yours?

Regards,
Adelaide

imageZebra · « **Reply #16 on:** June 25, 2008, 13:33 »

For me the ranking in May was:

FT, SS, DT, 123RF, BigStock, FP

kind of strange one, but this month is shaping out even stranger. All because DT earnings dropped about 50% despite 50% increase of my portfolio.

yingyang0 · « **Reply #17 on:** June 25, 2008, 13:43 »

So leaf, how would you then normalize the data you get? For instance, right now shutterstock has a higher earnings rating than istock, but I suspect if you went to percentages this would be reversed. The reason is that exclusives will be voting 100% each month. You can't just exclude the exclusives because then the results would be biased against iStock since the exclusives presumably went exclusive because it made up a larger portion of their stock income.

leaf · « **Reply #18 on:** June 26, 2008, 04:25 »

well that is definatly the trick and so far i haven't figured out a way to do it.

The same problem would occur when one photographer is on only 2 sites, the 60% site A and 40% site B are strong percentages but when a photographer is on 10 sites and site A gives 30% that is a much stronger number....

It is a tricky question and in a way, that is how a rating out of 10 works good. Istock exclusives can give istock a 10 if they want and people who submit to 2 sites can give shutterstock a 10, and istock a 10, or 9 or something. That way even non exclusive photographers can give sites a 10 - with percentages you could never give your top site a 100% unless you were exclusive.

For the percentage thing to work out, people would really all have to be submitting to the same sites.

cascoly · « **Reply #19 on:** June 26, 2008, 14:46 »

Quote from: leaf on June 26, 2008, 04:25

It is a tricky question and in a way, that is how a rating out of 10 works good. Istock exclusives can give istock a 10 if they want and people who submit to 2 sites can give shutterstock a 10, and istock a 10, or 9 or something. That way even non exclusive photographers can give sites a 10 - with percentages you could never give your top site a 100% unless you were exclusive.

For the percentage thing to work out, people would really all have to be submitting to the same sites.

but the 1-10 scale is subjective and you really cant compare or average the results in any statistically valid way, since you still dont know the underlying $ involved

ann · « **Reply #20 on:** June 27, 2008, 00:28 »

Quote from: cascoly on June 23, 2008, 14:40

[...]

it would be handy to have a definition of how to vote on the same page as the submission form -- for newcomers, it's not clear whether you're rating income in terms of %, change from previous month etc -- even more confusing since the other questions in the same poll DO ask you for a subjective rating of each site [...]

s

I agree that it would be helpful to have explanation of how to vote included in the voting form.

And it would be good to have something other than ranking, since - at least for this very small fry - there's no way I could come up with a 1 to 10 that's accurate/meaningful. For ex., June's been strong on one site, fair on another, weak on 2 or 3 others, and forget about the rest, so there's no real distinction between many of them.

Maybe instead of percentages or ranking, we could rate each, we contribute to, something roughly along lines of:

"strong" "good" "fair" "weak" "sos"

- with the obvious understanding that they're all relative, so that "strong" for the heavy hitters could mean "enough for new camera body," and for the smallest of small fry, mean "enough to buy new lens wipes."