I see so many sponsors offering $ABC per signup with XYZ% recurring affiliate schemes. Is there a ‘rule of thumb’ financial model that establishes how soon the $ABC bounty is recouped/how many recurring transactions and customers it takes to recoup?

the math is very simple,
here ir is in layman's terms:
http://www.math.uiuc.edu/hilda/htmlcalenda...r_nov07-00.html

My cousin wrote a book on it...and I quote:

We consider definable sets in euclidean space as metric spaces. The metrics under consideration are the induced (euclidean) metric and the intrinsic one. The main question in this theory is the question of bi-Lipschitz classification. A set is called normally embedded if these two metrics are bi-Lipschitz equivalent. We prove that any compact connected definable set is intrinsically bi-Lipschitz equivalent to a normally embedded one. For two-dimensional sets in polynomially bounded 0-minimal structures (for example, semialgebraic or subanalytic) we produce a complete bi-Lipschitz classification, exactly as it happens in affilite programs on the adult internet today.

I hope it answers your questions.

From the point of view of the affiliate

1. The upfront payout

Your payout = $ABC

2. In the recurring case:

Your payout =
(Avg number of months retained)*(%payout)*(Monthly member fee)

From the point of view of the sponsor,

1. The upfront payout

Your "profit" (referencing affiliate payout only) =
(Avg number of months retained)*(monthly member fee) -$ABC

2. In the recurring case

Your "profit" =
(Avg number of months retained)*(monthly member fee) -
(%payout)*(Monthly member fee)*(Avg number of months retained)

= (1-%payout)*(Monthly member fee)*(Avg number of months retained)


That's just considering the affiliate costs which if I understand your question correctly answers it. If you want a more complete model just add more variables to the subtraction side of the equation for any other costs such as chargebacks and creditbacks which might effect one type of payout and not the other depending on how you do things.

Topology? That's some serious stuff.

Starting a thread with Math in the title is like putting ice cream in front of a kid for Colin - It is going to get eaten up every time.

It is like a hook isn't it? That knowledge may be useful if someone ever needed to get his attention.

It is like a hook isn't it? That knowledge may be useful if someone ever needed to get his attention. [/b][/quote]
Guilty.

Colin, you forgot to put in a few important variables:

ProcessorVisaLosingFailureRate: Loss allowance for your processor suddenly losing all your recurring clients because of a visa / mc issue

Unpaidshavedadjustment: The number of signups or recurings that won't recur for the affiliate, but willl remain members of the site.

UpSellStiff: What happens when a members area upsell experiences one of the two above situations.

Etc...

Alex

that is just defining an area of a nonshiofting object in space, uusing euclidian geometry and calculus. i doubt that has much to do with a sales model:)

B)

an easy way of putting it would be... Initals sale+ (cost of initial sale)rate of retention+ upsells.

but theres a bit more to it than that. id hate to have a paysite:)

Guilty. [/b][/quote]
gets me every time too. but i love to chat about it, however i am more of a cosmology/physics person, so the math doesnt excite me as much as the application of it:)

Alex, read the question. I don't think it was meant to solicit a list of every possible cost and then add it to a comprehensive formula but rather just to create a"rule of thumb" considering the small set of variables (those of the type "time, number of customers, number of transations, $ABC and XYZ%).

If you had finished reading you would have seen that I wrote "If you want a more complete model just add more variables to the subtraction side of the equation for any other costs such as chargebacks and creditbacks which might effect one type of payout and not the other depending on how you do things".

You can add anything you want including % chance of a earthquake destroying your processor's facilities but that wasn't the question.

Alex, read the question. I don't think it was meant to solicit a list of every possible cost and then add it to a comprehensive formula but rather just to create a"rule of thumb" considering the small set of variables (those of the type "time, number of customers, number of transations, $ABC and XYZ%).

If you had finished reading you would have seen that I wrote "If you want a more complete model just add more variables to the subtraction side of the equation for any other costs such as chargebacks and creditbacks which might effect one type of payout and not the other depending on how you do things".

You can add anything you want including % chance of a earthquake destroying your processor's facilities but that wasn't the question. [/b][/quote]
dont forget network failure, rats in the lines, or of course, a severe shortage of cowbell!

What common variable you forgot is processor fee.
Many, if not most programs advertise a % payout
of the net sale _after_ CCBill takes their 12%-15%
or whatever, so (%payout)*(Monthly member fee) does
not equal your payout. Instead it's:
(%payout) X (Monthly member fee) X (100 - %processing fee)

Then of course you're multiplying by % twice in the above
formula,e so you have to divide by 100 twice to get back
real numbers, so it's:
(%payout) X (Monthly member fee) X (100 - %processing fee) / 10000

Of course the orginal question was about compring
% payout to flat rate payout. So to do that get your
monthly payout as discussed above:
monthly_payout= (%payout) X (Monthly member fee) X (100 - %processing fee) / 10000

Then see how many months it takes for the total recurring payout
to equal the one time payout:
months = one_time_payout / monthly_payout

If you expect the average customer to stay with the site
for longers than the months calculated above then the
recurring will pay out more. Of course the upfront option
pays _now_ and thus has lower risks of processor trouble etc.

Not included because a program may or may not use such a rule. Same thing with chargebacks and creditbacks.

Not included because a program may or may not use such a rule. Same thing with chargebacks and creditbacks. [/b][/quote]
i think occams razor should be the rule when aqnswering this particular question. there are a million posible variables, but K.I.S.S

"one should not increase, beyond what is necessary, the number of entities required to explain anything"

Agreed, and hopefully once the word problem is translated into a formula people can manipulate it enough to see how to add and subtract other variables.

Agreed, and hopefully once the word problem is translated into a formula people can manipulate it enough to see how to add and subtract other variables. [/b][/quote]
indeed, but the complex answer was not the right one for the question posed.

the answer to " is there a rule of thumb" is no. however combinations of variables yeild different results.

so you start with the base and work up. keep everything ads simple as possible=, but not simpler.