The payout is the single most consequential field on a campaign. Set it too low and the campaign is live, technically, and nobody sees it. Set it too high and you buy users who cost more than they bring.
There is a right way to find the number, and it starts at the far end.
Start from what a user is worth
Take the users you already have from any channel and work out what one is worth to you over a sensible window — thirty, sixty or ninety days. Revenue, not vanity. If you cannot calculate that yet, you are not ready to price a campaign; run a small one at a defensible number and use it to find out.
Then decide what fraction of that value you are willing to spend to acquire one. That is your ceiling.
Then discount for the channel
A rewarded user is not a user from your best channel. They came for the currency, and a proportion of them will never open your app again. Assume they are worth less, and buy them anyway if the arithmetic still works — the volume is real and it is fast.
The honest way to run this is to hold the model constant and let the price tell you: start at a fraction of your ceiling, measure retention against your other channels, and adjust. Two campaigns at different payouts on the same offer will tell you more in a week than any amount of modelling.
The floor nobody mentions
Your payout gets split before it reaches the user. The publisher takes their cut and rewards the user with the rest, in their own currency.
So the question that decides whether your campaign is ever shown is not "is this a good price for me" — it is "is what is left worth a user's time, and is the publisher's share worth their screen space?"
A wall sorts by what it earns. Your offer competes with every other offer for the same slot, and one priced below the going rate for that action simply loses that competition, every time, silently. There is no error message for "too cheap".
If a campaign is approved, funded, in the right countries and still not delivering, raise the payout before you change anything else.
Pay for difficulty
Price the effort, not just the outcome. An install with no requirement, a registration, a level 5 and a deposit are four completely different amounts of user time, and paying the same for all of them means the easy ones flood in and the valuable ones never happen.
A rough ordering that holds on most walls: a one-click or survey completion is cents; a straightforward install or sign-up is tens of cents to a couple of dollars; a game engagement with a real requirement is several dollars; a deposit or a purchase is tens.
Set the budget separately from the price
Payout is the price of one conversion. The daily and total budgets are how much you are willing to spend at that price, and the caps are how many conversions you will take.
Keep them independent. Under-pricing to make a budget last is the mistake this whole article is about — it does not spread the budget, it just means the budget is never spent at all.
How to test without wasting money
- Run one country, one device, one week, with a daily budget you would not mind losing.
- Change one thing at a time. Payout, or requirement, or targeting — never two.
- Give it enough conversions to mean something. Twenty is noise.
- Measure at the depth you care about, not at the install.
Then raise the price on what works and stop what does not. It is unglamorous and it is the whole job.