23 August 2026 · 9 min read
Setting ROI percentages your platform can actually pay
The admin panel will accept any percentage you type. It will not tell you whether you can pay it. That arithmetic is yours, and it is the difference between a platform that runs for years and one that does not.
What does the number actually commit you to?
Take a plan at 1.5 percent daily for 40 days, with capital returned at the end. A user deposits 1,000.
| Deposit | 1,000 |
| Daily return at 1.5 percent | 15 |
| Over 40 days | 600 |
| Capital returned at the end | 1,000 |
| Total owed on that one deposit | 1,600 |
You received 1,000 and owe 1,600. The gap of 600 has to come from somewhere that is not that user's own deposit. That is the entire question, and it is worth sitting with before you set any percentage at all.
Now the same plan without capital returned: you owe 600 and keep the 1,000, which is a very different business. Users read these two structures completely differently, so be explicit about which you are running.
Modelling a plan before you publish it
Before a plan goes live, work out three numbers on paper.
- Total owed on a single maximum investment, across the full duration, including capital if it is returned.
- Total owed if the plan fills. Pick a realistic number of investors, multiply, and look at the figure honestly.
- The day the outflow peaks: If everybody joined this week, when does the largest single day of payouts fall?
That third number is the one nobody calculates and the one that causes the problem. Payouts from a cohort of investors who all joined in the same period arrive in the same period.
An admin controlled investment platform, installed and configured for you. Full admin control, live in 6 to 12 hours, from $149.
Watching liability rather than deposits
Deposits are the number that feels like success. Liability is the number that tells you the truth.
Outstanding liability is everything you currently owe: returns accrued but not withdrawn, capital due back at the end of plans, and referral commission earned but unpaid. Deposits minus withdrawals is not the same figure and will flatter you.
Check it daily and compare it to yesterday. What matters is not the absolute number but the direction relative to deposits. If liability is growing faster than deposits, and it keeps doing so, the arithmetic has already decided the outcome even though everything currently looks fine.
If your software does not show this figure, that is a serious gap. It is one of the five things worth testing before you buy anything, listed in how to buy without getting burned.
Warning signs, in the order they appear
They arrive in a predictable sequence, and the early ones are visible in the panel long before anything is visibly wrong.
- The withdrawal queue stops emptying: You clear it and it is full again sooner than it used to be.
- Liability grows on a day when deposits were flat: Normal in small amounts, a signal when it repeats.
- Average withdrawal size rises: Users are taking more out per request, which usually means confidence is falling.
- Reinvestment falls: The proportion of returns being reinvested rather than withdrawn is one of the earliest honest indicators you have.
- Support questions change tone: From "how do I" to "when will I".
By the time you are delaying payouts, the decision was made weeks earlier by the numbers.
What can you actually change, and when?
Three levers, in order of how much damage they do to trust.
Change plans for new investors only: Cheapest and least visible. Add a new plan at a lower rate, disable the old one for new entries, let existing investments run out. This is what the disable-rather-than-delete rule exists for.
Lower maximum investment amounts: Reduces how fast new liability is added without touching anything anybody already agreed to.
Change terms on running investments: Almost always wrong. You are altering a deal after it was accepted, and it turns a business problem into a trust problem, which is much harder to recover from.
Use the first lever early. It only works when you still have time for it to matter.
The honest part
There is no percentage that is sustainable on its own. Every plan is a promise funded by something: your own capital, real revenue from somewhere else, or money arriving from new users. Only the first two are stable, and only you know which applies.
What software can do is show you the position clearly and early, so decisions are made with numbers rather than hope. That is what the reporting screen is for, and it is the reason to insist on one before buying anything.
Related: setting up plans and tiers and the admin panel walkthrough.
Common questions
How do I know if a return percentage is payable?
Model it. Take the maximum deposit, apply the rate for the full term, and look at the total you would owe on that one deposit. If the number is uncomfortable, the plan is wrong.
Why do high percentages look fine at first?
Because early deposits arrive before the first payouts are due. The gap between promises and obligations is invisible for weeks, then arrives all at once.
What is the earliest warning sign?
Outstanding liability rising while deposits stay flat. It shows up in reporting well before the withdrawal queue starts lengthening.
Should I match a competitor advertised rate?
Their number tells you what they advertise, not what they pay or how long they last. Matching a rate you have not modelled is copying a decision you cannot see the consequences of.
Can I lower a rate after launch?
For new investments, yes, and existing ones keep their terms. Doing it early and explaining it plainly is far better received than doing it under pressure.
Does a lower rate mean fewer users?
It means different users. A plan that survives long enough to pay reliably is a stronger position than a rate that attracts deposits you cannot service.
What should I do if the numbers are already wrong?
Change the rate on new investments immediately, clear the existing queue in the order it arrived, and be straightforward about what is changing and why.
Question this did not answer? Ask on Telegram. Reader questions get written up first.