Five ways prepaid calling leaks revenue
Prepaid is meant to be the safe model: the customer pays before they call, so there should be nothing to lose. In practice, margin still drains away in predictable, repeatable ways. Here are five of them, and how each one is closed.
The appeal of prepaid is simple. Customers fund their balance first, so the operator is never chasing payment for traffic that has already happened. The exposure on any account is supposed to be capped at whatever the customer has put in.
That guarantee only holds if the platform actually enforces it on every call, in real time, without exception. Where enforcement is approximate, the leaks below appear. None of them is dramatic on its own record, which is exactly why they go unnoticed and add up.
Where prepaid margin disappears
Each leak has the same shape: a small gap between what was charged and what was carried, multiplied across high call volumes. Each one closes with the same discipline of rating and enforcing in real time.
Overspend past zero
How it happens. Soft caps, delayed balance updates, and slow disconnection let an in-progress call keep running after the credit has gone. The customer has already paid for nothing more, yet the call carries on against credit that does not exist. On long international destinations a few seconds of lag per call adds up to real uninvoiced minutes.
How to close it. Decrement the balance in real time, per second, while the call is live, and enforce a hard cut-off the moment it reaches zero. With real-time per-second rating and hard cut-off enforcement, the maximum exposure on any single call is bounded rather than open-ended.
Stolen or shared PINs
How it happens. A single PIN that has been guessed, sold, or shared can be used by many callers at once. Each individual call looks legitimate, so nothing stands out until the balance drains far faster than one customer could ever spend it. Without a limit on simultaneous use, one compromised PIN becomes an open tap.
How to close it. Apply concurrency caps and velocity limits per PIN and per account, so a credential can only support a sensible number of simultaneous calls and a sensible call rate over time. Balance and PIN control plus fraud caps turn a runaway PIN into a contained, alertable event.
Rate-deck gaps and unmatched destinations
How it happens. Calls route successfully but the rating layer has no entry for the dialled destination, so the call either fails to rate at all or drops to a permissive default rate. Either way the call completes and consumes a carrier cost while charging the customer little or nothing. New number ranges and reshuffled prefixes open these gaps quietly.
How to close it. Maintain complete A to Z rate decks with longest-prefix matching, so every call resolves to the most specific destination rather than a fallback. Alert on any call that cannot be matched, so a gap is found in CDRs the same day rather than at month end.
Rounding and increment mismatches
How it happens. Charging increments that are too coarse, or that differ between the routing layer and the rating layer, quietly undercharge on every call. A one-minute call rated in sloppy increments, or a route where the supplier bills per second while the customer is metered in larger blocks, leaks margin on volume that no single record makes obvious.
How to close it. Use deterministic per-route increments and minimum charges, agreed up front and applied identically every time. Per-second rating with explicit increments and minimums keeps customer charging aligned with supplier cost, so rounding works for the operator rather than against it.
Negative balances and uncollectable debt
How it happens. When calls are rated after the fact rather than authorised first, a prepaid account can finish a call owing more than it held. That debt is not invoiceable: there is no postpaid relationship to bill against, so the negative balance is simply written off. Repeated across thousands of accounts it becomes a steady, silent loss.
How to close it. Authorise and reserve credit at call setup, and never let a balance go negative. With real-time rating and hard cut-off enforcement, the worst case on any account is that it reaches zero, so the platform never creates debt it cannot collect.
Real-time rating bounds the loss
The common thread is timing. Every one of these leaks comes from rating, checking, or disconnecting a fraction too late. When a call is rated per second while it is live, the balance is decremented as the minutes are used, and the call is cut off the instant credit runs out, the worst case on any account is bounded by the balance itself.
That same discipline is what contains calling fraud. Concurrency and velocity caps stop a shared PIN draining a balance, complete rate decks stop calls slipping through unrated, and an authorise-and-reserve model stops accounts ever going negative. The leaks and the fraud share a root cause, and a single control approach addresses both.
This is the principle Seshnova is built around: rate in real time, enforce balance and PIN limits as the call happens, keep CDRs that make every charge auditable, and never let a prepaid account spend credit it does not have.