The usual answer is that a business should automate collection once it becomes large enough. That framing is not useful, because size is not the variable determining the return. Plenty of businesses with two hundred customers benefit immediately. Plenty with two thousand see very little change.
What matters is the shape of the receivables rather than the volume of them.
Signal one: collections effort scales with customer count
Count the hours your team spends each month on reminders, follow-up calls and payment status updates. Divide by the number of customers billed. Then compare that figure with the same calculation from a year ago.
If the per-customer number has not fallen, the process contains no leverage. Every new customer adds a fixed quantum of collection work, and hiring becomes the only way to absorb growth. An automatic payment arrangement breaks that link, because authorisation happens once at onboarding rather than repeatedly at every cycle.
Signal two: days sales outstanding has stopped improving
DSO usually falls during a company’s early years as processes tighten and terms get enforced. Then it plateaus. The plateau is informative. It marks the point at which the remaining delay is structural rather than a matter of effort.
Chasing harder beyond that point yields very little. The residual delay exists because payment requires a customer action in every cycle, and no amount of reminder discipline removes that dependency.
Signal three: revenue recurs but collection does not
This mismatch is the single strongest indicator. Subscriptions, retainers, instalments, membership fees and trade credit on standing terms are all predictable on the revenue side. If the collection method for them is a fresh invoice and a fresh manual transfer each period, the business is doing repeated work to collect a repeating obligation.
Mandate-based collection is built for exactly this shape. It is a poor fit for genuinely one-off transactions, and for those a standard checkout remains the correct tool.
Signal four: failed payments are quietly costing you customers
In subscription and membership businesses, a portion of cancellations are not decisions at all. A payment fails, service lapses, nobody follows up in time, and the customer is gone without ever having chosen to leave. This stays invisible in most reporting because it lands in the churn number rather than the payments number.
If your cancellations cluster around billing dates, examine payment failures before examining product or pricing.
What the switch actually costs
Three things, stated honestly. There is integration work, which is modest. There is drop-off at the mandate approval step, because a share of customers will decline to authorise a standing debit and you lose them at onboarding. And there is a transition period during which existing customers remain on the old method while new ones register mandates, meaning both run in parallel for several months.
The drop-off deserves the most attention of the three. It is real, it is measurable, and in some customer segments it exceeds the efficiency gained.
When not to switch
Low customer counts paired with high invoice values do not justify it. Neither do billing models where amounts and dates vary unpredictably enough that customers reasonably expect to review each charge before paying it.
Businesses carrying unresolved billing accuracy problems should fix those first. Automating an inaccurate invoice only produces refunds faster, and refunds arriving at scale damage trust more than a late invoice ever did.
There is also a sequencing argument. A business still changing its pricing model every few months will end up re-registering mandates each time the structure shifts, which annoys customers and wastes the setup effort.
For everyone else the test is narrow enough to run this week. Calculate the per-customer collection hour for the last three months. If it is flat or rising while your customer count rises alongside it, automatic payment collection is already overdue, and the delay is costing more than the migration will.
Comments are closed.