Track Payments Effectively: Software Check for Businesses

If you search for software that keeps track of payments, you almost inevitably end up with household budgeting and savings apps. They're sensibly built for their purpose: they categorize what has already left the account. An invoice that hasn't been paid yet naturally doesn't show up there — there's simply no transaction for it to attach to.

For a business, the perspective is reversed. Your figures don't originate on the bank statement, but at the moment the invoice is sent. The invoice exists on the 3rd, the money arrives on the 28th, and during those 25 days the software is supposed to tell you that something is outstanding. Backward-looking expense trackers structurally cannot do that.

This leads to a simple pre-selection rule: if a tool has no data record called "invoice" with a due date, it cannot deliver a payment overview. Everything else — design, app ratings, feature scope — only becomes relevant after that.

The third row is underestimated. A spreadsheet is not a bad tool — it's a tool without memory. It doesn't remind you, it doesn't send reminders, it doesn't notice when a line is missing. As long as you keep the list in your head yourself, you won't notice.

MeinGewerbe belongs in the fourth row. Check it using the same scorecard as any other tool — see below.

Six criteria, 0 to 3 points each. The grid is deliberately small: it only asks about what carries the overview, not about depth of bookkeeping.

How to read the result:

• 0–7 points: A spreadsheet performs just as well. Save yourself the switch.
• 8–13 points: A transitional solution. Holds up as long as you still keep the exceptions in your head.
• 14–18 points: Holds up even with 50 or more simultaneously open items.

Criterion 3 is the one checked least often in demos and hurts most often in everyday use. Bank interfaces fail, access credentials expire, a customer pays cash. If a payment can't then be entered manually, the list is wrong from that day on.

The effort for payment reconciliation depends less on the software than on a single variable: the share of incoming payments that can be matched without having to think. Let's call it q.

Worked example with assumed values, 60 invoices per month:

• q = 0.85: 51 matches × 5 s ≈ 4 minutes, 9 exceptions × 90 s ≈ 14 minutes → about 18 minutes
• q = 0.50: 30 matches × 5 s ≈ 3 minutes, 30 exceptions × 90 s ≈ 45 minutes → about 48 minutes

Same software, same invoice volume, half an hour of difference. The lever isn't the program, it's q. Plug in your own values — your 90 seconds might be 3 minutes if you have to write an email first for every clarification.

And q doesn't rise through more features, but through the invoice number making it all the way into the payment reference. A QR code on the invoice that pre-fills the amount and reference does more for this than any dashboard. How structured invoice formats transport the same data in machine-readable form is covered in E-Rechnung Software: Formate, Funktionen und Auswahlkriterien.

Automatic matching works — until it hits one of these cases:

1. Combined bank transfer. One amount, three invoices. Without a split function, all three remain open or all three get closed incorrectly.
2. Partial payment. 50 percent deposit, remainder after acceptance. The invoice is neither open nor paid — a state many lists don't recognize.
3. Cash discount deduction. The customer deducts 3 percent, the amount no longer matches. What's left is a residual item of 14.70 euros sitting in the list for months.
4. Mangled number. "RE-2026-0000147-A" becomes "RE 2026 147" when typed in manually. Short, clearly separated numbers get miskeyed less often than long ones with suffixes.
5. Third-party sender. Payment comes from the spouse's account, the parent company, or the tax office. The name on the statement doesn't match any customer.
6. Cancellation and refund. An outgoing payment that belongs to an incoming invoice — the direction doesn't match, and the matching fails.

As a rule of thumb: if more than one in ten payments takes one of these six forms, automatic reconciliation in your business isn't a time-saver but a delayed source of errors. In that case, first check whether the root cause can be eliminated — for example through a fixed discount rule or individual invoices instead of combined ones — before switching tools.

A payment overview is only as good as the appointment at which someone actually looks at it. Five steps, fixed time slots:

Step four is the most important and the first one to get dropped. It's the only point at which you notice that the software has lost something — a gut-feeling number against a system number.

When you send a reminder and in what tone is a commercial decision, not a software question. Clarify the legal details on default and reminder fees with a lawyer if in doubt; the program only gives you the date and history for that.

Five patterns that are almost never the customer's fault:

• The total of open items grows while revenue stays the same. Probably unrecorded incoming payments, not worse-paying customers.
• Cent amounts and tiny sums stay open. Cash discount, rounding, or transfer fees — every residual item under five euros is a candidate for write-off.
• The same invoice appears twice. A classic after a CSV import or a migration.
• A single customer accumulates a conspicuous number of open items. Usually they're paying under a different name, and reconciliation doesn't find them.
• Nothing is older than 90 days. Could be true. Check anyway whether a filter or an archive is hiding the old cases.

Below about ten simultaneously open invoices, with a single payment channel and no partial payments, the spreadsheet wins — it's quicker to adjust, requires no onboarding, and you see everything on one screen.

It tips over as soon as two of these three conditions apply:

• You regularly work with deposits or installment payments.
• Payments come in through more than one channel (bank account, card, payment service provider, cash).
• More than one person issues invoices or maintains the status.

At two out of three, you no longer pay the price in software fees, but in follow-up questions. The third condition is the toughest one: two people working in a spreadsheet sooner or later create two truths.

As soon as a program reads account transactions, a tool question becomes an access question. Clarify this beforehand, not at the first incident:

1. How does the software get the transactions? Through its own interface, through a third-party provider, or through manual import? Ask for the name of the service provider, not just for "secure bank connection."
2. Does the tool also work without a bank connection? If not, your payment overview depends on an access point you don't control.
3. Where is the data stored, and is there a data processing agreement? Both should be available on request within a few minutes.
4. What does the export look like if you cancel? Test it once while you're still satisfied. An export nobody has ever tried is just a claim.
5. Is there two-factor protection for access — and for all users, not just the owner?

General, vendor-independent guidance on cloud and IT security questions is published by the Bundesamt für Sicherheit in der Informationstechnik. Which retention and traceability requirements specifically apply to your records is something to clarify with your tax advisor — no software can answer that question for you, no matter what the product page claims.

The payment overview doesn't originate at the invoice, but one stage earlier. A quote that already carries the payment term, validity period, and a consecutive number as its own fields passes these values on to the invoice. A quote where the same information sits in free text forces you to retype it later — and every retyping is an opportunity for a discrepancy.

That's exactly the useful criterion when you're looking for and want to use a professional quote template for your business: not the layout, but the question of whether the payment term, number, and service period exist as structured fields. Templates that don't separate these look good and cost you minutes per invoice afterward. Which fields a template should include is covered in Angebot erstellen software: 7 funktionen für rechtssichere angebote.

Is a household budgeting app enough for the self-employed?

For the private side: quite possibly. For open customer receivables, it lacks the data record a due date could attach to. Running both in parallel is unproblematic, as long as you know which of the two applications tells you who hasn't paid yet.

Which software is right for me?

The question can't be answered with a product name, because the answer depends on your payment structure, not on feature lists. Take the scorecard above, run two candidates through it, and calculate the reconciliation effort with your own q. At a similar score, the criterion you need most often decides — usually that's partial and combined payments.

Do I absolutely need a bank connection?

No. Below about 20 incoming payments a month, manual checking off with CSV import is often quicker to set up than clarifying an interface. The connection pays off where many small amounts arrive with a clean payment reference — that is, exactly where q is high.

What happens with cash and card payments?

Those don't arrive as individual account entries, but as a daily total or as a payout from the payment service provider, often reduced by fees and with a delay of one to three days. Before choosing, check whether a payout can be split across several invoices — that's criterion 2 of the scorecard, and it's where lean tools fail most often.

Don't use the provider's demo data. Take your ten oldest open invoices from the last three months — and deliberately add the two most uncomfortable ones: the one with the partial payment, and the one where a customer settled three invoices in a single bank transfer.

Enter all twelve, book the actual incoming payments against them, and time it. If after 20 minutes you have a list whose total is correct, you've found your tool. If you had to add the exceptions in a side table at the end, that won't get better in everyday use — so test the next candidate with the same twelve invoices.