Every record you keep is a message to a future version of yourself who has forgotten everything. The test of a record is not whether it looked thorough when you wrote it. It is whether, eight months later, it answers a question asked by someone who was not there.
Three questions, specifically: what happened, when, and how was this number arrived at?
Why the third question is the one that fails
Most records handle the first two well enough. A payment of KES 24,500 on the twelfth is recorded as a payment of KES 24,500 on the twelfth. What is almost never recorded is the derivation: that it settled 18,000 of August rent, 900 of water at six units by 150, and 5,600 of arrears carried from July.
Without that, the number is true but useless. It cannot be checked, cannot be reconciled against what was owed, and cannot answer the only question anyone will ever ask about it, which is "what was that for?"
What a durable record has
- A date that is the date it happened, not the date somebody wrote it down.
- The parties, named: who paid, who received, which unit, which landlord.
- The derivation, where a number was calculated: the reading, the rate, the multiplication.
- The document behind it, where one exists: the delivery note, the invoice, the receipt number.
- Who recorded it. Not to assign blame; to know who to ask.
Corrections should be entries, not edits
The most common way a good record becomes a worthless one is being tidied. Somebody notices last month's figure was wrong and fixes it in place. The books now balance and the history is gone. Worse, nobody can tell from looking whether anything was changed at all.
A correction should be a new entry that references the original and states the reason. It is slightly uglier and infinitely more defensible. This is the whole idea behind an append-only audit trail, and it is the difference between records that settle a dispute and records that become part of one.
Store it where it survives the person
A great deal of property record-keeping lives in individuals: the manager who knows which tenants pay late, the accountant with the spreadsheet, the caretaker who remembers which unit had the leak. All of it disappears when they do, at exactly the moment you need it most.
The fix is boring and effective: records in a shared system rather than a personal device, with access by role, and backups that somebody has actually tested restoring. An untested backup is a belief, not a control.
What "good" feels like day to day
You can answer a tenant's question about a charge from six months ago in under a minute. You can show a landlord where a figure on their statement came from without rebuilding it. You can hand a new member of staff the file for a unit and have them understand its history without a briefing.
That is the standard, and it is reachable with a notebook, a spreadsheet or a system. The format matters far less than the discipline of recording the derivation at the moment the thing happens, while it is still obvious. It is never obvious later. That is the entire problem.
How to prove you paid rent without a receipt
This is the question tenants ask most often, usually after the fact, and the answer is that you rebuild the trail from records other people made.
- The mobile money message and the transaction code. Keep the message, and note the paybill and account number it went to. The code is what lets the payment be traced by someone other than you.
- A statement from your bank or mobile money provider for the relevant period. A statement is stronger evidence than a screenshot, because it was produced by an institution rather than by a phone.
- Your own dated record. A note per month covering the date paid, the amount, what you understood it to cover, and the balance you believed remained afterwards. Contemporaneous notes carry real weight precisely because they were written before anybody was arguing.
- Messages where the landlord acknowledged payment. A WhatsApp message saying "received, thanks" is an acknowledgement. Export the thread rather than relying on it still being on a phone in two years.
- A witness, where cash was involved and nothing else exists.
The gap none of these close is what the money settled. A transaction code proves KES 15,000 reached a paybill; it does not prove whether it cleared August rent or last quarter's water. That is why an itemised receipt matters more than proof of payment does, as why itemised rent receipts prevent disputes sets out. It is also why the single most useful habit for a tenant is to ask for one every month, in writing, while the allocation is still uncontroversial.
The four records that settle most property disputes
Across tenancies, agencies and builds, arguments concentrate in four places, and each has one document that ends it:
What condition it was in. The dated, signed handover inventory with photographs. Nothing else substitutes, and its absence decides deposit disputes against whoever had the duty to make it.
What was owed and what was paid. The ledger showing each charge raised and each payment applied to specific charges, plus receipts that match it. Not a bank statement, which shows movement without meaning.
What was agreed. The signed agreement, plus every later variation kept with it. A varied term filed separately from the document it varies is a term nobody will connect in eighteen months.
What was said, and when. Notices with proof of service, and the messages that agreed a rent change, permitted an alteration or arranged a payment plan. Most of these now live in WhatsApp and almost none of them are filed anywhere.
If you keep nothing else systematically, keep these four per tenancy. They cover the overwhelming majority of what is ever actually questioned.
Turn WhatsApp into a record
Property in Kenya runs on WhatsApp, and pretending otherwise produces a filing system with a hole in the middle of it. Three habits make the messages usable later:
Confirm decisions in a message that stands alone. After a call, send: "Confirming: rent from 1 October is KES X, agreed today." A message that restates the agreement is evidence; a thread of "ok" replies to a call nobody recorded is not.
Export threads that matter, at the end of each tenancy at minimum, into the tenancy file. Phones are lost, numbers change, and a chat history is not a record if it exists only on a device that belongs to a person who can leave.
Move photographs off the phone. Meter readings, condition photographs and delivery counts are worthless in a gallery of eleven thousand images. File them against the unit, with the date, on the day they were taken.
The test, run backwards
The honest way to find out whether your records work is to pick one month from last year, at random, and try to answer four questions about a single tenancy without asking anybody:
- What was billed, and how was each charge derived?
- What was received, on what date, through what channel?
- What did each payment settle, and what balance remained?
- Who recorded all of that, and when?
Most portfolios fail on question one or three, and the failure is always the same shape: the numbers exist and the derivations do not. That is recoverable while the month is recent and unrecoverable once the people involved have moved on.
An hour spent running that test tells you more about your record-keeping than any policy document will. It is also, not coincidentally, exactly what a system like property management software built for Kenya is buying you: charges, payments, allocations and corrections held as one chain rather than as four records that have to be reconciled by hand. The habit of recording the derivation at the moment it is obvious is what makes either approach work.
Common questions
How do I prove I paid rent without a receipt?
Rebuild the trail from the mobile money transaction code, an official statement for the period, your own dated notes, and any message where the landlord acknowledged payment. None of them prove what the money settled, which is why an itemised receipt each month is worth insisting on.
How long should records be kept?
Long enough to cover the period in which a question could still arise, which is usually measured from the end of a tenancy rather than its start, and which differs by record type. The approach, and the reason for deleting some categories rather than hoarding everything, is in what to keep, and for how long.
Is a WhatsApp message a valid record?
It can be a very good one, provided it states the decision rather than assuming context, and provided it is exported somewhere durable. A thread on one person's phone is not a record of a business, it is a record of a phone.
Should corrections be edited or added?
Added, always, as a new entry that references the original and gives a reason. A record capable of being silently changed proves nothing, including the entries that were right all along.
What is the single most valuable record in a tenancy?
The dated, signed move-in inventory with photographs. It is twenty minutes of work and it pre-empts the dispute that costs the most and recurs the most often.
Who should be able to change records?
As few people as possible, each with their own login, and with every change attributable. Shared accounts remove the ability to say who did anything, which defeats the purpose of keeping the trail at all.