The Contract Nobody Owns Renews Itself
The core banking licence, the ATM switch, connectivity, guarding, cleaning, insurance — a SACCO runs on twenty contracts, each renewing automatically on a date nobody owns, at a price nobody compared.
Ask a SACCO how it buys things and you will hear about tender committees, thresholds and quotation comparisons — a governance apparatus built for the day a purchase happens — set out in full in our SACCO procurement governance guide. Now ask what the guarding contract costs this year against last year, when it renews, and who signed it. The apparatus has nothing to say, because the purchase happened in 2019 and has been happening quietly every month since. Recurring spend escapes procurement governance entirely, and in most societies it is the larger number.
The mechanism is worth naming precisely. Procurement controls are event-driven: they fire when somebody raises a requisition. A contract that renews automatically raises no requisition, generates no comparison, and produces an invoice that looks routine because it is routine. So the control never fires. Twenty contracts renewing silently at whatever escalation the supplier wrote into clause fourteen is not a procurement failure — it is procurement working exactly as designed on a category it was not designed for.
One-off purchase — well governed
- A requisition exists, raised by a named person with a stated need
- The value triggers a threshold and routes to the right approver
- Competing quotations are compared on the record, at the time
- The decision is minuted and retrievable three years later
- The approval chain cannot be skipped, by anyone
- The spend appears against a budget before the invoice arrives
Recurring contract — governed once, in 2019
- No requisition. The obligation predates whoever now pays it
- Each monthly invoice sits far below any threshold worth noticing
- The comparison happened once, against a market that has moved
- The signed agreement is in a drawer, a mailbox, or gone
- The renewal date is in clause fourteen and in nobody's calendar
- The escalation clause has compounded quietly for six years
The financial consequence is not dramatic in any single month, which is precisely why it persists. A guarding contract escalating 8% annually while inflation runs lower, on a base of two million, costs a SACCO more over five years than most of the purchases its tender committee spends its evenings scrutinising. And the escalation is invisible because it arrives as a slightly larger invoice in a category that was always there.
What a contract register has to hold
The register itself is unglamorous and does not need software to begin. What it needs is the right fields, because the wrong ones make it a document archive instead of a control. A document archive tells you what you agreed. A control tells you what you must decide, and when.
Field Why it is the control rather than the record
Renewal or expiry date The decision date
Not the signature date and not the anniversary — the last date on which you can act. A contract with ninety days' notice renewing on 1 April must be decided by 31 December, and it is that date, not April, that belongs in the register.
Notice period The window you actually have
The single most frequently missed clause. Miss the notice window and the renewal decision has been taken for you, whatever the register says. Record it as a number of days and derive the decision date from it.
Escalation basis What the price does on its own
A fixed percentage, an index, or negotiable. Write the actual clause reference. This is the field that turns "the guarding cost has gone up" into "the guarding cost went up because clause fourteen says 8% and nobody challenged it".
Named internal owner Who is accountable, not who signed
A contract owned by "management" or by a departed officer is unowned. One named current employee per contract, whose job it is to bring the renewal decision to a committee with a recommendation.
Value and its trend The comparison you will need
Annual value for the last three years, side by side. Three numbers in a row is the most persuasive evidence in any renegotiation, and it takes one column to keep.
Termination and exit terms The cost of changing your mind
Penalties, minimum terms, data-return obligations. Critical for core banking and ICT, where the switching cost is often the reason a bad contract survives. Know it before renewal, not during a dispute.
Last competitively tested The age of your price
The date this service was last put to market. A contract untested for six years is not evidence of a good supplier relationship; it is an untested price with a good relationship attached.
What we hold, and the honest workaround
There is no contract register in our system, and it is worth being blunt about that rather than pointing you at something adjacent. A vendor record carries the supplier's details, their prequalification history, their documents and their transaction trail — but it has no expiry date, no renewal date, no notice period and no contract value. Nothing anywhere holds a recurring obligation, and nothing reminds you that one is about to renew. If your reason for evaluating an operations system is contract lifecycle management, this is not it, and no configuration will make it that.
What does exist, and genuinely works, is a task with a due date and a recurrence — and it is a better answer than it sounds, because a renewal decision is a piece of work with an owner and a deadline, which is precisely what a task is. A recurring task reliably spawns its next occurrence, so an annual renewal review set once continues to appear every year without maintenance. Set the due date to the decision date rather than the renewal date, put the notice period and the current value in the description, assign it to the named owner, and you have converted a silent obligation into a dated piece of work somebody is accountable for. That is most of the control, achieved with a feature that exists.
The other half is that when the review does happen, the procurement machinery is properly good: an RFQ to three suppliers with responses compared on the record, a threshold routing the award to whoever should decide it, supplier prequalification behind each candidate, and the whole comparison retrievable when a member asks about it at an AGM three years later. So the pattern that works is a recurring task that forces the decision, feeding a real procurement event that governs it. The gap is only the reminder, and the reminder is the part you can build out of tasks.
Five questions to ask your own organisation this quarter
List every contract that renews in the next six months, with its notice deadline.
What a healthy answer sounds like
A list produced in under an hour, with dates, values and named owners.
What a hesitant answer is telling you
If this takes a week, no contract in the organisation is being decided — they are all renewing by default, and the ones you would most want to renegotiate are the ones you will discover too late.
For the guarding contract, what did we pay in each of the last three years?
What a healthy answer sounds like
Three numbers, from coded expenses or purchase orders against the same vendor.
What a hesitant answer is telling you
This one is answerable from your transaction history whether or not you keep a register, so a failure here is a coding problem — the spend is spread across categories or vendors that should be one.
When was each ICT service last put to market?
What a healthy answer sounds like
A date per service, and an uncomfortable silence about at least one of them.
What a hesitant answer is telling you
Long-untested prices are normal and not shameful. What matters is knowing which they are, so the board is choosing to continue rather than failing to notice.
Who is the named owner of the core banking contract today?
What a healthy answer sounds like
One current employee, by name, who knows they own it.
What a hesitant answer is telling you
If the owner is a former ICT manager or a committee, the contract is unowned. Unowned contracts renew, escalate and are discovered during disputes.
What would it cost us to leave our core banking provider?
What a healthy answer sounds like
A rough figure covering migration, data extraction, parallel running and penalties.
What a hesitant answer is telling you
Not knowing is how a society ends up with no leverage in a renegotiation. You do not need to intend to leave; you need to know the number, because the number is your negotiating position.
A contract discipline you can start this month
- List every recurring payment from twelve months of expenses and orders. The list is longer than anyone expects and that is the first finding.
- For each one, find the agreement. The ones you cannot find are the priority, not the exception — an obligation you cannot read is an obligation you cannot challenge.
- Record seven fields per contract: renewal date, notice period, decision date, escalation basis, named owner, three-year value, exit terms.
- Create one recurring task per contract, due on the decision date, assigned to the named owner, with the notice period in the description.
- Put anything untested for three years or more on a schedule to go to market — not all at once, three or four a year.
- Take the register to the audit committee quarterly, showing only what renews next quarter. One page, four rows, a recommendation each.
- When a contract is renegotiated, record what changed and what it saved. This is what turns the discipline from an administrative chore into a line the board notices.
One caution about ICT contracts specifically, since they are where the money and the lock-in both concentrate. The switching cost on a core banking system is genuinely high, and a supplier who knows that has little reason to sharpen a price. That does not make the renewal review pointless — it changes its purpose. You are not primarily shopping; you are establishing what the market rate is, what the migration would cost, and therefore what a fair price looks like. A society that has done that homework negotiates differently from one that has not, even when both end up renewing with the same provider.
What AWRA OpsHub does today
- Vendor records with contacts, documents, transaction history and a prequalification trail, so who a supplier is and what you have bought from them is properly held.
- Supplier prequalification as an applied-for, reviewed and approved pipeline, which is the right gate in front of any new service contract.
- Real procurement governance on the renewal event — RFQs to multiple suppliers, responses compared on the record, value-based approval thresholds, and an audit trail retrievable years later.
- Tasks with due dates, named assignees and recurrence that reliably spawn their next occurrence. This is the recommended home for a renewal decision, and it is a genuine control rather than a consolation.
- Expenses and purchase orders coded to a category and a vendor, so a three-year value trend for any supplier is derivable from your own transaction history.
- Document storage with access logging, so a signed agreement can at least live somewhere attributable.
What it does not do
- No contract entity of any kind. No contract record, no start or end date, no renewal date, no notice period, no contract value, no escalation clause, no termination terms. A vendor record has none of these fields.
- No renewal reminders. Nothing anywhere watches a date and warns you. The recurring task above is the mechanism, and you must set it up yourself, per contract.
- No supplier price list or agreed rate, so an invoice price is never automatically checked against a contracted rate. Overcharging against your own agreement is invisible.
- No subscription or recurring-cost register, so the list of everything that renews is a report you assemble, not a list the system keeps.
- No supplier performance scoring in practice. A performance model exists in the code with delivery, quality and pricing-trend fields, and nothing populates it — it is imported in one controller and never used. Treat supplier performance as a manual judgement recorded in minutes.
- No recurring invoice or recurring expense generation, so a monthly obligation is entered each month rather than raised on a schedule.
The straight version: we govern the day you decide, and we do it well. We do nothing at all to tell you the day has arrived. If contract lifecycle management is your core requirement, buy something for it or keep a spreadsheet — and if you keep the spreadsheet, mirror each renewal as a recurring task here so the decision lands in the same place as the rest of your team's work. That combination is honest, cheap, and closes most of the exposure.
Our take
Recurring spend is where a governed SACCO leaks money, because every control it has is aimed at the moment of purchase and a renewal has no such moment. Build the register in a spreadsheet this quarter — seven fields, an afternoon's work — and mirror each contract as a recurring task due on its notice deadline rather than its renewal date. That one distinction, deciding before the notice window closes rather than after, is worth more than any software feature discussed here.
Govern the renewal when it comes
RFQs compared on the record, value thresholds, supplier prequalification, and recurring tasks that put the decision in front of a named owner. A contract register with renewal dates is not built — the note above says so plainly.
See AWRA for SACCOsFrequently asked questions
How many contracts does a typical SACCO actually have?
More than the finance team expects, and the counting exercise is itself the value. Between core banking, the ATM or switch provider, connectivity, mobile-money integration, guarding, cleaning, insurance, audit, legal retainers, software subscriptions, equipment maintenance and branch leases, a mid-sized society with three or four branches will typically find between fifteen and thirty recurring obligations. The small subscriptions are the ones nobody has ever listed, and collectively they are rarely trivial.
Is it worth putting the guarding or cleaning contract out to tender every year?
Annually is usually too often — you spend real management time and destabilise a working relationship for a marginal price move. Every three years is a reasonable rhythm for services like guarding and cleaning, with an annual review that checks the escalation was applied correctly and the service level held. The annual review is the cheap control; the market test is the periodic one. What is not defensible is neither, for six years.
What do we do about contracts we simply cannot find?
Write to the supplier and ask for a signed copy, in writing, and record the date you asked. Most will provide it. Where nobody can produce an agreement, treat the arrangement as terminable on reasonable notice, take it to market at the next opportunity, and put the new one in the register properly. The missing-contract situation is common and the worst response is to leave it undisturbed because raising it feels awkward — an unwritten obligation favours whichever party has better records, and that is rarely the SACCO.
Can we use recurring tasks for statutory deadlines too?
Yes, and it is one of the better uses of the mechanism. Returns, licence renewals, insurance expiry, audit timetables and committee meeting cycles all have the same shape: a date, an owner, a piece of preparatory work. A recurring task with the deadline as its due date and the preparation described in the body works well, and because the recurrence spawns reliably you set it once. The limitation to know is that a task lives beside your records rather than attached to them, so the task will remind you and will not know whether the underlying thing was done.
Should the supervisory committee see the contract register?
Yes, and quarterly rather than annually. Recurring spend is exactly the category where management has the most discretion and the least scrutiny, which makes it a natural area of supervisory interest. Show only what renews in the next quarter, with a recommendation on each — a one-page paper with four rows. A committee that sees renewals coming can question them; one that sees an annual list of contracts already renewed can only note them.
Our core banking provider bundles everything into one annual fee. Does the register still help?
It helps more, not less, because a bundled fee is the hardest kind to challenge without preparation. Record what the bundle includes, service by service, and what each element would cost separately in the market. When the renewal comes, the conversation is about specific components rather than about a single number you can only accept or reject. Bundling is a legitimate commercial structure and it also reduces price transparency; the register is how you get the transparency back.
How do we stop the register going stale like the last one did?
Attach it to a rhythm rather than to good intentions. The reason registers die is that maintaining them is nobody's dated task, so make it exactly that: one recurring task per contract on its decision date, and one quarterly recurring task to take the next quarter's renewals to the audit committee. The register then updates as a by-product of decisions being made, which is the only kind of maintenance that survives a busy year.
