A referral channel going silent becomes a payroll problem when it supplies half of new revenue. The first move is to protect cash and diagnose the break in parallel, with a firm deadline for deciding whether the relationship can recover.
Consider Kojo, an illustrative composite of founders I have worked beside. At 7:12 on a Monday morning in Accra, he refreshed the partner dashboard for the third time. Zero qualified referrals. He still had Friday’s takeaway container beside his laptop and a payroll file waiting for approval.
The partner had driven roughly half of new revenue over the previous quarter. Nobody had promised that volume would continue, but Kojo had planned as if it would.
Two salaries were due before the next reliable customer payment. If the referrals stayed at zero, he could make payroll only by delaying a contractor, drawing from his personal savings, or closing a customer directly before Friday.
There was no comfortable option. Repairing the partnership might take weeks. Replacing the demand might take longer.
The silence was the last signal, not the first
Kojo’s first instinct was to message the partner lead: “We haven’t received anything today. Is something wrong?”
That question was too narrow. Monday’s empty dashboard could have been a tracking failure, a temporary pause, a change in the partner’s sales incentives, or a deliberate shift toward another vendor. He needed to know which problem he had before spending the week solving it.
He pulled the previous eight weeks into a plain spreadsheet. Referral volume had not collapsed overnight. It had fallen in steps: fewer introductions, lower-quality prospects, slower replies from the partner team. The channel had been weakening for three weeks. Zero was simply the first number nobody could explain away.
This pattern matters. Founders often call a channel dependable because it worked recently. Dependability requires visibility into why it works, who controls it, and what could stop it.
Kojo controlled none of those three things. He knew the partner contact, but not the internal owner setting referral priorities. He knew how leads arrived, but not why his company was selected. He had revenue history, but no agreed minimum volume or warning process.
The relationship still had value. It could no longer carry the forecast alone.
Repair the relationship with a decision deadline
Kojo sent a different message. He shared the decline he had observed, asked whether routing or priorities had changed, and requested a call that day. He also set an internal deadline: if he could not confirm a credible recovery path by Tuesday afternoon, he would remove partner referrals from the month’s cash plan.
That deadline changed the conversation.
Without it, every reassuring reply could buy another day. “We’re checking internally” sounds reasonable until payroll is close. A founder can spend a week preserving goodwill while the company loses the time needed to create another sale.
The partner lead replied before noon. Their team had changed how opportunities were assigned. Kojo’s product had not been removed, but it was no longer appearing in the default recommendation path. Restoring it required an internal review, and nobody could promise when that would happen.
The channel was repairable. It was not forecastable.
That distinction is the decision. A warm relationship can remain worth maintaining while contributing zero to the cash plan. Kojo scheduled the review, supplied the material the partner needed, and stopped treating a possible recovery as incoming revenue.
The same discipline applies when a partnership announcement has created confidence without reliable cash. I explored that gap in what Ama’s failed invoice taught Kojo about revenue.
Replace the demand before rebuilding the channel
By Monday afternoon, Kojo had three tempting responses. He could launch paid ads, offer a broad discount, or ask the team to contact every dormant lead.
Each created activity. None matched the immediate constraint.
Paid ads needed testing time. A public discount could weaken deals already in progress. A large outreach list would consume the week without telling the team which conversations could reach a decision before payroll.
Kojo chose seven prospects who had already described a costly problem, involved the person who could approve payment, and discussed a near-term start. He wrote to each one personally. No campaign language. He named the unresolved decision from their last conversation and offered a paid, tightly scoped first step.
By Wednesday evening, five had replied. Two declined. One wanted another month. Two agreed to calls.
One of those calls ended without a decision. The other prospect asked for a smaller scope and a clear delivery date. Kojo accepted the narrower work because it solved the customer’s immediate problem, could be delivered without hiring, and brought cash inside the payroll window.
On Thursday afternoon, the signed agreement arrived.
Payroll was still close. The company had not discovered a magical replacement for half its new revenue. It had converted one known problem into one paid decision, just in time.
That is the useful job of a runway response. It buys enough room to make the next decision without pretending the underlying risk has disappeared.
Build a warning system before the next Monday
Kojo kept working on the partnership, but he changed what counted as evidence. Conversations no longer entered the forecast because a partner said they were coming. A referral counted only when a qualified prospect entered the pipeline with a named owner and next step.
He also added two weekly checks: the share of new revenue controlled by any one outside party, and the number of direct customer conversations already close enough to become paid work.
Neither metric guarantees revenue. Together, they expose dependence before the dashboard reaches zero.
A related question deserves its own treatment: what to do when a referral partner stops sending qualified sign-ups. The short version is to separate relationship repair from revenue planning. Keep the first open. Make the second survive without it.
The following Monday, Kojo opened the same dashboard. It still showed zero.
This time, the payroll file had already been approved.
Comments
No comments yet.