The urgent message that sat in the queue
A distributor in Durban loses its refrigeration unit on a Friday afternoon. The manager emails support: the unit is down, stock is warming, and the delivery to a hospital kitchen is due at six on Monday. The email lands in the shared inbox between a request for a copy of last month's statement and a question about opening hours. It is answered on Monday morning, in the order it arrived, by a polite agent who has no idea that anything is burning.
Nothing in that story involves rudeness or laziness. The team worked through its queue faithfully. The failure is that the queue could not tell a customer who is losing money by the hour from a customer who is idly curious. When every message looks the same, urgency is decided by arrival time, and arrival time is a poor guide to what matters.
This article is about that specific gap: urgent customer issues that are handled like routine questions, and the revenue that quietly leaves because of it. It is distinct from support requests with no clear owner, where the problem is who picks the request up. Here the request has been picked up; what is missing is a shared way of deciding which one to pick up first.
Where the revenue actually goes
A slow answer to a routine question costs very little. A slow answer to an urgent one can cost a customer. The money leaves through several routes, and most are invisible on a month-end report.
- Cancelled and not renewed contracts. A customer who had a serious problem and waited days for a reply remembers the wait, not the eventual fix.
- Penalties and credits. Where a customer contract promises a response or restoration time, a missed urgent request can trigger a service credit or a dispute over an invoice.
- Escalation by the customer. When the support team does not recognise urgency, the customer does it for them: phoning the account manager, copying the managing director, or posting publicly. Senior time is spent firefighting a problem that a clear first response would have contained.
- Duplicate effort. An anxious customer sends the same message three times through three channels. Each copy is read, logged and answered as if it were new.
- Lost upsell conversations. A customer in trouble is not listening to a proposal for additional services, and an account manager who is busy repairing trust is not selling.
Because these costs land in different departments, nobody adds them up. Support sees a tidy queue, sales sees a lapsed account, and finance sees a credit note.
Why urgency gets flattened
In most small and mid-sized South African businesses, support arrives through a general inbox, a WhatsApp number, a web form and the telephone. Each channel produces a message that looks like any other. Several habits then flatten the differences.
- First in, first out. Shared inboxes sort by time. Unless someone actively re-orders the list, the oldest unanswered message goes first, whether or not it is the most serious.
- Urgency is declared by the customer. Customers write "URGENT" on everything, or on nothing. The word carries no information once it is used on every message.
- Priority lives in one person's head. The senior agent knows that a particular client is on a strict contract. The agent who covers on Friday does not.
- No definition of what urgent means. Without a short written definition, such as "customer cannot trade", "customer cannot be paid" or "a safety or legal exposure", each agent applies a private test.
- Everything has the same target time. If the only commitment is "we reply within one working day", an outage and a brochure request carry the same deadline.
What a working system looks like
Support platforms already treat priority as a first-class idea. Microsoft's documentation on defining service level agreements in Dynamics 365 Customer Service describes setting different response and resolution targets, and a way to flag cases that are approaching a deadline. You do not need that product to borrow the idea. The principle is that different kinds of request carry different commitments, and the system, not an individual's memory, tells the team which commitment is nearest.
Routing matters as well. The same vendor's guidance on setting up record routing shows requests being sent to the right queue or person based on rules rather than on who happens to look first. A small business can reproduce the principle with a rule that sends anything mentioning a named key account, or a trading-stopping keyword, to a named person's phone.
A short, usable priority scale
| Level | Plain-language test | Example | First human response |
|---|---|---|---|
| 1 - Critical | The customer cannot trade, deliver or be paid, or there is a legal or safety exposure | Point-of-sale down on a busy Saturday | Within the hour, by phone |
| 2 - High | Work is badly slowed but continues, or a deadline is within days | Reports failing before a month-end | Same business day |
| 3 - Normal | A fault or question with a workaround | A user cannot print from one machine | Within one working day |
| 4 - Low | Information, preference or cosmetic request | Request for a statement copy | Within three working days |
The times above are examples to adapt, not standards. What matters is that they are written down, short enough to remember, and agreed with whoever has to meet them.
The consumer-protection angle
For businesses that sell to consumers, prompt and honest handling of complaints is also a compliance matter. The South African Consumer Protection Act sets out consumer rights around the quality of goods and services and the remedies available when they fall short. A customer who has reported a serious fault and received no response for days is in a stronger position to escalate than one whose message was acknowledged and actioned at once. This is not legal advice; it is a reason to treat the first response to a serious complaint as more than a courtesy.
Estimating what it costs you
You do not need precise figures to see whether the problem is worth fixing. Take last quarter's support requests and answer four questions.
- How many were genuinely urgent by the plain-language test above?
- For those, how long did the first human response take, compared with the average for all requests?
- How many of those customers later escalated, complained or left?
- What was the annual value of the customers who left or reduced their spend?
Even a handful of lost accounts, set against the cost of a half-day of process work, usually settles the question. If the numbers are uncertain, treat them as a range and say so. A conservative estimate is more persuasive to a sceptical owner than a confident one that cannot be checked.
Common objections
"Our team is too small for priorities." A small team benefits most, because there is no spare capacity to absorb a misjudged urgent request. A four-level scale on a sticky note is enough to begin.
"Customers will abuse the urgent label." They will, if the label is the customer's choice alone. Use the plain-language test, apply it on triage, and tell customers what qualifies. Most abuse stops once the criteria are visible.
Related problems worth reading
If your team also struggles to agree what to work on first internally, unclear work priorities and misdirected effort describes the same flattening inside the business. If the underlying cause is customers who cannot see what is happening, customers who do not know the status of a request covers the follow-up workload that results. And where the service depends on whoever picks up the ticket, see when service quality depends on who picks up the ticket.
Practical next actions
- Write a four-level priority scale in plain language, with one real example from your own customers at each level, and get your support team to agree it in a single twenty-minute meeting.
- Name the accounts and request types that always count as Level 1 or 2, and list them where every agent can see them, including the person who covers lunch and leave.
- Set a different first-response target for each level and add a visible marker, such as a colour tag, a flag or a separate folder, so urgent items are not buried by arrival order.
- Decide who is phoned, and at what hour, when a Level 1 request arrives outside office hours, and write down the backup if that person does not answer.
- Review last quarter's genuinely urgent requests, record how long each waited, and repeat the review monthly for three months to see whether the waits are shrinking.
Treating urgent and routine requests differently is not extra bureaucracy. It is a way of making sure the few messages that carry real revenue risk are seen by a person while there is still time to act.
Sources and further reading
- Configure service-level agreements in Dynamics 365 Customer Service (last verified 10 October 2026)
- Set up unified routing for records
Related Mellorca reading
- Answering the Same Customer Questions Daily: The Labour Cost of Hidden Knowledge
- Customers on Hold While Staff Search Five Systems: The Cost of Scattered Answers
- Customer Records in Salespeople's Spreadsheets: The Revenue and Continuity Cost