The second call has a first cause.

Consider an illustrative delivery enquiry. A customer checks a tracking page, starts a chat and calls the following morning. The chat was answered quickly. The call is handled politely. Both channels meet their response targets. The customer still needed three attempts to find out when the parcel would arrive.

The useful unit of analysis is the unresolved need across those contacts. Was the delivery information missing? Was the promise unclear? Did the adviser lack authority to resolve it? The answer determines whether the improvement belongs in knowledge, a system, a supplier process or the conversation itself.

In ellio's service, a digital agent uses connected order, fulfilment and interaction knowledge to handle the enquiry through an appropriate channel, including digital voice. It queries a human expert when clarification is needed, retaining the conversation where the task remains within its authority. If a specialist must take over, the history and unresolved issue travel with the customer, and the digital agent continues supplying knowledge and guidance. The completed outcome returns to the shared knowledge.

Start with a defined customer need and a sensible period in which another contact would indicate that it remains unresolved. A parcel enquiry and an insurance claim may require different observation periods. Match contacts across the agreed channels and distinguish a planned follow-up from another attempt to obtain the same answer.

Keep the definition stable while testing a change. A repeat rate can fall because customers have stopped trying, because a channel has become harder to reach or because the reporting window has shortened. Check completion and customer effort alongside volume before calling the result an improvement.

Identity gaps matter too. If phone calls and web chats cannot be linked reliably, report that limitation. A precise-looking percentage does not repair missing evidence.

Here is a deliberately illustrative calculation. An operation handles 4,000 repeat contacts a month at an average fully allocated cost of £6 each. That represents £24,000 of monthly activity. A verified reduction of 1,000 contacts releases activity valued at £6,000 on that basis.

That £6,000 is not automatically a cash saving. Salaries, property and platform costs may remain. The benefit might first appear as capacity for other work, less overtime, shorter waits or room for growth. Finance should distinguish the cost attached to activity from the expenditure the business can actually remove.

The same discipline applies to customer value. A prevented complaint or retained relationship may be valuable, but it needs its own evidence. Adding an assumed retention benefit to every avoided call can exaggerate the result.

Choose one recurring customer need and gather the permitted records that explain it. Put the original question, connected evidence, recommended action, expert consultation, any transfer and the later contact in time order. Include the task record where the action should have happened. This distinguishes a missing answer from a failed action, an unnecessary transfer or guidance that did not reach the person handling the case.

Review a selection of completed needs as well as repeat contacts. They help distinguish the common conditions of the service from the conditions associated with failure. Ask the responsible team to challenge the proposed cause. A pattern in the transcript is a useful lead; the business process must support the explanation.

Separate repeated attempts about the same need from a new question. Keep planned updates visible without treating every update as failed resolution. Where identity matching is uncertain, retain an unresolved category rather than forcing every contact into a tidy customer history.

Suppose the investigation suggests that customers call again because a promised refund never becomes an assigned task. A proportionate test could introduce a confirmed task owner and a visible completion check for one eligible journey. Keep the customer access routes and the repeat-contact definition stable during the test.

A comparable team or a staged introduction can strengthen the evaluation, provided differences in workload, case complexity and staffing are understood. If all teams change at once, use the available baseline carefully and record other changes. The result may still be useful, but it supports a weaker claim about what caused it.

Report the share of the original eligible-need cohort generating a further same-issue contact over the agreed observation period. Show pending and unresolved cases separately. A count alone is misleading when the number of customers served changes. Allow enough time for the promised action to occur and for the customer to discover whether it worked.

Return to the illustrative £6,000 of monthly activity value. Operations should show what the released capacity can actually do. It might cover an existing backlog, avoid an additional shift or support higher demand with the same team. Those are different decisions, with different financial consequences.

Finance can then identify which costs are variable, which commitments can be changed and when the effect will reach expenditure. Include the cost of the fix, ongoing monitoring and any extra work created in another team. A faster contact that leaves more back-office work has moved part of the cost.

The review should end with a decision about the process and a decision about the capacity. If the fix holds but expenditure is unchanged, report a service and capacity improvement. Claim a cash saving only when there is evidence of the relevant reduction in spend. That distinction makes the case more useful to the business.

The table shows possible causes and owners, not findings about a real operation. Each hypothesis needs checking against the records and the people responsible for the process. A confusing letter might contribute to repeat contact, while an underlying billing error creates most of it.

Give an agreed change a named owner, a delivery date and a measure. Compare the repeat rate for the affected need before and after the change, allowing for shifts in volume and customer mix. Where feasible, use a comparable group or a staged introduction to strengthen the evidence.

ellio's approach connects customer support with this improvement work. Resolving the immediate need matters. So does making the recurring cause visible to the person who can address it. Some causes sit with the client or another supplier, which makes clear ownership and follow-through essential.