Back to all articlesCustomer feedback

Decide when repeated complaints should become one shared issue

Alactic Editorial9 min read

Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.

Several customers say their collection updates were misleading. One received a notice too early, another saw an outdated delivery estimate, and another was sent to the wrong location. Counting them separately can hide a recurring failure. Combining everything under “delivery communications” can conceal problems requiring different action.

Create a shared issue when complaints show a specific failure pattern, supporting evidence, and a sufficiently common corrective action to justify prioritising them together. Use candidate status when the connection is plausible but a decisive fact remains uncertain.

This method addresses operational grouping for complaint prioritisation, not contact deduplication, technical incident management, or detailed qualitative coding. The worked example shows how to make, challenge, and revise a grouping decision.

1. Separate the shared issue from the individual complaint

A complaint records a customer's experience, its consequences, the response requested, and what your team has promised. A shared issue records a recurring failure that an owner should investigate or improve. Linking these records should not replace either one.

The shared issue needs a failure statement, inclusion and exclusion rules, supporting evidence, an owner, and a review point. Each complaint needs its own account, requested remedy, response owner, next update, and resolution status.

Use a relationship such as “linked to shared issue” rather than merging complaint accounts into one summary. Preserve the original account and record later interpretations separately. “Customer experienced inconvenience” loses the distinction between a wasted journey and a call made before leaving home.

This distinction also prevents false closure: completing an operational improvement does not settle every customer's complaint, and settling one complaint does not demonstrate that the recurring failure has stopped.

2. Apply evidence gates before counting links

A rising complaint count warrants investigation; it does not establish a shared issue. Similar language can describe different failures, while different language can describe the same failed step. Apply these gates:

  • Failure match: Identify the expected step that failed. “Collection notice preceded parcel availability” is more useful than “poor communication.”
  • Context match: Compare conditions that could change the explanation or response, such as collection point or process route.
  • Evidence match: Check whether message records, staff notes, or transaction histories support the sequence. Establish what each decisive timestamp means.
  • Action match: Ask whether the proposed correction would reasonably address the observed failure in each case.
  • Contradiction check: Look for evidence supporting a different sequence or materially different correction.

Use three states. Linked means checked evidence supports the current boundary, not that a root cause has been proven. Candidate means a relevant fact is missing or ambiguous. Separate means available evidence puts the complaint outside the boundary.

You need not prove a root cause before opening an issue. Distinguish the observed failure from its suspected explanation. However, an unchecked interpretation of decisive evidence is not a provisional link: it is a reason to retain candidate status.

3. Use this filled issue-linking worksheet

All identifiers, records, times, roles, counts, and commitments below are fictional and hypothetical. This is a worked example, not an Alactic customer result. Review days refer to one hypothetical working week.

Shared issue record

  • Issue: SI-014 — North collection notices sent before parcels were available.
  • Observed failure: Notices were sent after site arrival but before the recorded handoff making parcels available at the counter.
  • Suspected explanation: Notification may be tied to site arrival rather than counter availability; unconfirmed.
  • Include: North cases where a notice preceded counter availability, supported by records whose meaning has been checked.
  • Exclude: Wrong-location notices, outdated delivery estimates, and parcels already available when notified.
  • Candidate gate: Missing timing or uncertain availability evidence prevents linking.
  • Proposed shared action: Review the handoff between site arrival, counter availability, and permission to send a notice.
  • Issue owner: Collection operations lead.
  • Review: Friday at 10:00, or sooner if contradictory evidence arrives.

Complaint decisions

Worksheet review: Monday at 14:00, by the complaint review lead. Role-based owners below represent assigned people in this fictional team.

  • C-201 — Linked. Notice at 09:10; customer arrived at 10:00; staff could not release the parcel; checked counter-handoff record at 11:40. Match: notice preceded availability at North. Contradiction: none identified. Impact: wasted journey. Response owner: travel-cost case handler. Customer update: decision or progress report on the travel-cost request by Thursday at 16:00. Evidence task: none outstanding for linking.
  • C-202 — Linked. Notice at 13:00; checked counter-availability record at 15:20. The customer called rather than travelling. Match: same sequence at North. Difference: avoided journey, not a contradictory failure. Response owner: collection enquiries adviser. Customer update: confirm current availability and explain the known sequence by Wednesday at 12:00. Evidence task: none outstanding for linking.
  • C-203 — Candidate. Notice timing is recorded; the customer says staff could not find the parcel; the availability record is missing. Possible match: reported sequence and location fit. Uncertainty: the parcel may already have been available but difficult to locate. Evidence owner and deadline: North collection supervisor to reconstruct the handoff from shift notes and parcel records by Wednesday at 15:00. Response owner: missing-parcel case handler. Customer update: Friday at 12:00, even if evidence remains incomplete.
  • C-204 — Separate. A notice directed the customer to North, but the parcel was available at South before notification. Surface match: inaccurate collection information. Contradiction: location assignment failed, not notice timing. Routing: location-accuracy review. Response owner: location support adviser. Customer update: confirm collection arrangements by Monday at 17:00. Evidence task: none outstanding for this classification.
  • C-205 — Separate. A delivery estimate remained unchanged after a delay. Surface match: misleading information. Contradiction: no collection notice or counter handoff occurred. Routing: delivery-estimate review. Response owner: delivery support adviser. Customer update: revised status by Tuesday at 12:00. Evidence task: none outstanding for this classification.
  • C-206 — Candidate. Notice at 10:00; a field labelled “available” shows 12:20. The customer was turned away and also complained about staff conduct. Possible match: apparent premature notice at North. Uncertainty: the field may record retrieval rather than initial counter availability. Evidence owner and deadline: North records supervisor to verify the field meaning and check handoff and storage records by Thursday at 10:00. Response owner: service recovery lead. Customer update: address both collection and interaction concerns by Friday at 15:00.

Decision summary and reusable structure

Initial position: two linked complaints, two candidates, and two separate complaints. Neither candidate contributes to the linked count. Different consequences do not invalidate the two supported links.

To reuse the artifact, copy the shared-issue fields and repeat this complaint structure: identifier; decision; observed sequence; matching evidence; contradiction or uncertainty; evidence task, owner, and deadline; response owner; customer-update commitment. Add a decision date and reviewer whenever a classification changes. “No outstanding evidence task” is more informative than an unexplained blank.

4. Write a boundary another reviewer can apply

Use this formula: observed failure + defined process + relevant conditions. Then add an exclusion for the nearest likely confusion.

For SI-014, that becomes: “At North, a ready-for-collection notice preceded counter availability.” Its nearest-confusion exclusion is: “Exclude cases where the parcel was already available but staff could not retrieve it.”

“Delivery problems” is a reporting category, not an operational boundary. Avoid adding conditions merely because they appear in the first complaint; a condition belongs in the boundary only if it changes the inclusion decision or proposed action.

5. Keep the customer-response work actionable

The issue owner coordinates improvement; response owners handle individual replies. At each review, use this operational checklist:

  • Confirm who owes the next response and when it is due.
  • Carry forward requested remedies and unresolved promises.
  • Record the actual consequence rather than inheriting a group description.
  • Assign additional concerns, such as staff conduct, to an appropriate reviewer.
  • Share only the customer detail needed for the operational task.
  • Check individual resolution conditions before closing any complaint.

Separate evidence-task deadlines from customer-update commitments. In C-203, Wednesday's evidence deadline supports Friday's update but does not replace it. If the supervisor cannot reconstruct the handoff, the response owner still owes an update explaining what is known, what remains uncertain, and the next step.

A customer reply should distinguish established facts about that case from the wider investigation. A suspected shared explanation should not be presented as a confirmed cause.

6. Report scale without inflating it

Report linked and candidate counts separately, with the review date. Also distinguish complaint records from affected service events. Several contacts about one failed collection indicate repeated customer effort, not several failed collections.

In this hypothetical example, two linked complaints establish two supported complaint records within SI-014's boundary. They do not establish the proportion of North collections affected. A failure-rate calculation would need a comparable count of affected collection events and a relevant total for the same process and period. Neither denominator nor complete event coverage is available here, so report counts and limitations rather than a percentage.

Prioritisation should also consider consequences, continuing exposure, and recovery effort. A lower-volume issue involving wasted journeys may warrant attention alongside a more frequent issue causing minor confusion. Record the reason for priority so a larger but loosely defined category does not automatically displace a smaller, well-evidenced problem.

7. Resolve candidates and revise decisions visibly

In the hypothetical continuation, C-206's records supervisor finds a completed counter handoff at 09:40, before the 10:00 notice. Storage records show the parcel was placed in the counter area but on the wrong shelf. The supervisor confirms that the 12:20 “available” field recorded retrieval from that shelf, not initial counter availability.

Change C-206 from candidate to separate. The checked evidence supports a retrieval failure rather than premature notification. Record Thursday's decision, the supervisor's evidence, the corrected interpretation, and routing to a retrieval-process review. Its staff-interaction concern remains assigned to the service recovery lead.

The revised position is two linked complaints, one candidate, and three separate complaints. C-203 remains the sole candidate. C-206 was never part of the linked count, so this is a resolved uncertainty, not an unlinking correction.

If later evidence overturns an existing link, preserve that earlier decision and explain the change. If a case shows both premature notification and a retrieval failure, it may support both issues. An additional failure does not erase evidence for the first.

8. Run a bounded quality review

Start with one theme and a manageable set of recent complaints. Ask another reviewer to challenge the evidence, especially the strongest-looking candidate and the nearest excluded case. Check whether they can apply the boundary without relying on the original reviewer's memory.

At the next review, record:

  • Whether every linked case has checked supporting evidence.
  • Whether each candidate has a specific evidence request, assigned owner, and deadline.
  • Which candidates received supported decisions and which remain uncertain.
  • Whether changed decisions reflect new evidence or an unclear boundary.
  • Whether any customer-update commitments are overdue, including in separate cases.
  • Whether the proposed correction still fits every linked case.

Do not target the highest linking rate. The useful outcome is a defensible group: another reviewer can explain each inclusion, identify unresolved evidence, and see what action follows. A review that produces several separate issues can be more useful than one large category that nobody can act on.

Customer feedbackComplaint managementOperational improvement

Put the next conversation to work.

Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.

Explore Communications