The meeting ends with a page of customer suggestions. The company still has to decide what each suggestion means.
“Build a portal” might describe a visibility problem, an access problem, or a preference for a different interface. Put the phrase on a roadmap without examining the problem and you have chosen a solution before understanding its job.
Create a feedback register after the meeting. Separate what participants report, what they propose, what remains unknown, and what the company decides.
GitLab describes its executive advisory program as a forum for exchanging insights with company leaders. That public description provides advisory context. The register below is an editorial method for handling input, not GitLab's documented follow-up procedure.
Use it for strategic input across customers. Keep complaints about an agreed deliverable inside the relevant account's delivery process. The advisory group does not reopen another customer's scope or acceptance decision.
Preserve Four Different Things
Write the reported problem in terms of a task and an obstacle. Keep the participant's suggested solution in a separate field.
Then add supporting context: the role involved, the process stage, the example available, and any limits on what the note owner can verify. Finally, record the company's assessment and decision.
These fields answer different questions:
| Record | Question it answers | Fictional example |
|---|---|---|
| Reported problem | What becomes difficult during the work? | A reviewer cannot tell which intake questions still need an answer |
| Suggested solution | What did the participant propose? | Add a customer portal |
| Supporting context | Where does the problem occur, and what supports the account? | A coordinator describes a handoff between two roles; no sample has been reviewed |
| Company decision | What will the company do with this input? | Investigate the handoff before considering an interface change |
The final row belongs to the company. A participant proposing a portal does not authorize its construction or establish that other customers need it.
Keep the original account accessible within its permissions. Summaries should preserve meaning without copying sensitive records into a broadly shared worksheet. Mark interpretations as interpretations.
For deeper investigation, use the discipline in customer research interviews: ask about a recent task and keep observations apart from inferred explanations. A follow-up question should test the problem rather than recruit agreement with a proposed feature.
Give Every Input an Owner and a Boundary
A useful register can be a table in the company's existing workspace. It does not need a new customer-facing system.
Start with the following fields. Split wide tables into linked views when the document becomes hard to read. Keep one stable input identifier across those views.
| Field | What to record |
|---|---|
| Input ID and meeting | Stable reference and the session that produced it |
| Problem | Task, obstacle, and reported consequence without invented measurements |
| Proposed solution | Participant's suggestion, kept separate from the problem |
| Evidence and limits | Supporting account or permitted record, plus what remains unverified |
| Affected segments | Context where the problem is reported; other segments remain unknown |
| Owner | Person responsible for the next assessment step |
| Next investigation | Specific question or evidence needed before a decision |
| Status and reason | Investigate, decline, already planned, or deferred, with an explanation |
| Decision owner | Person authorized to choose the company's response |
| Communication boundary | What can be shared, with whom, and which commitment must remain unmade |
| Review trigger | Date or event that causes the owner to inspect the row again |
Use a named owner in a real register. The fictional example below uses roles because it contains no actual people.
The communication boundary matters even for internal records. “Investigating” describes an assessment. It does not establish a delivery date, an approved scope, or permission to disclose another participant's example.
Keep evidence links in a controlled location. Record a non-sensitive summary in the register if its audience is broader than the original notes.
Read a Fictional Six-Item Register
Consider a fictional service firm that documents supplier intake processes. Its customer advisory group has discussed readiness and handoffs. Every record below is invented. No entry describes a real customer or completed project.
The program owner creates two linked views of six inputs. The first view holds the problem, proposal, and evidence. The second holds responsibility and disposition.
| ID | Problem and proposed solution | Evidence and affected segment |
|---|---|---|
| A1 | Reviewers cannot identify unanswered intake questions. Proposal: build a portal. | One coordinator describes a handoff; no artifact reviewed. Reported in a multi-team setting; other segments unknown. |
| A2 | Reported obstacle: unknown. A manager requests visibility into individual workers' activity. Proposal: add person-level activity ranking. | One request without a demonstrated process requirement. Segment fit and effects on workers remain unverified. |
| A3 | Teams use different names for the same handoff field. Proposal: create a shared glossary. | Two participants describe different labels. In this fictional scenario, the documentation owner has checked both labels against the approved glossary scope and confirmed coverage. |
| A4 | A distributed team needs review material in another language. Proposal: translate the complete service library. | One described review task; required languages and document scope unknown. Reported for distributed teams only. |
| A5 | A reviewer cannot distinguish a revised instruction from its previous version. Proposal: send a change summary. | One permitted synthetic example identifies two versions. Reported by a receiving reviewer; frequency unknown. |
| A6 | Reported obstacle: unknown. A delivery contact requests every open intake item in a daily message. Proposal: create daily reminders. | One preference; no evidence about recipients' needs. A quieter participant prefers exception-only updates. |
Now attach a responsible next step to each input:
| ID | Owner and next investigation | Status and communication boundary |
|---|---|---|
| A1 | Service designer: compare the described handoff with a permitted synthetic walkthrough. | Investigate. Describe the problem under study; promise no portal or implementation date. |
| A2 | Sponsor: record the mismatch with the forum's process-documentation purpose. | Decline this ranking proposal. Preserve the stated visibility need; do not imply worker monitoring is planned. |
| A3 | Documentation owner: retain the scope comparison and check for changes at the next glossary review. | Already planned. Link the prior internal decision and confirmed coverage; promise no new scope or earlier release. |
| A4 | Program owner: establish the review task, required language, and available evidence. | Deferred pending a defined need and assessment capacity. State the review trigger; give no translation date. |
| A5 | Delivery standards owner: test whether a change summary answers the version question. | Investigate. Keep the example synthetic; make no service commitment. |
| A6 | Sponsor: retain both preferences and assess the decision context. | Decline a universal daily message. Preserve exception-only input; do not claim consensus about cadence. |
Set review triggers separately. A1 returns when the walkthrough exists. A4 returns when its task and language needs are defined. The owner should also inspect unresolved rows at the next internal program review.
A deferred entry without a trigger can disappear into a storage bin. Write the condition that would make another assessment useful.
A2 and A6 contain requests without a described work obstacle. Keep that gap visible. Declining their proposed solutions does not establish what caused the requests. Ask about the task before adding a problem statement. A3 has a different evidence boundary: “already planned” follows a scope check. A similar project name alone would leave the input under investigation.
Keep Priority Separate from Applause
Do not turn the number of people supporting an idea into its entire business case.
In this method, support counts describe the discussion. They do not establish how common a problem is outside the room. The advisory membership was selected, and attendance defines who could express a preference.
A minority concern can expose a different process requirement. In A6, the quieter participant's preference for exception-only updates remains attached to the daily reminder proposal. Neither preference proves the right cadence for every recipient.
Compare inputs against a named company decision. Consider relevance, available evidence, dependencies, effort, and the consequence of leaving the problem unresolved. Mark estimates and assumptions so the sponsor can challenge them.
Atlassian's prioritization matrix compares work by urgency and impact. Its preparation also records expected outcomes, effort, risks, and dependencies. Those are practitioner prompts, not a mandatory scoring formula for advisory programs.
Use a matrix when it helps the team explain a tradeoff. Do not manufacture a numerical score when the evidence is still a single unexamined account. “Unknown pending walkthrough” can be the more useful entry.
Name Who Can Make the Decision
Separate coordination from authority. The program owner can maintain the register without approving new service work.
Atlassian's DACI method distinguishes a driver, an approver, contributors, and people who need the outcome. Its approver makes the decision; contributors provide recommendations. Apply that distinction when it clarifies the company's own responsibility.
For A1, the service designer gathers evidence. The sponsor decides whether the problem warrants further service design. Advisory members supply context. Delivery and documentation owners receive the resulting decision if their work is affected.
Record the reason alongside the status. “Decline” needs enough explanation for the next internal reader to understand what was rejected. In A2, the company declines the ranking proposal while preserving the underlying visibility question.
For “already planned,” link the actual internal decision. Similar words do not prove that an existing project covers the same problem. Check scope before merging the input with prior work.
If the company changes its decision later, retain the earlier disposition and its reasoning. Add the evidence that changed the assessment. Do not rewrite the record to imply the company always intended the new direction.
Close the Loop at the Right Level
Treat acknowledgement, assessment, decision, and delivery as separate events.
The program owner can record that an input was captured. An investigator can record what evidence was examined. An authorized sponsor can decide what follows. Delivery needs its own scope, owner, and acceptance conditions.
Keep those boundaries visible in consulting project management. Advisory interest does not change an existing statement of work. A program discussion should not become an invisible addition to someone's delivery queue.
Before sharing an outcome, check what the recipient may see. Remove other participants' private details. Confirm the current decision and any uncertainty. Use a separate authorized communication process for outward messages.
For client retention work, the register supplies a record of what the company heard and assessed. It does not prove satisfaction, renewal intent, or future revenue. Keep those questions tied to their own evidence.
A closed input means the company has recorded a disposition. It does not necessarily mean the suggested feature exists. Define that meaning so an internal status does not become an external promise.
Write an Internal Follow-Through Note
The program owner can close the internal review with a note like this fictional example. It is an internal working record, not a message to customers.
Internal advisory review: readiness and handoffs
Six inputs are recorded as A1 through A6. The source view preserves each problem, suggestion, and evidence limit.
A1 and A5 need investigation. The service designer owns the A1 walkthrough. The delivery standards owner owns the A5 version example. Both return when their evidence is available.
A2's activity-ranking proposal is declined because it falls outside this program's process-documentation purpose. The visibility question remains recorded without a monitoring commitment.
A3 is covered by the approved glossary scope. The documentation owner has checked both labels and retains the comparison with the prior decision. Scope changes return at the next glossary review. No new release date is recorded.
A4 is deferred until the task and language needs are defined. The program owner checks that condition at the next internal review.
A6's universal daily-message proposal is declined. The record retains the exception-only preference and the unresolved question about recipient needs.
The sponsor owns company decisions. No portal, translation release, new delivery scope, or external roadmap commitment is approved by this note. Any participant-facing account needs a separate accuracy and disclosure check.
Leave the note where the next owner can find the evidence and current status. An orderly register still needs someone to inspect its open rows.
Keep the problem visible until the decision is recorded.
Sources
- GitLab Handbook: Growth Community Programs. Advisory forum context only. Checked 2026.10.28.
- Atlassian Team Playbook: Prioritization Matrix. Urgency, impact, effort, risks, and dependencies as comparison prompts. Checked 2026.10.28.
- Atlassian Team Playbook: DACI. Separate coordination, contribution, and decision authority. Checked 2026.10.28.