Collaboration Scaling Matrix: From Ratings to Agreement

A collaboration scaling matrix keeps different people’s accounts of one shared task visible. It records what each person means, what information matters to their role and which next agreement would make the work more usable. A number is optional; an explanation is essential.

The purpose is not to rank departments or produce an objective collaboration score. A sales lead’s seven and an engineer’s two may expose different parts of a handover, not different levels of commitment. The complete example below follows those differences through an accepted pilot, a changed request, a failed update and a mixed review.

Understand the conversational method and its limits

Solution-focused scaling comes from the clinical tradition associated with Steve de Shazer, Insoo Kim Berg and colleagues. The SFBTA Research Committee’s 2025 treatment manual describes scaling as an interviewing tool for exploring impressions, goals, past successes and future possibilities. The workplace matrix in this guide is an original adaptation, not a validated team diagnostic or a clinical protocol.

In Scales and Scaling, published in January 2001, de Shazer questions treating the same number from different people as the same experience. He also emphasizes the conversational context surrounding an answer. Preserve the explanations rather than assuming a shared numeral provides a common measurement unit.

That does not mean factual evidence is irrelevant. The conversation can identify a missing approval, an outdated brief or an unresolved dependency. Investigate those matters through the proper operational process. A solution-focused question is not a replacement for technical assessment, incident analysis or a formal conduct route.

Keep personal descriptions, work evidence and authority separate. “I feel more confident,” “the sample was received” and “the delivery date was approved” establish different things. A matrix becomes misleading when it silently converts one into another.

Choose one task, period and use for the record

Begin with an interaction the participants can actually discuss. “How usable is the current pilot brief for deciding the next assessment?” is more precise than “How collaborative are Product and Engineering?” The latter invites people to judge a whole department from an undefined set of events.

Agree the period and descriptive anchors, then check what each person understands. Offer words, a range or an example instead of a number. Do not insist that people defend low ratings by praising the situation.

Before collecting accounts, explain who will see them, whether a manager participates and what the record will be used for. A small named group cannot be promised anonymity. Do not turn a developmental conversation into an undisclosed employment rating or sponsor report.

The ICF Code of Ethics provides relevant standards for agreements, roles and information exchange. Translate those principles into the actual audience and purpose. A promise that the discussion is “safe” cannot remove real power differences or guarantee what colleagues do afterward.

The HSE Role Management Standard emphasizes clear responsibilities, compatible requirements and routes for raising uncertainties or conflicts. That supports examining a responsibility problem directly instead of interpreting every disagreement as a poor relationship.

Use four stages to build the matrix

1. Hear the separate accounts. Ask each role what its number or words describe, which evidence matters and what remains uncertain. Check your summary with the person before filling a row.
2. Explore a relevant useful occasion. Ask about a specific comparable interaction, including what people did and what conditions helped. If no useful occasion comes to mind, do not demand one. Identify missing information or support instead.
3. Describe the next useful sign. Ask what the receiving role would need to assess the request responsibly. Make incompatible requests explicit. A sender’s wish for a date and a receiver’s need for assessment are not yet a shared commitment.
4. Agree a bounded test and review. Name the action, owner, receiver, actual acceptance, decision authority and review point. State what the test will not authorize. Keep unresolved questions visible.

A half-point or one-point shift can be shorthand for a client-defined sign, but first describe the sign. It is not a promise of an effortless improvement or a percentage change. A small sentence in a matrix can involve substantial work and another person’s consent.

For deeper investigation of an earlier occasion, use the separate exception scaling matrix guide. This article concentrates on the current shared task and the different role perspectives that must inform its next agreement.

Complete example: a pilot brief changes before delivery is authorized

All people, dates, requests and results below are fictional. Anya leads sales, Hugo product work and Kiran engineering assessment for a proposed customer pilot. Dana coordinates delivery decisions within the organization’s agreed process. Customer contact Leah wants a pilot on 20 November 2026.

Facilitator Nell agrees a forty-five-minute joint session on 10 November and a thirty-minute review on 24 November. All four participants agree the purpose and the named shared action record. They can use words or decline a rating. This is not an employee assessment or a customer-facing technical review.

Practical session arrangements use email to Nell; additional between-session coaching is not included. Customer messages, samples and technical work stay in the organization’s approved systems. The coaching discussion uses descriptions and an invented placeholder example, not real customer records or credentials.

Make the different meanings explicit

The agreed question is: “How usable is the current pilot brief for deciding what must be assessed before a delivery date can be authorized?” Zero means there is no usable common brief. Ten means the relevant roles have enough shared information to identify the next assessment, its owner and the decision route. Ten does not mean the product is ready for delivery.

Anya says seven because the customer has named a contact and a target date. Hugo says four because the first use case is unclear. Kiran says two because the material does not yet support deciding the assessment work. Dana chooses words: the date is not approved.

Nell checks each row aloud. Anya initially hears Kiran’s two as “Engineering is unwilling to help.” Kiran explains that he is describing the brief, not refusing the project. Anya accepts that correction. The disagreement becomes more specific without requiring matching ratings.

Use a previous occasion without assuming it will repeat

Hugo recalls a previous pilot that used a one-page brief and an example before anyone promised a date. Kiran says that made the unknowns visible. Anya notes that the previous customer’s request was narrower, so copying its delivery schedule would be unjustified.

They choose the useful element, a shared brief before commitment, rather than treating the earlier project as proof that the new pilot will be simple. Nell does not claim the exception reveals a universally successful process.

The discussion also reveals that Anya’s 9 November email called 20 November “confirmed,” although Dana had not authorized it. The action cannot merely say “communicate better.” Anya agrees to correct that exact status with the recipient.

Complete the first agreed test

The accepted actions are:

Anya sends the correction on 11 November:

“I described 20 November as confirmed. That was incorrect; it is your requested date and has not been authorized by our delivery process. We are clarifying the scope and assessment needed. I cannot replace it with another confirmed date today.”

Leah acknowledges the correction on 12 November but still wants 20 November. Her acknowledgment is not agreement to abandon the request or evidence that she is satisfied. Dana retains the unapproved status.

Hugo completes the first brief on 12 November. It describes reading a status value and explicitly excludes writing changes and transferring earlier records. Leah confirms that this describes the request as it then stands and supplies the invented example through the work route.

Kiran uses the accepted thirty-minute period on 13 November. He confirms that the material is sufficient to list further review questions. He does not approve implementation or promise an estimate. The three preparation actions have occurred, while the delivery decision remains open.

Keep the failed change handoff in the record

On 16 November, Leah changes the request: the pilot would also need to write a status change. Hugo notes this privately but does not update the shared version-one brief. Anya sends that old brief again on 17 November, and Kiran’s current account still concerns the status-only request.

The gap becomes explicit on 18 November when the roles compare Leah’s new question with their shared brief. The previous preparation was completed for an earlier request; it is not current agreement about the changed request.

Hugo takes responsibility for the missing shared update. He creates version two on 18 November, labels the writing request as new and unassessed, and retains version one as history. He does not overwrite it as though the new scope had always been agreed.

On 19 November, Hugo sends the changed brief and a clear change note to Anya, Kiran, Dana and Leah through the organization’s work route. Leah acknowledges the changed request on 20 November. That establishes the current requested scope, not its feasibility or acceptance for delivery.

Dana had accepted Kiran’s request for another protected thirty-minute read on 18 November, after he was told that the request had changed. Kiran uses it on 19 November to read version two and identify further assessment needs. He does not claim that an hour split across two versions completes the engineering work.

No pilot is delivered on 20 November. The earlier confirmation has been explicitly withdrawn, and the requested date remains unmet. The case does not describe that outcome as customer approval or a successfully completed implementation.

Review the matrix without inventing consensus

At the 24 November review, Anya declines another number and says that the customer’s request is clearer but the timing remains unresolved. Hugo chooses five because the changed scope is now shared. Kiran chooses three because the current brief supports identifying the next assessment, while substantial work remains.

They retain the original question and discuss the meanings. The ratings do not quantify a common increase, and Anya’s words are not a missing value to be filled with an average. Dana again states that no delivery date is authorized.

Nell initially proposes the closing phrase “The teams are now aligned.” Hugo objects: they share the changed request, but have not agreed feasibility, assessment resources or timing. Nell removes the overstatement. The completed shared review says:

“The original date correction and first brief review occurred. A later scope change was not shared promptly, and the old brief was resent. Version two is now acknowledged as the current request. Further assessment, its resources and any delivery date remain undecided. We have not measured overall collaboration or customer satisfaction.”

Dana, Hugo and Kiran accept an operational review for 3 December to decide the next assessment and its resourcing. It is not a delivery appointment. Anya accepts responsibility for explaining that distinction to Leah.

Anya emails the current state and next update date to Leah on 25 November. Leah acknowledges receipt that day, while retaining her concern about delay. The 3 December review, further assessment and any eventual delivery remain future events. Nell’s two agreed sessions are complete; future operational work does not imply additional coaching.

Use a complete short or extended agenda

The fictional first meeting used forty-five minutes. An original allocation is five minutes for purpose, authority and audience; eight for separate accounts; ten for a comparable occasion; twelve for the proposed agreement; five for receiving-role acceptance and dependencies; and five for owners, review and corrections. The total is 45 minutes.

For a more complex task, a ninety-minute option can allocate ten minutes to purpose and participation, fifteen to accounts, fifteen to useful occasions, five to a break, twenty to building and rehearsing the agreement, fifteen to difficult responses, five to the review plan and five to closing. That totals 90 minutes, including the break.

These are facilitation plans, not required clinical doses. Adjust participation and access arrangements with the people concerned. A written contribution, ordinary verbal account or suitable prior preparation can serve the purpose. Do not confuse a person’s willingness to speak in front of their manager with the completeness of their experience.

If a serious conduct, safety or grievance issue emerges, use the appropriate organizational route. Do not keep collecting more positive ratings as an alternative to addressing it. An operational matrix may need to pause or narrow its purpose.

Use a matrix template that preserves the receiving role

Build each row around six fields:

1. Role and current account: the person’s words or optional rating, with its meaning.
2. Evidence and period: what actual interaction or version supports the account?
3. Useful difference: what was relevant about another occasion, without assuming identical conditions?
4. Next sign: what would make the task more assessable or the decision clearer?
5. Accepted action: who actually agreed to send, receive or decide what, and when?
6. Limits and review: what remains unknown, who sees the record and what the test does not approve?

Keep a separate action record when several rows depend on one owner. That avoids counting the same shared brief as three independently completed outcomes. Preserve version dates and later corrections rather than rewriting a historical row.

A useful receiving response might be: “I can examine this version for assessment questions by Friday. That is not an implementation estimate.” The sender should check what the response actually accepts. A receipt, a review commitment and a delivery authorization are distinct states.

Organize preparation and follow-through in Coachful

Coachful is our #1 recommendation for organizing the original matrix preparation, sessions and selected follow-through. This is our editorial recommendation as its publisher. It does not turn personal ratings into a validated collaboration dashboard or make the app the authority for a customer commitment.

1. Prepare the actual question and audience. A worksheet can hold the six original fields. Explain which accounts will be discussed together and what stays in the organization’s systems. Do not upload technical credentials, customer records or sensitive personnel information merely to fill the matrix.
2. Assign the intended resource. Use Clients > the client > Worksheets & resources > Assign for the appropriate account. Copying a private link does not grant access. Staff preview does not submit a participant’s answer.
3. Check recap delivery before the discussion. Recaps send automatically by default. Sessions > settings gear > Session AI > Session recaps includes Approve before sending; each Session Type can override the workspace choice. Review the actual setting and intended audience rather than assuming every draft stays unsent.
4. Separate coaching notes from the shared agreement. Review the selected record, actual accepted actions and unresolved matters before any deliberate send. Holding a recap is not recording consent and does not authorize sharing every participant’s reflection with a sponsor.

The case uses explicitly named organizational work routes and live meetings, not proof that a real customer workspace was configured. Explore Coachful’s features for the service and record arrangements you actually intend to provide.

Troubleshoot the discussion

Someone interprets a low rating as blame. Return to the task and let the person explain what their account describes. Correct the facilitator’s summary if necessary. Do not demand a higher number to keep the meeting pleasant.

The proposed small change requires substantial effort. Name the real work and ask the owner whether they can accept it. Revise the scope or timing when needed. A half-point label does not create capacity.

A changed request invalidates the brief. Label the version and what changed, notify the relevant roles and establish the next review. Do not keep an old acceptance attached to new work without checking.

Confidence needs its own conversation. Ask a separate question rather than changing the meaning of the collaboration scale mid-record. Our confidence and action guide addresses readiness for a chosen action.

Frequently asked questions

What should I do with ratings of two and seven?

Ask what each person is describing, which evidence matters and what the receiving role needs. Preserve the accounts before deciding an action. Do not treat their average as an objective team result.

Can participants use words instead of numbers?

Yes. A clear account can serve the purpose without a rating. Do not impute a score to someone who declines or require a private explanation for that choice.

Does a half-point shift measure better collaboration?

It is a conversational label with the participant’s meaning. It does not establish a percentage improvement, equal change across people or a causal result from coaching.

Is a received brief an accepted delivery commitment?

No. Check what was actually accepted. A recipient may agree to read a version, ask for more information or identify a further approval. The shared record should preserve those limits.

What happens when a serious concern emerges?

Use the appropriate conduct, safety, grievance or other specialist route. Clarify the facilitator’s role and whether the coaching purpose needs to pause. Positive inquiry should not suppress the concern.

How often should the matrix be repeated?

When a real task or decision makes it useful. Review changes and agreements rather than collecting numbers because a weekly form exists. Stop using the matrix if it adds reporting without helping the work.