← Team Lead: coordination, people, and outcomes
10 / 10 · 60 MIN

Cross-team disagreement and learning from outcomes

Represent positions accurately, distinguish review from approval, and interpret outcomes without overstating conclusions.

Build a shared problem

APS wants to defer an intervention because coverage is missing; ADM wants to meet a date promised to the business. Saying one team blocks work or the other ignores operations hides relevant objectives. Ask each side to explain the other’s position in terms that side recognizes as accurate. Then identify the needed decision, options, consequences, and missing information. The Team Lead facilitates this preparation; decision authority needs to be explicit. If agreement does not emerge within the useful timeframe, inform the parties and bring the problem to the shared decision maker designated by local process. Timing suggested in a provider playbook is not a bank SLA. These examples concern planned decisions; an active incident follows its applicable response process. Escalation should enable a decision rather than punish disagreement. Preserve the rationale so later changes can be discussed with context.

Handle feedback without inventing approval

A draft review session needs context, audience, and questions. If APS requests operational detail and the sponsor requests brevity, the requests may be compatible through a summary with a linked appendix. The author should show treatment of accepted, deferred, and unresolved suggestions. Session participation does not justify writing “validated by everyone”. Confirm necessary validation and who can provide it. After a valid decision, documented disagreement need not be erased to execute. Clarify owners, conditions, and signals justifying reopening the decision. Do not prohibit new information merely because the issue was discussed earlier. In international communication, a short note can distinguish the selected option, the remaining concern, and the review point. Brevity should preserve these distinctions rather than turn questions into apparent consensus. Feedback has value even when contributors do not prefer the same document structure.

Distinguish observation, effect, and explanation

A local incomplete-handover measure records two out of twenty and then one out of ten: 10% in both samples. The count fell, but the observed proportion did not change. In another example, rework changes from four out of forty to one out of twenty, from 10% to 5%. If requests also became simpler, the decrease does not isolate the new guide’s causal effect. Keep denominator, definition, and composition visible. These measures are invented for the exercise rather than alternative names for DORA metrics. A service outcome also does not automatically assign credit or failure to one person on duty. Seek contextual evidence before reaching that conclusion. When reviewing improvement, publishing a form demonstrates artifact delivery; observing correct, usable contacts in subsequent handovers moves analysis toward the intended effect. Learn from choices and conditions without erasing follow-up responsibility.

Disagreement session and decision note

Prepare 35 minutes with APS and ADM representatives, a facilitator, and an observer. During five minutes, each side reads its objective and constraint. In the next ten, the facilitator gathers options and requests confirmation of positions. Then introduce a decision by the common owner and use ten minutes to record conditions, owners, and the remaining concern. During the final ten, present the two rework datasets and revise an overstated causal claim. Deliverables are an escalation note, a decision record, and a data interpretation with limitations. The observer looks for faithful representation, a distinction between feedback and approval, correct arithmetic, and no invented causality. This guide has no passing points or certification. The session has not been conducted with human participants; examples and arithmetic received editorial review without observation of actual negotiation. Repeat the exercise after feedback to compare specific behaviors rather than global impressions.

Decision needed: choose a window with understood coverage and impact.
APS position: preserve response capability within the window.
ADM position: meet the identified business commitment.
Options: retain with confirmed cover; defer with agreed impact.
Decision owner: role defined in the local process.
Status: feedback collected; decision not yet confirmed.
IN PRACTICE

Decision note: option B selected by the owner; APS retains a coverage concern. Before execution, confirm substitute and contact. Reopen if agreed coverage ceases to exist.

Common pitfalls

Escalating only one’s own position; participation treated as approval; more activity treated as effect; a lower rate treated as proven causality.

Related topics: Decision escalation · Improvement and metrics

Take this idea with you

A useful decision preserves positions, authority, and conditions; a useful conclusion respects what data and observation actually establish.

Create account

Reference: Clean escalations · Atlassian Team Playbook, GitLab leadership handbook, Google SRE, Kanban Guide May2025 and DORA five-metric model; inspected 2026-10-01