Run mobile launches like a conductor: score, rehearsals, cues
Treat a launch like an orchestra: one conductor's score, section rehearsals, explicit cues and a tempo check one week out to avoid timing failures.

Who this is for: product managers owning consumer or mobile features that cross product, design, engineering, QA, marketing and support with a tight launch window. At the end you will have a simple conductor's score template, a rehearsal plan and an explicit cue checklist to use in launch week.
In brief
The conductor's score
Do not write another task list. The score is a single page that shows who does what and when, how systems must align, and where the decision points are. It sits above JIRA and the sprint boards.
What to include:
- Milestones with exact times, not just days.
- Dependencies: which service must be ready before the SDK rollout starts.
- Cues: explicit actions tied to measurable conditions.
- Fallbacks and rollback triggers.
Make a one-page version for stakeholders and a detailed score for teams. A One-pager works well for the summary card you hand to execs.
Steps to prepare the score
1. Map the timeline
List each team's key delivery with times. Include marketing sends and support hours.
2. Mark interfaces
Note every handoff: SDK init, feature gates, token exchange, API contract checks.
3. Write cues
For each handoff, write precise cues and tolerances. Example cue: disable feature flag X if payment error rate > 2% for 10m.
Rehearsals: sectionals then full dress
Run sectionals first. Those are team-only rehearsals to validate their piece of the score. Then run a full dress rehearsal with all teams present and the conductor (the PM) in charge of cues.
Sectionals are fast. They focus on the interfaces and test data. Use production-like tokens where safe.
Define cues and tolerances
The goal is autonomy. Teams must act without waiting for instruction. Cues must be specific and measurable.
Good cue format: trigger condition → immediate action → escalation path. Example: "if SDK initialization failure rate > 5% for 5m, rollback SDK version and notify infra channel X."
Tempo check — one week before launch
A tempo check is a short run to validate metrics, mock traffic and rollback paths. It is not a full dress run. It validates that monitoring, alerting and runbooks work under load.
- Run synthetic traffic through critical flows.
- Validate alarms and on-call paging.
- Confirm feature gate controls and rollback steps.
Freeze and final dress rehearsal
Freeze noncritical merges 48 hours before the final dress rehearsal. The dress rehearsal should use the production feature gates and be the last point where teams can refuse go/no-go.
"A score gives teams permission to act. Vague instructions slow the launch."
Comparison: task list vs conductor's score
| Aspect | Task list | Conductor's score |
|---|---|---|
| What it shows | Individual tasks and owners | Timeline, interfaces, cues and fallbacks |
| Timing | Loose dates | Exact times and sequencing |
| Autonomy | Low — teams wait for instructions | High — explicit actions at cues |
| Risk on launch day | Hidden until late | Exposed early via rehearsals |
Checklist before launch
- One conductor's score shared with all teams.
- Sectional rehearsals completed and issues logged.
- Full dress rehearsal completed using production gates.
- Tempo check passed one week out.
- Noncritical merges frozen 48h before final dress.
- Cue list published with escalation channels and owners.
Common mistakes to avoid
- Making the score a task list with no interfaces or cues.
- Skipping sectionals and relying only on a full dress run.
- Writing vague cues like "notify team" instead of specific actions tied to metrics.
- Over-rehearsing to the point teams stop treating the launch as real.
Worked example: mobile checkout launch
Conductor's score (summary)
- SDK rollout cadence: staged rollout at 09:00, 10:00, 11:00 UTC. Each stage lasts 30 minutes with health checks.
- Backend feature gates: enable payments_api_v2 at 08:30 UTC behind a feature flag; verify contract tests at 08:45.
- Marketing sends: promotional push scheduled for 11:15 UTC after the 11:00 SDK stage completes and payment success rate > 98% for 15m.
- QA sign-off cues: QA approves each staged SDK after passing end-to-end payment flow with production-like tokens.
- Rollback triggers: disable feature flag X if payment error rate > 2% for 10m, or if latency p95 increases by defined threshold in runbook.
Rehearsal plan:
- Sectionals: payment team runs tokenized payments; SDK team validates init across device list; marketing rehearses the send timing.
- Full dress: simulate staged rollout with mock traffic, validate alerts and rollback. All teams join the call with read-write access to the score.
- Tempo check (7 days out): run peak synthetic traffic on the critical checkout flow and exercise rollback steps end-to-end.
- Freeze noncritical merges 48h before the final dress rehearsal.
After the launch: the conductor's debrief
Run a structured review within 48 hours. Measure three things:
- Phrasing: user flows — did users move through the checkout as intended?
- Tempo: latency and throughput — did the system keep the intended pace?
- Ensemble balance: were any teams overloaded during launch and handoffs?
Capture improvements to the score, the rehearsal plan and the cues. Publish the revised score for the next launch.
Use the conductor approach for any multi-team mobile launch. If you need a single-page summary to hand to execs, create a One-pager and attach it to the score.


