On-site delivery: making requirements clarification workshop stick

Context: This write-up is for colleagues progressing project delivery. The client is in wholesale & retail, roughly two hundred+ staff, based around economic development zones. We hit a case of metrics not matching finance numbers. Troubleshooting started with logs on object storage, then followed our reproduction steps. Prefer a go-live window outside peak traffic, and prepare the rollback package early.

Background & Goals

This write-up is for colleagues progressing project delivery. The client is in home & building materials, roughly about thirty people, based around the Middle East. Business asked for a clickable prototype within two weeks, while engineering wanted to lock fields related to requirements clarification workshop first. Archive screenshots and config change IDs for this requirements clarification workshop round so next month's retrospective stays concrete. Last week we reviewed acceptance checklist with the client's IT lead. They care more about stability than buzzwords. QA walked the main path three times on real devices and flagged edge cases around acceptance checklist. Prefer a go-live window outside peak traffic, and prepare the rollback package early.

Risks & Mitigations

Last week we reviewed go-live rollback plan with the client's IT lead. They care more about stability than buzzwords. QA walked the main path three times on real devices and flagged edge cases around go-live rollback plan. Archive screenshots and config change IDs for this go-live rollback plan round so next month's retrospective stays concrete. Last week we reviewed backup and recovery drills with the client's IT lead. They care more about stability than buzzwords. Editors know the CMS well enough, but backup and recovery drills still needs a one-page rule sheet to cut verbal rework. Add a field dictionary to the docs so new joiners ask fewer repeat questions.

Testing & Acceptance

The same requirement can sound different across departments; writing it down helps align decisions on integration environments. Business asked for a clickable prototype within two weeks, while engineering wanted to lock fields related to integration environments first. Add a field dictionary to the docs so new joiners ask fewer repeat questions. Last week we reviewed acceptance checklist with the client's IT lead. They care more about stability than buzzwords. QA walked the main path three times on real devices and flagged edge cases around acceptance checklist. Next, we recommend to have business users record a screen walkthrough. If more features are added, scope the blast radius before scheduling. Last week we reviewed data migration windows with the client's IT lead. They care more about stability than buzzwords. Editors know the CMS well enough, but data migration windows still needs a one-page rule sheet to cut verbal rework. Next, we recommend to log API status codes and latency. If more features are added, scope the blast radius before scheduling.

Delivery Approach

Last week we reviewed prototype sign-off checkpoints with the client's IT lead. They care more about stability than buzzwords. Editors know the CMS well enough, but prototype sign-off checkpoints still needs a one-page rule sheet to cut verbal rework. Prefer a go-live window outside peak traffic, and prepare the rollback package early. The same requirement can sound different across departments; writing it down helps align decisions on prototype sign-off checkpoints. QA walked the main path three times on real devices and flagged edge cases around prototype sign-off checkpoints. Add a field dictionary to the docs so new joiners ask fewer repeat questions.

Takeaways: The same requirement can sound different across departments; writing it down helps align decisions on invoice and payment checkpoints. Editors know the CMS well enough, but invoice and payment checkpoints still needs a one-page rule sheet to cut verbal rework. Prefer a go-live window outside peak traffic, and prepare the rollback package early.