You have your methodology decision. You have board sign-off in principle. Now you actually have to run the programme. This guide is the operational playbook — what to do, when, and who needs to sign what — to get from kickoff to published TSMs without sliding the timeline and without the Regulator finding gaps in your summary of approach. It assumes you have already read the cornerstone TSM guide (the regime, the 22 measures, the compliance pitfalls) and the methodology deep-dive (which method to choose for your stock profile).
Plan ~12 weeks end-to-end from kickoff to NROSH+ submission, plus 4–6 weeks of pre-launch if you are selecting a vendor or revising your questionnaire. The week-by-week breakdown below is the pattern that survives first contact with a real tenant population.
The 12-week timeline at a glance
| Phase | Weeks | Primary owner | Decision gate at end |
|---|---|---|---|
| Pre-launch | −6 to 0 | Insights / Research lead | Board sign-off on methodology + questionnaire |
| Fieldwork launch | 0 to 2 | Operations + vendor | First-wave response rate review |
| Fieldwork main wave | 2 to 6 | Operations + vendor | Quota representativeness check |
| Fieldwork tail + follow-up | 6 to 8 | Operations + vendor | Cut-off decision |
| Analysis + weighting | 8 to 10 | Insights / Research lead | Internal sign-off on weighted figures |
| Reporting + publication | 10 to 12 | Insights + Comms + Board | Board approval + NROSH+ submission |
Phase 1 — Pre-launch (weeks −6 to 0)
Week −6: Methodology lock
Confirm the survey mode mix you will use this cycle (informed by the methodology decision framework). Document the rationale in a one-page note that will become the opening of your summary of approach. This is the moment to flag any mode change versus last year — the Regulator and the Board both expect rationale, not surprise.
Week −5: Sample frame extract
Pull tenant data from your housing management system: tenancy ID, household composition, stock type (general-needs / sheltered / supported), build type, tenancy start date, geographic patch, language flag where held, contact preferences (postal address, phone number, email address, opt-out flags). Lock this snapshot — fieldwork must reconcile back to this list. Tenant churn during the programme is the most common cause of week-9 data-quality headaches.
Week −4: Vendor selection or in-house ramp
If outsourcing, this is the latest point to brief vendors and confirm contracts. Five questions every vendor must answer (from the methodology deep-dive): verbatim wording fidelity, mode-mix transparency, representativeness weighting approach, year-on-year defensibility, recording and audit trail. Get these in writing.
Week −3: Questionnaire finalisation
Use the verbatim wording from the RSH Tenant Survey Requirements (March 2023, updated April 2024). The ordering rules matter — TP01 first, TP02 before TP03 — and no warm-up questions. The Stoke-on-Trent and Cannock Chase exclusions in 2024/25 came from materially altered wording. Pilot the questionnaire on 20–50 tenants in the same week to catch wording problems before main fieldwork.
Week −2: Tenant communications + accessibility
Draft pre-notification letters and emails. Translate into the languages relevant to your portfolio (the Regulator’s representativeness expectations explicitly include ethnicity / language as a characteristic to test). Coordinate with your residents’ comms team — surveys that arrive on the same week as a rent-increase letter perform badly.
Week −1: Board sign-off
Board approval (or delegated-committee approval) of: methodology, questionnaire, sample frame, vendor appointment, communications. This is the gate before fieldwork — the Regulator’s text says Board responsibility is “ultimate” for accuracy, so an explicit pre-fieldwork sign-off is hard to leave out.
Phase 2 — Fieldwork launch (weeks 0 to 2)
Week 0: Launch
Send invitations. For mixed-mode designs, push-to-web invitations go out first (with a 14-day window). For postal-only designs, the main mailing leaves this week. Set up daily response-rate dashboards from day one — the biggest operational mistake is monitoring weekly when first-week patterns predict the final outcome.
Weeks 1–2: First-wave reading
By the end of week 2 you should know whether you’re on track. Compare achieved response counts to your forecast for each quota cell (age band, stock type, geographic patch). If a cell is materially behind, flag it now — do not wait until week 6 when remedial fieldwork compresses your analysis window. Vendors should be producing daily or weekly response reports automatically.
Phase 3 — Fieldwork main wave (weeks 2 to 6)
Weeks 3–4: Reminders + postal sweep
For push-to-web designs, this is when the postal sweep goes out to non-responders. For postal-only, this is when reminder 1 is mailed (reminder 2 follows in week 5). Telephone backstop calls begin for under-represented quotas identified in the week-2 reading.
Weeks 5–6: Quota fill
Active management of under-represented cells. AI-phone or human telephone is best deployed here — targeted at specific demographic gaps rather than blanket coverage. Hold a weekly cell-by-cell review with the vendor; cut cells that have hit target so reminder spend doesn’t over-shoot.
Phase 4 — Fieldwork tail (weeks 6 to 8)
Week 7: Final-week chase
Final reminders, final phone calls, final email nudges. The marginal response in week 7 typically comes from the hardest-to-reach cells — sheltered, supported, ESL — so this is where your translated and accessible materials earn their keep.
Week 8: Cut-off
Hard cut-off date. Document the date in the summary of approach. Any response received after this date is excluded from the headline TSM but can be retained in the operational dataset. Communicate the cut-off to the vendor explicitly — accidental post-cut-off responses creep in and create reconciliation issues later.
Phase 5 — Analysis and weighting (weeks 8 to 10)
Week 8: Data cleaning
Reconcile responses back to the locked sample frame. Apply the household cap (one response per household). Resolve partial responses (count for the questions answered, exclude from the questions skipped). Flag any responses where TP02/TP03 or TP09/TP10 filter inconsistencies are apparent (e.g., respondent answered the repair-satisfaction question after saying they had no repair).
Week 9: Weighting
Calculate weights against the six representativeness characteristics named in the Technical Requirements: stock type, age, ethnicity, building type, property / household size, geographic area. Apply weights consistently across all TSM perception measures (the Regulator’s text is explicit on this). Document the weighting approach in the summary of approach — do not just describe it as “weighted”.
Week 10: Internal sign-off
Insights / Research lead signs off the weighted TSM figures. Cross-check that LCRA and LCHO have been calculated separately where both are owned in material volume. Flag any management-information measures (BS01–BS05, RP01, RP02, CH01, CH02, NM01) that are at risk of late delivery from finance / repairs / complaints — these come from operational records and often need chasing in parallel with the survey.
Phase 6 — Reporting and publication (weeks 10 to 12)
Week 10: Summary of approach drafting
Write the summary of approach in parallel with the headline report. Every required element must be present: achieved sample size, survey timing, collection method(s), sampling method, representativeness assessment, weighting applied, named external contractors, households excluded with reasons, reasons for not hitting minimum samples, any incentives used, and methodological issues with material impact.
Week 11: Board approval
Board-level approval of: the headline TSM results, the summary of approach, the publication artefacts (website page, downloadable PDF, tenant communications). The Regulator treats Board approval as the accountability checkpoint, so document the approval minute and date.
Week 12: Publication + NROSH+ submission
Publish on your own website first. Include the questionnaire that was actually used (this is a requirement, not a nice-to-have). Submit to the Regulator via NROSH+ within the annual window — the exact date is set in the CEO data-requirements letter each year. Communicate the published results to tenants via the channel that worked best in the survey wave (often a tenant newsletter or post-survey email).
Anti-patterns we see most often
- Sample frame drift.Tenant data refresh during the campaign without re-running quota calculations against the new frame. Lock the frame at week −5 and reconcile in week 8 — don’t move during fieldwork.
- Late representativeness reading. Discovering at week 6 that a quota cell is critically under-represented. The remedy — emergency phone follow-up — works but compresses analysis. Build the check into week 2.
- Summary of approach written last. Drafting it in the final week before publication leads to gaps and Board push-back. Start drafting in week 10 in parallel with weighting.
- Management-information measures out-of-sync.The 10 MI measures come from operational systems, not the survey. Run a parallel workstream so they’re ready when the perception data is.
- Tenant communications collision. Survey arrives the same week as a rent increase or a service-charge letter. Response rates drop materially. Coordinate with the comms team on a rolling 8-week view of outbound tenant communications.
“You should be hitting 100% compliance on all of your safety indicators. If you’re not… we will come and ask.” — Will Perry, Director, RSH
The quote above is about the BS01–BS05 building-safety measures, but it applies to the operational discipline of the whole programme. The Regulator’s questioning starts where the published data has gaps — fix the process, not just the number.
Sources and further reading
- TSM cornerstone guide — the regime, the 22 measures, scope thresholds, full compliance pitfalls.
- Methodology deep-dive — postal vs phone vs online vs AI cost-per-response, mode-effect mechanics, decision framework.
- RSH Tenant Survey Requirements — verbatim question wording, response scales, ordering rules.
- RSH Technical Requirements — sample size accuracy bands, six representativeness characteristics, weighting requirements.