Appendix — the evidence
The data behind every number
Every figure in documents 01–07 traces to a row on this page. The agency keeps all of it, exportable.
A · The measurement grid — four weekly cycles, per step, in minutes
Office hands-on minutes per step, from the forms platform’s submission log, SharePoint’s file timestamps, and timed shadowing (July–August). The median row = the figures used throughout. The grid is wider than the page — it slides sideways.
| Week | 03 intake | 04 chase | 05 re-key | 06 review | 07 corrections | 08 form & filing | total | in by deadline | returned | filed |
|---|---|---|---|---|---|---|---|---|---|---|
| W1 | 70 | 295 | 410 | 145 | 235 | 225 | 23 h 00 m | 53/70 | 13 | +7 d |
| W2 | 75 | 330 | 420 | 150 | 255 | 225 | 24 h 15 m | 52/70 | 15 | +8 d |
| W3 | 80 | 370 | 435 | 155 | 290 | 235 | 26 h 05 m | 49/70 | 17 | +9 d |
| W4 | 75 | 330 | 420 | 150 | 255 | 220 | 24 h 10 m | 52/70 | 15 | +8 d |
| Median | 75 | 330 | 420 | 150 | 255 | 225 | 24 h 15 m | 74% | 15 | +8 d |
Steps 01–02 are the workforce’s and sit outside office minutes: ~45 m per worker per week, measured with the senior DSP and cross-checked against submission timestamps. "Filed" = the Sunday the week ends to the last packet reviewed, official, and filed. W3 was a mid-year plan-update week — more returns, longer review. The 16-day December week (document 02) was reconstructed from that month’s records and deliberately kept out of this grid: medians, no averages, so one bad week cannot distort the baseline.
B · Forms & spreadsheet inventory
Everything the weekly cycle runs on, as collected. "Last touched" from file metadata.
| Artifact | Kept where | State |
|---|---|---|
| The official weekly documentation — state revision 2025-11 | office master copy — the coordinator’s desktop | current |
| The same form, revision 2025-06 | still circulating with nine workers | retired · in use |
| The same form, revision 2024-01 | one worker’s phone | retired 2024 · in use |
| Forms-platform form set — daily time & activity, weekly sign-off, mileage, incident | the platform’s template library | touched 5 months ago |
| The pay sheet — "Payroll 2026.xlsx", one tab per period | Excel, office manager’s OneDrive | weekly |
| The billing sheet — "DDD Billing 2026.xlsx" | Excel, office manager’s OneDrive | weekly |
| The authorized-hours sheet — ceilings per participant per week | Excel, office manager’s OneDrive | on plan changes |
| The tracker — "Submissions 2026 FINAL(3).xlsx" | Excel, shared drive | every Monday |
Finding in passing: three revisions of the official form in circulation, and the tracker’s filename ends in FINAL(3). The build pins the official layout to one versioned place (rule R6, chapter 08) and retires the tracker entirely.
C · Interview log
On small screens the log is wider than the page — it slides sideways.
| Session | With | Length | Form |
|---|---|---|---|
| Initial interview | the owner | 60 m | video call · recorded, transcribed, consented |
| Deep dive — day 1 | office coordinator | 75 m + shadowing a Monday | recorded, transcribed, consented |
| Deep dive — day 1 | office manager | 50 m | recorded, transcribed, consented |
| Deep dive — day 2 | senior DSP | 45 m, after shift | call · notes, consented |
| Log replay | coordinator + office manager | 2 × 60 m | screen sessions over the submission log and SharePoint timestamps |
D · Method note
Measured: office hands-on times from the submission log and SharePoint file timestamps where steps happen in-system; off-system steps (paper drops, calls, the Sunday reconstruction) timed live during shadowing and cross-checked with the people who do them. Medians used throughout; no averages (the December week would distort them).
Estimated, and marked as such: hourly rates (the agency’s own loaded figures — coordinator $28/h, office manager $34/h), the out-of-scope item sizes in document 06 (owner estimates — recorded, not audited), and option A’s software pricing (vendor list price at the agency’s user count).
Not measured: participant identities or the contents of care notes — the audit reads how documentation moves, not what it says. Packets were observed inside the agency’s own systems and never copied out; the data-handling agreement signed before the deep dive governs all of it.
Proposed-state times were engineering estimates until the parallel run and production measured them — chapters 08 and 09 report the measured result against this baseline.
Review: every figure and finding in this workspace passed a business and a technical review by two partners before delivery.
open daily · 24/7 · +1 (786) 370-9416 · hello@slashanalog.com · privacy notice