Raw

**To:** COO — **Fieldbook: extension to North and Islands**

You asked to see the two-quarter rollout review before the quarterly meeting. Across the four live regions Fieldbook has done what it was bought to do — jobs per technician are up from 4.6 to 5.3 a day, travel time down 18%, overtime down 12% — but only Central and East have paid back their running cost. South and West sit at break-even, and the reason is not the platform: it is that onboarding was a single classroom session per region and two known system gaps were never assigned an owner. North and Islands would add £13,500 a month in licences and would need winter travel for classroom training, so the same pattern there would be more expensive than it was in the West.

**Extend to North and Islands in Q1, but release the go-live only once onboarding is rebuilt and the two integrations are ordered — the economics are favourable at the Central and East adoption pace and negative at the West pace.**

## 1. The operating gain is real and repeatable, and it is what pays for the extension

- Route-aware assignment lifted completed jobs per technician per day from 4.6 to 5.3 and cut travel time between jobs by 18% across all four live regions.
- Emergency insertion, previously a dispatcher ringing round until someone accepted, is now automatic with a median of 4 minutes.
- Overtime is down 12% quarter over quarter and fuel per job down 9%, which finance attributes to the scheduling gains.
- Central and East recovered the running cost within the second quarter; the gain arrives with adoption, not with go-live.

## 2. Onboarding, not the platform, decides whether the new regions reach that pace

- Where onboarding held, adoption was immediate: Central and East hit full adoption in six weeks, and all 41 dispatchers were live in week one because the old spreadsheet was switched off on day one.
- Where it did not, adoption stalled: South needed eleven weeks to pass 80% of orders, West is still at 71% after nine weeks, and 34 of 118 West technicians never attended any session — the course ran once per region, so later hires and absentees learned second-hand.
- The gap resurfaces as cost elsewhere: 44% of the 1,940 support tickets are password and login problems concentrated among technicians who missed the course, and the improvised shadowing that covered for it consumed an estimated 380 technician-hours of lost route time in South and West.
- North and Islands are the two regions where a repeat is most expensive, since any classroom session there means winter travel. Rebuilding onboarding as a continuous, remote-first programme — the desk's weekly webinar already reaches 20 to 30 technicians and is best attended exactly where classroom slots were missed — removes the largest single driver of the West outcome.

## 3. Two integration gaps are being paid for in manual labour, and both have solutions waiting on our decision

- **No feed from the legacy asset register.** Technicians open every job with an empty equipment-history panel: they call the back office or ask the customer, and in 9% of audited jobs they replaced parts already replaced under warranty within the year. The same gap makes the panel's audit trail useless in disputes. The export was requested in month two and is still in the asset register team's backlog.
- **Completion codes do not match the invoicing system.** The billing team re-keys about 300 job records a week by hand, which added a temporary clerk at £8,400 for the two quarters, puts transcription errors onto customer invoices, leaves reporting and invoicing disagreeing on completed-job counts by 2–4% a week for finance to reconcile at month end, and generates a standing trickle of tickets the desk can only forward to the vendor. The vendor has a configurable mapping table that takes roughly a day to set up.
- **Customer search matches on exact postcode only.** Technicians who cannot find a record create a second one: 19% of all tickets are duplicates, the desk merges about 70 records a week, and affected customers' job history stays split until it is merged.
- The vendor's catalogue lists supported connectors for both the asset register database and the invoicing import format. Neither has been evaluated, because the programme was staffed for deployment and training only and no one on our side owns the integration. Both blockers are Kestrel decisions, not vendor work.

## 4. Nothing above was a surprise, which is the defect to fix before the next go-live

- Every negative pattern in this review was visible by week four of the Central go-live.
- The programme had no mechanism to turn early field observations into programme-level fixes, so each issue replayed in every subsequent region.
- Without a weekly review that can change the rollout while it runs, North and Islands will reproduce these findings a third and fourth time.

## What I need from you at the quarterly meeting

Approve the Q1 extension with three conditions before either go-live: a continuous onboarding programme replacing the single classroom session; a named owner for the Fieldbook integrations with the mapping table and the asset-register connector scheduled; and a standing week-four review with authority to change the rollout in flight.

Two smaller items were left out of the argument above: the parts-ordering screen is unpopular and mostly bypassed by phone, though it works correctly in testing and looks like habit; and the vendor relationship itself is sound, with response times inside contract.

---

```text
Top: Extend to North and Islands in Q1, releasing go-live only after onboarding
     is rebuilt and the two integrations are ordered.
  1. The operating gain is real and pays for the extension
     - 4.6 → 5.3 jobs/technician/day; travel time −18%
     - Emergency insertion automatic, median 4 minutes
     - Overtime −12%, fuel per job −9%
     - Central and East repaid running cost in quarter two
  2. Onboarding decides whether the new regions reach that pace
     - Central/East full adoption in 6 weeks; dispatchers live in week one
     - South 11 weeks to 80%; West 71% after 9 weeks; 34/118 never trained
     - 44% of 1,940 tickets are logins from untrained technicians
     - 380 technician-hours lost to improvised shadowing
     - Winter travel makes a repeat costlier in North and Islands
  3. Two integration gaps are paid for in manual labour and are solvable now
     - No asset-register feed: 9% of audited jobs re-replaced warranty parts
     - Code mismatch: 300 records/week re-keyed, £8,400 clerk, 2–4% variance
     - Postcode-only search: 19% of tickets, 70 merges/week
     - Vendor connectors and mapping table exist; no Kestrel owner assigned
  4. The absence of a feedback loop is what multiplied all of it
     - Every pattern was visible by week four of Central
     - No route from field observation to programme fix; issues replayed
```

Structurally: the review's six descriptive sections were replaced by four groups that answer the extension question — the proven gain, the condition that unlocks it, the cost of the two open gaps, and the mechanism that let both spread — with each defect consolidated into one branch and its evidence gathered from wherever it had been reported.