Turning Complexity into Clarity
Can your depot actually maintain the fleet it's been given?
Willow answers in hours, berths and competent staff — built on the rail industry's own depot performance method, and it challenges the numbers it's fed instead of trusting them.
No account? Sign up
Already have an account? Log in
Your account has been created. The administrator needs to approve you before you can access Willow.
Back to login
© 2025–2026 Lee Lockwood · Willow™ · Terms of use · Privacy
Willow is proprietary software. © 2025–2026 Lee Lockwood. All rights reserved. Access is provided under licence for evaluation and planning use by authorised users; you are granted a right to use the service, not any right in the software itself.
You agree not to copy, download, scrape, decompile, reverse-engineer or reproduce the software or any part of its calculation method; share login credentials; or use the service to build or assist a competing product.
Your data stays yours. Operational data you enter belongs to you or your employer. We claim no rights over it beyond what is needed to run the service.
Planning support, not an instruction. Willow's outputs are decision support. They do not replace your organisation's safety, engineering or operational approval processes, and must not be used as the sole basis for safety-critical decisions.
No warranty during evaluation. The service is provided as-is during the evaluation period, without warranty of any kind, and liability is limited to the fullest extent the law allows.
Licensing enquiries: lockwoodlee7@googlemail.com
What we hold: your name and work email (to run your account), the operational planning data you choose to enter, and basic usage/audit events (sign-ins and saves) kept for security.
Where it lives: with our hosting providers (Netlify for the application, Supabase for accounts and data). Your organisation's data is isolated from every other organisation's at the database level.
What we never do: sell your data, share it with other customers, or use one operator's data to inform another's results.
Your choices: ask us to export or permanently delete your account and data at any time: lockwoodlee7@googlemail.com
--
Turning Complexity into Clarity
Select your operator to load depots, fleet allocations and maintenance data. You can add, remove and cascade fleets once inside.
Or configure a custom operator manually
You can add more depots later. Start with the one you know best.
Class, number of units, and typical weekly miles per unit. These drive exam demand.
Headcount on the tools and one main shift. Add more roles and shifts later in Resources.
Honesty note: exam durations and intervals start from industry defaults so you get a number today. Calibrate them to your own working times in the Tasks tab — that's what makes Willow accurate for your depot.
Walk the shed. Tap what you see.
Willow works out how much maintenance your depot has to do, how much your team can actually deliver, and where the gap is — then says it in numbers a director can act on.
Its capacity model follows the method in RDG-ENG-GN-009, the Depot Performance Handbook, so the figures it produces can be defended in a capacity review rather than just asserted.
Four questions — depot, fleets, team — and a real capacity answer in five minutes. Correct everything afterwards.
A worked multi-depot dataset with every screen alive. Replaces what is loaded — Export first if you have work to keep.
Download the Willow template, fill it from your own systems, upload it back. Or paste your Rules of the Plan straight in.
Three ways in, depending on what you arrive with. Any of them can be undone — use Export on the Resources tab first if you have work you want to keep.
Sign in to an empty account and Willow opens a four-step wizard: your depot, your fleets, your team, and then a real capacity answer from the live engine. Five minutes, and you can correct everything afterwards.
Use Load sample data at the top of this page. It fills every tab with a worked six-depot network so you can see what a finished model looks like before typing anything. The depots and figures are invented.
Best start. Open Stabling › Import roads from your Rules of the Plan and paste your road table straight out of Word, Excel or a PDF. Willow reads the columns and shows you what it understood before importing anything.
Five places. The first three drive every number; the last two make the depot-capacity side work.
| Tab | What goes in | Why it matters |
|---|---|---|
| Fleet | Units and vehicles per class at each depot, weekly mileage, defect backlog hours | Mileage drives how often every exam falls due. Get this right first. |
| Tasks | Hours per task and mileage interval, per exam type, plus the work-arising buffer | Ships with industry defaults. Replace them with your own figures — they are the biggest single source of error. |
| Resources | Roles and headcount, shift patterns, sickness, leave, training, efficiency | Efficiency is the biggest lever in the whole model. Measure it rather than guessing — see below. |
| Stabling | Roads with their lengths in metres, who controls each one, limitations, and what stables overnight | Length is what makes the Appendix B calculation work. Without it Willow falls back to a typed unit count. |
| Upload | Diagram dashboard and end-of-day stabling report, as Excel | Saves retyping mileage and overnight vehicle counts each period. |
Weekly miles ÷ task interval × hours per task, across every fleet and every task type, plus defect backlog, fuel-point work and the work-arising buffer.
Gross shift hours cascaded down through sickness, leave, training, breaks and meetings, efficiency and siting delay, to the hours actually available on the tools.
Capacity minus demand, converted into the people it would take to close it and what that costs per week.
Road metres turned into units that can physically berth, by the Appendix B method, then tested against what actually stables there overnight.
| If you want to know | Go to |
|---|---|
| Can we deliver the plan, and where do the hours go? | Analysis |
| How many B exams can we sustain, and what would fix it? | Depot Model |
| Something to put in front of a director | Director’s Summary |
| Will the units physically fit tonight? | Stabling |
| What happens if the timetable, the fleet or the staffing changes? | Scenarios |
| Which depot across the network is under most pressure? | Depot Compare |
| Is it getting better or worse over time? | Scenarios › trend history |
| A report to send someone | Export |
Efficiency is the single biggest lever in the capacity cascade, and almost everyone starts with an estimate. Analysis › Efficiency Study replaces it with a measurement: walk the depot at random intervals and record whether each technician you see is on the tools or not. The ratio is your real figure.
Willow shows the sampling error as you go, so you can see when you have enough observations to trust it — roughly a hundred, though a lopsided split settles sooner. Applying it records that the figure was measured, by whom, and on which shift.
Willow installs to a home screen and runs full-screen with no browser bar. iPhone or iPad: Share, then Add to Home Screen. Android or desktop Chrome: the install icon in the address bar, or the menu. No app store, no download.
On a phone you will see a Depot floor mode button at the bottom of the screen. It puts the eighteen tabs away and leaves three things: whether tonight’s units berth, this week’s gap, and two large buttons for running an efficiency study as you walk. The button switches you back.
Willow keeps a flight recorder of what it saved and loaded, and it survives a refresh. Admin › Diagnostics › Copy diagnostics puts it on your clipboard — send that with any problem report and it is usually enough to find the cause.
The save chip in the corner tells you most of what you need on its own:
| An old version number | Your browser is holding a stale copy. Hard-refresh. |
| “Shared data loaded” with no revision | It loaded the wrong record. Sign out and back in. |
| Revision goes backwards | Another tab or device is writing too. Close the others. |
| Same revision but the screen differs | The data is fine and a control is out of step. Refresh the page. |
Willow marks anything it had to assume as PROVISIONAL, on screen, where you will see it. Exam intervals, task hours and vehicle lengths all ship as industry-typical defaults, and every one of them should be replaced with your own figures before a number leaves the building.
The same applies to depot capacity: where a road has no recorded length, Willow says so rather than quietly inventing one. A figure that is derived is always labelled as derived.
For each fleet at the selected depot: weekly miles ÷ task interval = tasks per week, then tasks per week × hours per task = weekly hours. Summed across every task type and every fleet, then added to defect backlog, daytime fuel-point work, and the work-arising percentage that covers what an exam finds once it is opened up.
Gross shift hours are reduced in order: sickness, then annual leave, then training, then mandatory inefficiencies such as briefings and breaks, then the efficiency factor, and finally siting delay — the time lost moving units to where the work happens. What is left is net productive hours. Each step is shown separately so you can see where the hours go.
Net hours minus all non-B-exam demand, divided by average B exam hours. This is the number of B exams the depot can actually sustain each week — usually the figure that matters most, because B exams are what fall over first when capacity is short.
Capacity minus demand. Negative is a deficit. FTE needed = deficit hours ÷ net hours per person. Cost is shown on an overtime basis, which is deliberately conservative; agency cost is listed separately for comparison rather than mixed in.
Appendix B of the handbook sets out a framework for determining the capacity of a depot. Willow implements its stabling steps, and the Stabling tab shows the chain rather than just the answer:
That last step is why a unit count on its own misleads. A four-car unit in a six-car road wastes two cars of road, and the handbook is explicit that a depot is effectively full well before it is physically full. Willow plans to a practical occupancy you can set yourself, and keeps a road clear for a failed train and for running round — warning you when neither is recorded.
Willow uses the industry's own words. If you don't live in a depot, this is what they mean.
Willow
Turning Complexity into Clarity
Configure task hours and intervals for fleets at this TCC
Add depot-specific maintenance tasks beyond the standard 15
Set hours and intervals for standard tasks. Use Custom Task Builder above to add depot-specific tasks (PIVTS, etc.)
|
Inc |
Fleet |
B Exam |
A Exam |
A2 |
A3 |
FP/Svc |
UFC |
Tyre Turn |
PHC ⏸️ |
W/Set ND |
W/Set D |
W/Set Full |
Engine/Raft |
Transmission |
Panto |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Interval | Interval | Interval | Interval | Nightly | Interval | Interval | Interval | Interval | Interval | Interval | Interval | Interval | Interval |
Predicted task volumes based on fleet mileage and maintenance intervals. Validated against Maximo: B exam prediction accuracy ±2%.
Tasks per week for each fleet class at this TCC -- scroll right for component changes
| Fleet | B | A | FP | UFC | Tyre | WS-ND | WS-D | Eng | Transmission | Panto | Hrs/wk |
|---|
Planned heavy component changes per week -- use this to size a dedicated heavy maintenance team
Every task type rolled up across all fleets -- occurrences and hours per week/day
| Task | Per Week | Per Day | Hours / Week | Hours / Day | % of Demand |
|---|
Import weekly operational data to auto-populate fleet mileages and night shift arrivals
Diagram_Dashboard_-_Dec_2025_-_WC_22_Feb.xlsxEnd_of_day_locations___stabling_report_-_Dec_2025_-_WC_220226.xlsx
Every operator's internal reports look different, so Willow doesn't chase formats: download the template, fill the Fleets, Roads and Staffing sheets from your own systems, and upload it back. Plain language, no mapping, problems named by row. Only the depots in your file are touched.
Weekly fleet mileage data -- updates the Fleet tab with actual miles run per fleet per TCC
Drag and drop or click to select your Diagram Dashboard .xlsx file
End of day locations -- provides nightly arrival counts for fuel points and service checks
Drag and drop or click to select your Stabling Report .xlsx file
Manage fleet allocation and operational data for this TCC. 18 fleets configured. Use the TCC selector above to switch depots.
Define each shift at this depot -- start/end times, staff count, and days per week. The capacity calculation sums across all shifts. Add or remove shifts to model different patterns (2-shift, 3-shift, 24hr coverage etc).
Configure staffing by role for this location. Add or remove roles as needed -- each depot can have its own structure.
Compare actual work completed against your calculated capacity to validate your efficiency setting
Enter exam counts by fleet type -- hours are pulled from your Task tab automatically
Automated cross-checks to ensure input accuracy and prevent narrative-driven manipulation
All parameter changes are recorded for transparency
How many staff hold each task competency — identifies skill bottlenecks
28-day cycle showing team deployment — click a pattern to apply
Forecast hours per week that sit outside exams and reactive defects. These feed demand directly — no work-arising uplift is applied, as they are already forecast figures.
Complete workload picture for this TCC
Where your workload comes from:
Configure your staffing levels in the Resources tab to see detailed capacity analysis.
Auto-generated insights based on your current data
Projection based on current backlog and weekly clearance allocation
Forward projection of heavy task demand based on mileage accumulation rates
Theoretical availability based on maintenance dwell times — per fleet class
Record key information for the incoming shift
Record actual weekly overtime to validate the model
Activity sampling to measure actual wrench-time efficiency on shift
How it works: Walk the depot at random intervals throughout the shift. For each technician observed, record whether they are doing productive wrench-time work or non-productive activity. After enough observations (aim for 100+), the ratio gives you a statistically valid efficiency percentage. This replaces the estimated figure in the Resources tab with a measured one.
No observations recorded yet.
Unit journey through the depot — time cost, staff, and flagged inefficiencies
Diagram_Dashboard_-_Dec_2025_-_WC_22_Feb.xlsx
End_of_day_locations___stabling_report_-_Dec_2025_-_WC_220226.xlsx
Per-shift capacity, B exam throughput, and recovery scenarios for —
Drag the sliders to see how changes affect B exam throughput
1,000 randomised weeks -- what's the probability you can deliver?
Hard limits from depot infrastructure that cap throughput regardless of staffing
Defects from exams, driver reports, and daily checks. Configurable per depot.
Planned exam volumes, units visiting, and hours required -- by fleet type. Aggregates from weekly through to annual with heavy maintenance flagging.
Each row shows one fleet type. Units/vehicles visiting = how many need to attend the depot in the period. Heavier tasks are flagged.
| Fleet | B Exams | A Exams | Units Visiting | Exam Hrs | Total Hrs | Heavy Mx Notes |
|---|
Component changes and heavy tasks in this period -- these require dedicated resource over and above routine exam teams.
Select a period above to see heavy maintenance breakdown.
Estimated total maintenance hours across each 4-weekly period of the year. Spikes indicate when heavier component work clusters.
* Annual profile assumes constant weekly mileage. Actual peaks will vary with timetable changes and component life accumulation.
Model the TOM phase by phase. Route tasks away from TCCs and see the network impact instantly.
Use the slider or type a number. Green = capable now. Amber = needs investment first.
Ask a question. Willow works out the answer.
Timetable change models the case named in the RDG Depot Performance Handbook: a mileage uplift turns, say, 52 B-exams/week into 80 at an already-stretched depot. Willow shows the exam-count change and the capacity gap per depot, then lets you plan the staffing response.
Save your current configuration as a named scenario. Adjust parameters in Resources or Depot Model, then save different versions to compare.
Infrastructure and competency matrix for TCCs. For outstation capabilities, use the Outstations tab. This data is cross-referenced by the Scenario engine -- when you ask "what if we move fleet X?" it checks whether the receiving depot has the right infrastructure and competency.
Physical facilities at this location
Which fleet classes can this depot maintain? Tick all that apply.
What level of maintenance can be performed here?
What does each fleet class need from a depot? This is shared across all TCCs.
Known restrictions at this location
Ask questions about your depot data, fleet, capacity, and resource planning
Define depot infrastructure, model overnight stabling requirements, and detect conflicts before they happen.
Visual representation of stabling roads. Green = available, amber = occupied, red = over capacity. Each block = 1 unit space.
Copy the road table straight out of your depot's rules of the plan — Word, Excel or PDF — and paste it below. Willow reads the columns, shows you what it understood, and lets you correct anything before it imports. Nothing changes until you press import.
Define each road's length, status and limitations. Length drives the Appendix B calculation above; the typed capacity is only a fallback for roads with no length recorded.
| Road | Capacity | Length (m) | Type | Stabling status | Reserved for | Use (day / night) | Fouling / m lost | Out of use | Restrictions | Allocated | Spare |
|---|
Units requiring stabling at this location overnight, derived from fleet allocation. Override manually if needed.
At-a-glance stabling status across all depots
Can the depot take the trains? Arrival timings, movement restrictions and driver resource decide how many units can actually come in tonight — and which of those is the one holding you back.
Match staffing reality against workload demand — planned and unplanned
Configure your shift groups and staffing levels. Staff counts should reflect actual available headcount per shift (after accounting for rest days in your rotation).
Planned demand from Willow's beat rate engine, plus an unplanned buffer derived from historical variance. Capacity is your net productive hours per shift after losses.
Enter or upload planned unit movements per shift window. Each arrival generates maintenance demand based on the task library.
Historical unplanned arrivals and reactive work. Adjust the variance % to stress-test against worse-than-average weeks.
Recommended headcount per shift to meet total demand (planned + unplanned). Compares against your current allocation.
Test different staffing levels side by side.
Side-by-side analysis of all Train Care Centers -- demand, capacity, staffing & fleet allocation
Red bars exceeding green = capacity deficit at that depot
| TCC | Units | Vehicles | Weekly Miles | Demand (hrs/wk) |
Capacity (hrs/wk) |
Gap (hrs/wk) |
Utilisation | FTE Surplus/Deficit |
Status |
|---|
Unit count per fleet class at each depot
Visual overview of where pressure points exist across the network
Set staffing and efficiency for each TCC to enable accurate comparison
Generates a draft Depot Rules document for the selected depot in the RDG Appendix A structure, populated from live Willow data. Sections Willow cannot know are marked [TO BE COMPLETED] rather than guessed. Opens as a printable HTML document.
Generate professional branded reports from live Willow data. Reports include executive summary, demand/capacity analysis, fleet breakdown, staffing configuration, and actionable recommendations.
One-page overview for senior leadership. Key metrics, status indicators, and top-line recommendations.
Comprehensive analysis with demand breakdown, resource configuration, per-fleet detail, and gap analysis.
Cross-TCC comparison showing demand, capacity, utilisation, and staffing across all six depots.
Depot stabling layout, road allocation, capacity conflicts, and overnight parking requirements.
Each operator is an isolated customer with its own data. Create one, then assign their people to it in "All users" below.
| Operator | Users | ID |
|---|
People waiting for you to approve them.
| Name | Signed up |
|---|
| Name | Status | Role | Operator |
|---|
| User | Event | When |
|---|
| User | Action | Detail | When |
|---|
Validates all inputs against expected ranges and flags anomalies
Architecture: Willow is a single-file HTML application with no backend dependencies. All calculation engines, operator data, and UI run in the browser. Data persists in localStorage -- each browser/device maintains its own state. No data leaves the user's machine unless API integrations are configured.
Operator Isolation: Every user signs in with their own email + password (secure cloud login). Each user belongs to exactly one operator, and the database itself enforces the wall (Row-Level Security): a user can only ever read or write their own operator's data — isolation does not depend on the interface.
Adding a New Operator (customer): No code needed. Open the Users & Logs tab: create the operator in the "Operators (customers)" panel, have their staff sign up on the login screen, approve them, and assign each person to the operator from the dropdown. They then log in and see only their own operator's data. The built-in operatorDB is sample/demo data only.
Calculation Engine: Willow calculates maintenance demand using beat rates: (weekly fleet miles ÷ exam interval) × exam hours = weekly demand hours per task per fleet. These cascade across all tasks and all fleets at each depot. Capacity is calculated from shift hours × staff × days, minus sickness/leave/training percentages, multiplied by an efficiency factor (default 50% -- accounting for non-wrench time). The gap is demand minus capacity.
Efficiency Factor: The 50% default means only half of available hours are productive wrench-turning time. The rest covers toolbox talks, walking to/from units, paperwork, break overlaps, handovers, material collection, etc. This is configurable per depot in the Resources tab. Most UK TOCs run between 45-55%. Note: This is an estimated figure -- to establish your actual efficiency, conduct timed activity sampling studies on shift. Use the Efficiency Study tool in the Analysis tab to record observations and calculate a measured baseline.
Scenarios: The scenario engine creates deep copies of baseline data and applies fleet moves, staffing changes, and efficiency adjustments without touching the live data. Users can compare Scenario A vs B vs Baseline side-by-side. Fleet cascades properly redistribute vehicles, mileage, and demand proportionally.
Fleet Allocations: Each fleet row has: TCC (depot), fleet class, unit count, vehicle count, target (units needed for the timetable), weekly miles, and B/A exam intervals. The beat rate calculation divides weekly miles by the interval to get exams per week, then multiplies by task hours. Users can add/remove fleet rows and cascade units between depots.
Task Library: Each fleet class has task hour definitions for 14 maintenance activities: B Exam, A Exam, A2, A3, Fuel Point, UFC, Tyre Turn, PHC, Wheelsets (non-driven), Wheelsets (driven), Full Unit Wheelsets, Engine/Raft, Transmission, and Pantograph. Each task has an hours value and a mileage interval. Tasks with zero interval are excluded from calculations. Fleet classes can share task data via aliases.
API Integrations: The Integrations tab is designed for future connections to IVU (live mileage data), Maximo (defect tracking and work orders), and TMS (staffing and absence data). These would replace static fleet mileage figures with real-time data, dramatically improving forecast accuracy. Each integration needs: API endpoint URL, authentication credentials, data mapping configuration, and refresh frequency.
Data Export: The Export tab generates reports from live calculation data. For commercial deployments, this should produce professional PDF reports with charts for COO/director-level presentations. Current implementation supports text/CSV export with PDF generation planned.
Password Management: TOC passwords are stored in localStorage and can be changed below. The admin password (your login) can also be changed. If a TOC user forgets their password, change it here and share the new one. Passwords are checked at login against both the default hardcoded values and any stored overrides.
Users currently logged in to Willow
As admin, you can switch between any TOC's data. This reloads all depots, fleet, and resource data for the selected operator.
Switch the AI assistant between Claude (Anthropic) and Azure OpenAI (Microsoft Copilot). Both use the same system prompt and Willow data context -- only the endpoint and model changes.
2024-02-01. Data sent to Azure stays within your tenant -- no data leaves to Anthropic when using this setting.
Add, edit, and configure outstations. Changes appear in the depot selector immediately.
Each operator has a unique login password. Change them here -- TOC users will need the new password next time they sign in.
These rates drive the Gap Cost calculation in the sticky bar and analysis. Set per-operator so each TOC sees their real cost exposure. Only visible here in admin -- TOC users see the £ figures but not the underlying rates.
Change the secret key required to access this admin panel. Current key is needed to access admin.
Every load / save / apply and every task toggle, with the value of A3 and the fleet count at that moment. Survives refresh. Use "Copy" and paste it to support.
Enable or disable features for all users
Connect Willow to live data sources. Configure endpoints and credentials below. Settings are saved locally per browser.