Batteries and fuel cells
Findings

Battery-pack manufacturer scaling from pilot to series · 1 plant in Hauts-de-France · 140 employeesThis War Map shows how we turn an operational review into an implementation plan. The dates show a possible sequence; the scope and schedule of a real engagement are agreed with you.

Findings

Where work gets held up, who has to step in, and what it costs in this example.

Changes reach the line late

About 14 engineering changes a month. Each takes 23 days on average to become effective, and 3 last quarter were built on a superseded BOM: 46 packs reworked.

Shortages set the plan

The shortage list is rebuilt by hand every morning. 31 hours of line stops for missing parts in three months, most on parts with a confirmed date that slipped.

Traceability gap on line 3

Rebuilding the cell history of a line-3 pack takes 2 days. From February 2027, industrial batteries over 2 kWh need a battery passport.

Documents behind the ramp-up

9 open PPAP packages, about 35 class 9 shipments a week and a grant report every six months all rely on documents assembled by hand.

Illustrative time allocation for 11 interview roles

Chasing parts, data and approvals
26%
Engineering and quality work
24%
Re-entering data between systems
19%
Firefighting on the line
17%
Customer and funder reporting
14%

These scenario percentages total 100%. They describe the 11 interview roles only, without extrapolating to all 140 employees.

45% of the time goes to chasing parts, data and approvals and to re-entering data between systems. Firefighting on the line takes 17%, reporting 14%. 24% remains for engineering and quality work.

Key Insight: Some firefighting comes with any ramp-up. The 19% spent re-entering data is the clearest target, because the same BOM, serial number or test result is typed into two or three systems.

SOP vs. Reality

Engineering Change to ProductionSOP: 4 steps → The Reality: 11 steps, 5 handoffs
The SOP
Change approved in Jira
BOM revised
Line updated
Change effective
4 steps, linear
The Reality
11
Steps
5
Handoffs
23 days on average, 10 planned
Cycle time
~14 changes/month
Volume

Engineering decides quickly. The change then has to be copied into four places by different people, with no shared list of what is done. With 14 changes a month, one missed copy is enough to build a batch on the wrong revision.

The Gap
No link between PDM, Business Central and the MES. Each BOM revision is typed again by hand.
Nobody owns the full list of tasks after approval. Progress lives in Jira comments and emails.
3 changes last quarter were built on a superseded BOM, leading to 46 reworked packs.
Some changes affect a customer's PPAP or the UN38.3 test summary. That check depends on someone remembering to ask.
Shortage ManagementSOP: 5 steps → The Reality: 9 steps, 4 handoffs
The SOP
MRP run in Business Central
Shortages listed
Suppliers expedited
Plan adjusted
Line informed
5 steps, linear
The Reality
9
Steps
4
Handoffs
~2.5 h of preparation every morning
Cycle time
~31 part numbers on the list, 85 suppliers
Volume

The MRP knows what is missing. It does not know what suppliers promised yesterday, which lives in buyers' inboxes. So the list is rebuilt by hand every morning before the meeting can start deciding.

The Gap
Supplier confirmations sit in buyers' inboxes. Business Central shows order dates, without the latest promise.
64% of line-stop hours involved a part with a confirmed date that slipped.
The list is rebuilt from scratch each morning, and yesterday's decisions have to be remembered.
Customer delivery promises change after the meeting and are passed on by hand.
Customer Traceability Request (Line 3)SOP: 3 steps → The Reality: 9 steps, 3 handoffs
The SOP
Request received
Genealogy extracted from MES
Report sent
3 steps, linear
The Reality
9
Steps
3
Handoffs
2 days elapsed, 3 person-days
Cycle time
~3 requests/month
Volume

On lines 1 and 2, the same request is a query. On line 3, it is an investigation across scan files, bench PCs and supplier certificates. When a scan is missing, the answer to the customer comes with caveats.

The Gap
Line 3 started production before the MES could cover it. Scans go to CSV files on a station PC.
Bench results and scans share no key, so matching is done by serial number and time stamp by hand.
About 1 pack in 40 on line 3 has at least one missing scan.
The same data will be needed for every pack once the battery passport applies.
PPAP PackageSOP: 6 steps → The Reality: 10 steps, 4 handoffs
The SOP
Customer requirements received
Documents prepared
Dimensional and material results
Test reports
Part Submission Warrant signed
Submitted
6 steps, linear
The Reality
10
Steps
4
Handoffs
~10 weeks per package
Cycle time
9 open packages, 4 customers
Volume

The engineering content is solid. The time goes into finding the current version of each element, chasing suppliers for certificates and reformatting for each customer portal. A single outdated element sends the package back.

The Gap
No list of which controlled document version belongs to which package.
Supplier certificates are chased one by one by email.
2 of the last 7 submissions came back for a missing or outdated element, each adding about two weeks.
Each customer uses a different template and naming rule for the same 18 elements.
Key Insight: The four sample comparisons show 175%, 80%, 200% and 66.7% more steps, respectively: (actual ÷ documented − 1) × 100. Some extra steps come with any ramp-up; the costly ones are retyping between systems and searching for the latest version.

Work that needs attention

Engineering + ProductionChange implementation: 14 changes/month × 6 h retyping BOMs, instructions and routings = 84 h/month; 84 × 12 ÷ 52 = 19.4 h/week19.4h
Supply ChainShortage list rebuilt by hand each morning: 1.5 h/day for Sarah plus 1 h/day across two buyers12.5h
QualityTraceability reports for claims and audits: 3/month × 21 h = 63 h/month; 63 × 12 ÷ 52 = 14.5 h/week14.5h
ProductionLine-3 test results matched to serial numbers by hand: ~40 packs/week × 6 min4h
Quality + ProgrammesPPAP packages: 9 open × 1.5 h/week chasing and formatting documents; excludes engineering work13.5h
FinanceGrant milestone reports: 2/year × 110 h assembling timesheets, invoices and evidence; 220 ÷ 52 = 4.2 h/week4.2h
LogisticsClass 9 shipping documents and UN38.3 test summaries: ~35 shipments/week × 10 min5.8h
AllBattery passport: line-3 packs cannot yet show the complete cell-to-pack data required from February 2027ongoing
73.9 person-hours/week across 7 quantified rows; 1 unquantified risk73.9 h
Key Insight: Change, traceability and grant rows are converted from monthly or yearly hours to weekly hours. The traceability row counts quality and test time per report; the line-3 matching row covers routine production only. Rows are scoped to avoid double counting, and engineering and test work itself is excluded.

Risk Register

Regulatory ReadinessCritical
Battery passport from February 2027
The EU Battery Regulation requires a digital passport for industrial batteries over 2 kWh. Line 3 cannot yet produce complete cell-to-pack data.
QualityCritical
Builds on superseded BOMs
3 changes last quarter reached part of the line late, and 46 packs were reworked. At double the volume, the same gap doubles the exposure.
Single Point of FailureHigh
Sarah Kone, shortage list
Only Sarah knows how to turn the MRP output into a list the meeting trusts. When she is away, the meeting works from the previous day's list.
Transport SafetyHigh
UN38.3 summary after a change
Some design changes require the UN38.3 test summary to be re-evaluated. The check depends on Hugo being asked; nothing in Jira triggers it.
FundingHigh
Grant evidence assembled late
Milestone payments depend on complete reports with evidence by work package. Timesheets are collected after the fact, which makes funder audits harder.
System FragmentationHigh
PDM, Business Central and MES unlinked
The same BOM exists in three systems kept by three people. During ramp-up, they drift apart within days.
Customer ExposureMedium
PPAP rejections
2 of the last 7 PPAP submissions came back for a missing or outdated element, each adding about two weeks before series approval.
Key-Person DependencyMedium
Hugo Lambert, test data
Wrote the nightly script that links bench results to serial numbers on lines 1 and 2. It is undocumented.
Next Chapter
Implementation plan

Strategy Session · 30 min

A real engagement is scoped in the Strategy Session and may begin with a War Map or a bounded Feasibility Sprint.

Request a Strategy Session