Consumer goods
Findings

Maker of natural personal-care and household products · 1 plant in the Loire valley · 190 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.

Launches slowed by paperwork

A new product takes 11 weeks from approved formula to first listing. 4.5 of those weeks are data rejects and extra artwork rounds.

Promotion money paid without a check

Retailers claim €2.4M of promotion funding a year. Only claims above €2,000, 38% of the value, are checked against what was agreed.

Peak season handled by hand

In December, ticket volume more than triples. First reply reached 3.4 days, and almost half the tickets only asked where an order was.

One reviewer for every pack

Every claim, ingredient list and hazard pictogram passes through one regulatory manager. Her archive is the only record of past decisions.

Illustrative time allocation for 11 interview roles

Copying data between systems
24%
Review and approval loops
21%
Chasing retailers, printers and agencies
18%
Planned brand, sales and supply work
25%
Decisions
12%

These scenario percentages total 100%. They describe the 11 interview roles only and are not extrapolated to the 190 employees.

Copying data, review loops and chasing take 63% of the interviewed roles' time. Planned work takes 25% and decisions 12%.

Key Insight: Part of the review time is protection: regulatory and brand checks must stay. The target is fewer proof rounds and less copying, with the same people signing off.

SOP vs. Reality

New product listingSOP: 5 steps → The Reality: 12 steps, 5 handoffs
The SOP
Product data created in PIM
Data published to data pool
Retailer accepts listing
Marketplace listing created
Product sellable
5 steps, linear
The Reality
12
Steps
5
Handoffs
11 weeks, formula to listing
Cycle time
~60 launches and renovations/yr
Volume

The same product data is typed into Divalto, the PIM, the data pool and Amazon. Retailers check weights, dimensions and ingredient lists, and each rejection adds a round trip of about a week.

The Gap
The PIM to data pool connection has failed since a version upgrade in March. Manon keys every change again by hand.
17% of SKUs carry at least one retailer data warning, mostly weights, dimensions and ingredient lists that differ between systems.
Amazon listings are built by hand from a PIM export. Nine were suppressed last quarter for a missing attribute, each for 2 to 6 days.
Scenario: data rejects add about 2 weeks to a launch and extra artwork rounds about 2.5 weeks. Together, 4.5 weeks of an 11-week path.
Artwork changeSOP: 5 steps → The Reality: 11 steps, 6 reviewers handoffs
The SOP
Change brief
Designer update
Internal approval
Printer proof
Print release
5 steps, linear
The Reality
11
Steps
6 reviewers
Handoffs
23 days on average
Cycle time
~22 changes/month
Volume

Every pack change, whether regulatory, promotional or requested by a retailer, goes through an Excel log, an agency, six reviewers by email and a printer proof checked by eye. Average 2.6 proof rounds.

The Gap
Six reviewers comment on separate copies of the PDF. Fabrice merges the comments and settles conflicts by phone.
Claims, recycling information and CLP pictograms are checked by Bérénice from her own archive of past decisions.
Printer proofs are compared with the approved PDF by eye. One wrong barcode reached print in 2025; 18,000 sleeves were scrapped.
Pack changes are planned without the remaining stock of old packaging. Run-out is negotiated case by case with planning.
Promotion funding claimsSOP: 4 steps → The Reality: 9 steps, 3 handoffs
The SOP
Promotion agreed
Claim received
Claim checked against agreement
Paid or disputed
4 steps, linear
The Reality
9
Steps
3
Handoffs
45-90 days to close a dispute
Cycle time
~1,400 claims/yr
Volume

Promotion terms are agreed in the annual negotiation and in mid-year deals, then kept in Romain's Excel file. Retailers invoice or offset promotion funding weeks later. Claims under €2,000 are approved without a check.

The Gap
About 1,400 claims a year for €2.4M. Claims above €2,000 represent 38% of the value and are the only ones checked.
In a sample of 210 claims, 6.8% of the claimed value had no matching agreement or exceeded the agreed terms: duplicates, wrong rates, promotions that never ran.
Scenario: 6.8% of €2.4M is about €163K a year. Every dispute stays a decision of the finance controller and the key account manager.
Offsets against payments carry only a short reference, so matching them to a promotion is guesswork.
Forecast to productionSOP: 4 steps → The Reality: 8 steps, 6 sources handoffs
The SOP
Monthly forecast
S&OP meeting
Production plan
Components ordered
4 steps, linear
The Reality
8
Steps
6 sources
Handoffs
Monthly, reworked weekly
Cycle time
640 SKUs
Volume

Laurent rebuilds the forecast every week from four retailer sell-out files, Shopify and Amazon exports and promotion emails. The monthly S&OP meeting spends most of its time agreeing on the numbers.

The Gap
Forecast accuracy at SKU and month level is 61%. Late promotion confirmations explain a large share of the misses.
14% of the 2025 promotions had a shortage on at least one SKU.
Packaging components take 6 to 10 weeks, longer than the window in which promotions are confirmed.
The forecast file has one owner. During his August leave, the plan was rolled forward unchanged.
Key Insight: The four sample comparisons show 140%, 120%, 125% and 100% more steps, respectively: (actual ÷ documented − 1) × 100. Some extra steps are regulatory checks that must stay; most are copies of the same data between systems.

Work that needs attention

Product dataProduct data keyed again in the data pool and Amazon after the PIM sync failed: ~60 SKU updates/week × 9 min9h
PackagingArtwork proofs sent to 6 reviewers and comments merged by hand: 22 changes/month × 3 h; 22 × 3 × 12 ÷ 52 = 15.2h/week15.2h
FinancePromotion funding claims matched to agreements by hand, above €2,000 only8h
Supply chainForecast rebuilt weekly in Excel from 4 sell-out files and 2 e-commerce exports7h
Customer serviceOrder status tickets answered by hand: ~6,800/yr × 5 min; 6,800 × 5 ÷ 60 ÷ 52 = 10.9h/week, far higher in peak weeks10.9h
E-commerceAmazon refunds checked against items received: ~120 cases/month × 12 min; 24h × 12 ÷ 52 = 5.5h/week5.5h
RegulatoryClaims and CLP pictograms checked by a single reviewer from her own archiveongoing
55.6 person-hours/week across 6 quantified rows; 1 unquantified risk55.6 h
Key Insight: Monthly and yearly volumes are converted to weekly hours (× 12 ÷ 52, or ÷ 52). The order status row is a yearly average: December weeks run at three to four times that load. Rows are scoped to avoid double counting; time spent is not automatically recoverable.

Risk Register

Single point of failureCritical
Bérénice Carré, regulatory
Sole reviewer for cosmetic claims, ingredient lists and CLP labels, working from her own archive. When she is away, pack releases stop.
Single point of failureHigh
Laurent Chevalier, forecast
The forecast file has one owner. During his August leave, the plan was rolled forward unchanged.
MarginHigh
Promotion claims paid unchecked
€2.4M a year of funding claims; 62% of the value is paid without a check against agreements.
PackagingHigh
Print errors reach production
Printer proofs are checked by eye. A wrong barcode in 2025 led to 18,000 scrapped sleeves.
Retail relationshipHigh
Data warnings at retailers
17% of SKUs carry a data warning. Retailers can delay a listing or hold replenishment orders on wrong dimensions.
Customer experienceMedium
Peak backlog
First reply reached 3.4 days in December 2025. Marketplace ratings fell during the same weeks.
Data integrityMedium
Four versions of each product
Divalto, the PIM, the data pool and Amazon disagree on weights or dimensions for about 1 SKU in 6.
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