Logistics & Supply Chain

₹1.4 crore of margin nobody could see

A third-party logistics operator found out which of its lanes actually made money, two weeks after starting to look.

This is an illustrative composite. It is drawn from delivery patterns across our work rather than published as a named client reference, and the figures describe the pattern rather than one audited engagement. We would rather label it than imply a reference we cannot put you in touch with.

Size
210 staff
Platforms
ForgeAtlas

Context

A 210-staff 3PL running warehousing and distribution for eighteen clients across eleven lanes. Revenue was growing and margin was not, and the finance team's lane-level view was a quarterly spreadsheet built by hand from four sources.

The problem

Nobody could answer 'which lanes make money' inside a quarter. Costs sat in the TMS, the warehouse system, fuel cards and a payroll export, each with a different notion of a job. The quarterly reconciliation took three weeks and was stale by the time it landed, so pricing decisions were made on last quarter's guess.

What we did

  1. 01Deliberately scoped as a two-week Opportunity Map, not a build. The question was whether the data supported lane-level margin at all.
  2. 02Reconciled four cost sources onto one job identity, which is where most of the fortnight went and where the answer actually came from.
  3. 03Produced a lane-level margin picture for the trailing four quarters and had finance challenge every number in it.
  4. 04Quantified the leakage and named its three causes, with the evidence for each attached.
  5. 05Delivered a costed build plan for the permanent version — and an explicit statement of which two lanes did not need one.

Architecture

Nothing here is a black box. You can read the architecture before you sign, and you own it after.

  1. 01IdentityOne job identity reconciling TMS, warehouse, fuel card and payroll records — the hard part, and the whole answer.
  2. 02Cost modelAllocated cost per job with the allocation rule stated per line, so finance can dispute a rule rather than a total.
  3. 03Margin viewLane-level margin across four trailing quarters, rebuilt from source rather than from the old spreadsheet.
  4. 04EvidenceEvery leakage finding carries the rows behind it. A number finance cannot trace is a number finance will not act on.

Results

The numbers the work closed against.

annual margin leakage a lane-level model is designed to surface
₹1.4 Crannual margin leakage a lane-level model is designed to surface

Basis: Modelled across 11 lanes from four reconciled cost sources over four trailing quarters. It is what the arithmetic identifies as unallocated, and every finding carries the rows behind it for finance to dispute.

design target from kickoff to a costed build plan
14 daysdesign target from kickoff to a costed build plan

Basis: The scoped length of an Opportunity Map. Most of it is reconciling four cost sources onto one job identity, which is where the answer comes from.

the leakage is expected to decompose into, each with its evidence
3 causesthe leakage is expected to decompose into, each with its evidence

Basis: A structural output of the method rather than a forecast: findings are grouped by cause and each carries its source rows.

the engagement is designed to rule out as needing further work
2 lanesthe engagement is designed to rule out as needing further work

Basis: Stated because ruling work out is an output we commit to. A map that only ever finds more to build is not a map.

Stack

  • Python
  • Supabase Postgres
  • dbt-style transforms
  • Next.js
  • Vercel

Does this look like your business?

Tell us what is slow and what it is costing. We will point you at the closest precedent we have and say plainly where yours would differ.

Scope it with us