Orus Studio
Manufacturing

Learning what a finished part actually cost

An auto-component manufacturer in Pune with about 180 employees

A manufacturer quoting from estimated costs, discovering after implementation that two of its highest-volume parts were being sold below cost.

Client
An auto-component manufacturer in Pune with about 180 employees
Industry
Manufacturing
Duration
16 weeks
Team
4 engineers
Year
2024

The situation

Material consumption was recorded at the end of the month against total production, so per-part cost was an average rather than a measurement.

Significant work went to outside job workers with no systematic tracking of what was sent, returned or lost.

Rejection and rework were recorded on paper at each station and rarely reached costing at all.

Quoting relied on an estimated cost that had not been validated against actual consumption in years.

What we did

01

Capturing consumption at issue, not at month end

Material issues post against a specific work order at the store counter. Actual versus standard consumption then becomes a variance per work order rather than a monthly aggregate that hides which part is responsible.

02

Shop-floor terminals with a deliberately minimal screen

Operators scan the job card and enter quantity produced and rejected. That is the entire interaction. Every additional field we considered was rejected, because anything taking more than a few seconds per job would have been worked around within a week.

03

Job work tracked as material in transit

Material sent outside is tracked against a challan with an expected return quantity and date, reconciled on receipt. Losses at job workers became visible and quantified for the first time, and the GST obligations under Rule 45 are handled as a consequence of the same records.

04

Rolling actual cost per part

Material, labour and machine overhead roll into a per-unit actual cost, compared against standard. Two high-volume parts turned out to be below cost, which changed the company's quoting immediately.

The parts that were actually hard

Problem

The shop floor had metal dust, vibration and operators wearing gloves. Standard tablets did not survive.

How we handled it

We moved to fixed industrial panel PCs with physical scanners at each station, and designed the interface for large targets usable with gloves on. The first rollout with consumer tablets failed within a fortnight, which we should have anticipated.

Problem

Multi-level BOM changes were being made without version control, so historical costing became meaningless whenever a design changed.

How we handled it

BOMs were versioned with effectivity dates, so a work order references the BOM version current when it was released. Historical costs stay correct after a design change, which is the whole point of measuring them.

What shipped

  • Multi-level versioned BOM with effectivity dating
  • Work order release, scheduling and operation-level progress tracking
  • Shop-floor capture terminals at each production station
  • Job-work tracking with challans, reconciliation and ITC-04 data
  • Actual versus standard costing with variance analysis per part

Outcomes

  • Two high-volume parts were found to be selling below actual cost and were repriced
  • Job-work losses became measurable, and one processor was discontinued as a result
  • Quoting moved from estimated cost to measured cost, which management described as the main value of the project

Outcomes are described qualitatively where no clean measured baseline existed before the work started.

Built with

Next.jsPostgreSQLPrismaRedisDocker

Have a problem shaped like this one?

Tell us what you are dealing with. We will come back with a scoped estimate and an honest view on whether we are the right fit.

Explore