DiscvrAI
Manufacturing & FMCG

Your Should-Cost Isn't Last Year's Benchmark

Live should-cost on DiscvrAI: a current number from your consumption and cost drivers, not last year's spreadsheet. A human still commits the price.

Shubham Srivastava · 1 September 2026 · 5 min read

Ask a category manager what number they took into the last steel or polymer conversation. More often than not it is last year's clean-sheet, a cost engineer's file, or the supplier's quote. None of those is a should-cost. One of them is a historical document. One of them is the other party's opening bid.

The index has moved. The lag you modelled last April is not the lag you have now. The tail of the item master never got a number at all. So the room is anchored on whatever was easiest to print.

A should-cost is a current argument

Material, conversion, freight, duty, a margin band. Each line with a source and an as-of date. Mapped to the driver that actually moves that item, with a pass-through you can defend. That is the same work a good cost engineer already does. The failure is cadence and coverage, not skill.

If the number cannot be opened, it is not a should-cost. It is a guess with formatting.

What changes when it is live

  • The material line re-prices when the driving index moves, within a modelled lag, instead of waiting for next year's exercise.
  • Gaps versus contract are ranked by rupee impact, so the next conversation is the one that is worth having.
  • The same number travels into the RFx as a landed-cost brief, then into AP as the number invoice-match can enforce.

Frequently asked questions

Start with one outcome. Scale from there.

Most engagements begin as a single product on a single workflow, with a measurable result inside 8-12 weeks.