The pricing-as-a-person problem
In most Singapore SMB manufacturers, the real pricing engine is a senior salesperson. She knows which customers get which rate, which orders get a margin override, which products carry hidden assembly costs, and which freight terms require approval. That knowledge is not written down anywhere a computer could read. It lives in her head and a few master spreadsheets she opens twice a day.
For a long time this works. The business grows because she is competent and the customer relationships are real. Then volume doubles, two junior sales hires join, and the team realises that every custom-spec quote still has to route through her. Quotes take days. Deals are lost to faster competitors. She works late.
What a custom pricing engine actually does
A pricing engine is a structured representation of your pricing logic that any authorised user — or the customer, in a self-service portal — can use without asking the senior salesperson. It captures: customer-specific rate cards, product variants and forms, dimension-based scaling, MOQ tiers, freight rules, approval thresholds, and exception logic. It produces a price for any valid combination, instantly.
The hardest part of building one is not the code. It is forcing the business to make implicit pricing rules explicit. The senior salesperson knows that this customer gets a 5% bump if the order goes above 50,000 units, but only if delivery is to their warehouse, and not in December. Until that rule is written down, it cannot live in software.
A real case
We rebuilt the platform with the pricing logic baked into the system itself. Each customer sees their own negotiated rates the moment they log in. Configurations recalculate live as they adjust width, length, and quantity. The admin team gained a panel where non-technical staff could onboard customers and adjust rates without developer tickets.
The numbers that changed:
- Quotation cycle: 2 days → 2 minutes. Sales productivity up ~85%.
- Mobile order share: 12% → 45% of total volume within three months.
- Order errors: −75%. Pre-sales support calls: −60%.
- Average order value: +40% — customers self-optimized to better pricing tiers.
- Customer onboarding time: 3–5 business days → under 5 minutes.
- Scaled from 50 to 200+ customers without adding admin headcount.
The deeper shift was qualitative. The senior salesperson stopped being a calculation engine and became the strategist she was hired to be. Junior sales could quote confidently within seconds.
What good pricing engines have in common
- Customer-scoped views — logged-in users see only their pricing, no exposure of competitor rates.
- Form + variant modelling — the unit of pricing is not a SKU; it is a configured order line.
- Live recalculation — every spec change updates the price instantly without a round-trip.
- Admin self-service — sales managers onboard customers and tweak rates in a UI, never a database.
- Bulk import + export — Excel is still the team's muscle memory; pricing data should move both ways.
- Approval thresholds — orders below a margin floor or above a discount ceiling route to an approver automatically.
- Audit trail — every rate change has an author, a timestamp, and a reason.
What to do before you build one
The discovery is more about the business than the software. Sit with the senior salesperson and write out, in plain language, every rule that affects a price. Walk through ten recent custom quotes and ask what made each one what it was. Half the rules will be obvious; half will be implicit and emerge only when you stress-test edge cases.
The output is a rule sheet. That sheet is what becomes the pricing engine. If the sheet is hard to write, that is the sound of business risk you have been carrying without seeing.