Ask five software houses what an ERP costs and you will get five refusals to answer, followed by a request for a meeting. That is not always evasion — the honest range genuinely is wide — but it leaves you unable to budget, which is unhelpful when you are trying to decide whether to start at all.
So here are real numbers, what sits behind them, and what is usually missing from the quotation you will be shown.
The ranges
| What you are buying | Typical range (PKR) | Time to live |
|---|---|---|
| POS / basic inventory, single location | 120,000 – 300,000 | 2–4 weeks |
| Inventory + accounts, multi-user | 300,000 – 700,000 | 4–8 weeks |
| Multi-module ERP with production | 600,000 – 2,500,000 | 2–4 months |
| Multi-unit / multi-location ERP | 1,500,000 + | 4–8 months |
| HRM, payroll & biometric attendance | 200,000 – 600,000 | 3–6 weeks |
| FBR e-invoicing integration | 150,000 – 400,000 | 1–4 weeks |
| AI chatbot or calling agent (setup) | 175,000 – 500,000 | 2–6 weeks |
| One custom report | 15,000 – 60,000 | 1–5 days |
These are development costs for genuinely custom work, excluding hardware, third-party licences and tax. If somebody quotes you far below the bottom of a range, that is worth understanding rather than celebrating — see the section on cheap quotes further down.
The five things that move the number
1. How many modules — and how much they talk to each other
Cost does not scale linearly with module count. Two independent modules is roughly twice one. But production feeding costing feeding accounts is not three times one, because the connections between them carry rules, edge cases and reconciliation. Integration is where the effort lives.
2. How unusual your process is
Standard buy-and-sell is well-trodden. A dyeing floor tracking lots across shades, or a garment unit tracking an order through cutting, stitching and finishing with rejections at every stage, is not. Unusual is not bad — it is often exactly why off-the-shelf failed you — but it costs more to build and more to test.
3. The state of your existing data
This is the most consistently underestimated line in every project we have done. Clean Excel with one row per item imports in an afternoon. Fifteen years of records where the same supplier appears under four spellings, half the items have no unit, and opening balances do not tie to the ledger — that is weeks, and it must happen whoever builds the system.
4. Users and locations
Five users at one location is a small deployment. Fifty users across three units means rights matrices, network and server work, more training and much more testing, because there are more ways to do things wrong.
5. Integrations
Each external system — FBR, a bank file format, a website, a biometric device, a weighbridge — adds its own scope. Each has its own failure modes to handle, and each needs a plan for the day it changes without warning.
What is usually missing from the quotation
- Data migration and cleaning. Often quoted as a line item at a token price, then discovered to be the biggest task. Ask specifically what happens to your history.
- Training. Not a demo — sessions with your actual staff on your actual data, and again for the people who join later.
- Server and backups. Somebody has to set up the machine, secure it, schedule the backups and verify they restore. That is real work.
- Support after the warranty. Ask the monthly figure before signing, not after.
- Change requests. Your process will evolve. Ask the hourly or per-change rate up front.
- The parallel-run month. Most businesses run old and new side by side for a few weeks. That costs your team time, and it should be planned rather than discovered.
Why the cheap quote is usually expensive
Somebody will offer to build your ERP for PKR 150,000. That number is real, and here is what it typically buys:
- A student or a very junior developer, working alone
- No data migration — you start from zero balances
- No handling of the awkward cases: returns, adjustments, corrections, closing
- No documentation, so nobody else can ever maintain it
- A disappearance when the next project comes along
We have rescued enough of these to say it plainly: the second attempt costs more than doing it once properly, because you pay to understand what was built before you can replace it. If budget is genuinely tight, buy a smaller scope from someone serious rather than a large scope from someone cheap.
How to control what you spend
- Buy in stages tied to delivery. Pay for a module when you can use it. It aligns everyone’s interest with going live rather than going long.
- Start with what hurts most. Whichever problem costs you money or sleep today. Momentum from one working module carries the rest of the project.
- Get scope in writing, including what is excluded. The exclusions list prevents more arguments than the inclusions list.
- Clean your data before you start. This is free if you do it, expensive if you pay someone else to, and it is the single biggest lever on timeline.
- Name one decision-maker. Projects do not stall on code. They stall waiting for someone to decide how discounts should work.
A worked example
A textile unit with 120 staff, one factory, wanting production tracking, store, and accounts, with FBR compliance and payroll:
| Stage | Scope | PKR |
|---|---|---|
| 1 | Store, purchase, issue, stock ledgers | 250,000 |
| 2 | Production tracking and daily output | 350,000 |
| 3 | Accounts: vouchers, ledgers, aging, trial balance | 300,000 |
| 4 | HRM, payroll and biometric attendance | 250,000 |
| 5 | FBR e-invoicing integration | 180,000 |
| — | Data migration, server setup, training | 150,000 |
| Total, spread over ~5 months | 1,480,000 |
Roughly the cost of two mid-level salaries for a year — against a system that removes considerably more than two people’s worth of manual work, and keeps doing so.
Want a real number for your business? Tell us the modules you need and we will send a written fixed quote, usually within two working days. The scoping call costs nothing. See our published pricing or send us the details.