ERP
Contents
Every company eventually hits this moment: the warehouse manager quotes one number, accounting quotes another, sales quotes a third. None of them is lying — each one is just working from their own file.
That's exactly when the question "do we need an ERP" comes up.
This article answers it: what ERP is, what problems it solves, when you need one, when you don't, what it costs, and how implementation actually goes.
Short answer. ERP is a system that brings all of a company's operational processes into a single database. Warehouse, sales, production, cash flow, and payroll all run in one place, and data is entered once. It pays off starting around 30–50 employees. In Uzbekistan, launching on a ready-made system costs $5,000–9,000; building from scratch costs $45,000–90,000.
What ERP is, and what it isn't
ERP stands for Enterprise Resource Planning. The name is dated — a modern ERP doesn't just plan, it runs day-to-day operations.
In practical terms, ERP is one database with several modules on top of it.
A salesperson enters an order — the warehouse balance drops immediately. Goods ship — an invoice is generated. An employee clocks in — attendance is logged, and payroll is calculated from it at month end. Nobody enters the same thing twice.
What ERP is not:
- Not accounting software. Accounting handles official reporting; ERP handles operations. They connect through integration — one doesn't replace the other.
- Not a CRM. A CRM works with customers and deals. An ERP can include a sales module, but running the sales pipeline is a CRM's job.
- Not a magic wand. Feed a chaotic process into an ERP, and you get digitized chaos out.
What problems ERP actually solves
The table below covers the most common situations: current state on the left, after ERP on the right.
| Right now | With ERP |
|---|---|
| Stock levels are known only at the end of the day | Stock levels are visible in real time |
| One item gets sold to two customers | Stock is reserved the moment an order is placed |
| Payroll takes three days to calculate | Calculated automatically from attendance |
| Each department keeps its own spreadsheet | One database, each department has its own screen |
| No way to tell who changed what | Every action is logged |
| Pulling a report takes a full day | Reports generate in seconds |
| An employee leaves, and the data leaves with them | Data stays in the system |
The biggest shift is in the last two rows. The rest saves time; those two remove the business's dependence on specific people.
When you need ERP: five signs
- You have more than 30 employees and departments can't see each other's data
- The same data gets entered manually multiple times a day
- Preparing a monthly report takes several people several days
- Disputes over stock levels or debts come up multiple times a month
- You're opening a new branch or line of business and the current setup doesn't scale
Three of these — it's time to think about it. Four — you're already losing money.
Do the math. Say payroll takes three days a month to calculate. An accountant's salary is $500; three days is roughly $70. That's $840 a year. Add warehouse disputes, delayed decisions, and lost orders, and a 50-person company usually lands around $5,000–15,000 a year.
When you don't need ERP
Most articles skip this section, because it doesn't benefit ERP vendors. But it's worth knowing.
If you have fewer than 20 employees. People talk to each other directly at this size. A system becomes extra overhead for little payoff. A well-organized spreadsheet and one simple warehouse tool are enough.
If your processes haven't settled yet. When how you work changes every month, it's too early to lock it into an ERP. Stabilize the process first.
If you have exactly one problem. For example, if all you need is attendance tracking, a full ERP is overkill — a standalone tool solves it and costs ten times less.
If leadership isn't ready for change. ERP changes how work happens. If the requirement is "keep everything exactly as it is, just put it on a computer," the project will fail.
ERP modules: which ones you actually need
You don't need every module at once. Most companies start with two or three.
| Module | What it does | Who needs it first |
|---|---|---|
| Warehouse | Receiving, shipping, stock levels, inventory counts | Almost everyone |
| Sales & orders | Orders, invoices, customer database | Trade and distribution |
| Production | Planning, resources, cost accounting | Manufacturing companies |
| Cash & settlements | Cash flow, receivables/payables | Almost everyone |
| Payroll | Calculated from attendance | Companies with 50+ employees |
| Procurement | Suppliers, purchase orders, price history | Manufacturing and trade |
| Reporting | Analytics across every module | Almost everyone |
Practical tip: start with the two modules causing the most pain. Run them for three months, then add the next one. Rolling out seven modules at once is too big a change for staff, and resistance grows with it.
Ready-made ERP or build from scratch
This is the question we get most, and the answer depends on your processes.
Configuring a ready-made system
A good fit when your process is standard. The system already works and has been proven at other companies.
- 3–4 weeks
- $5,000–9,000
- Lower risk — the product is known to work
- Constraint: you adapt to the existing logic
Building from scratch
A good fit when your process is your competitive edge.
- 14–16 weeks
- $45,000–90,000
- Higher risk, but fully adaptable
- Everything is built around how you work
The most common mistake is doing it backwards. Companies build custom systems for standard warehouse accounting, then force a unique production process into a ready-made template. Both choices waste money.
The right question is: does this process set me apart from competitors? If not, buy ready-made. If it does, build from scratch.
What it costs and how long it takes
Real ranges from the Uzbekistan market:
$5,000-9,000On a ready-made system base, 3-4 weeks
$45,000-90,000From scratch, 6-10 modules, 14-16 weeks
What affects the price:
| Factor | Impact |
|---|---|
| Number of modules | Each extra module adds 1–3 weeks |
| Integrations | Each one adds 1–3 weeks |
| Number of branches | More branches mean separate configuration |
| Data migration volume | Depends on the old system's format |
| Staff training | Scales with the number of groups |
What's not included in the budget: server or hosting ($50–300/month), third-party licenses, SMS gateway. Ask about these separately when you get a proposal.
How implementation actually goes
A good project has six stages.
01
Discovery
2 weeks
02
Architecture
03
Sprints
04
Testing
05
Launch
06
Support
1. Discovery and scope — 2 weeks. The process is studied, a written document is produced, and an exact budget comes out of it. It's normal for this stage to be paid — it determines what system you actually need, and the document is yours to keep.
2. Architecture. Screens, data model, technology choices.
3. Sprints. A working version ships every two weeks. You open it and give feedback. This is the most important condition — you're not reading a report, you're opening the system.
4. Testing. Checklist-based review, load, and security.
5. Launch. Deployment, data migration, staff training.
6. Support. Warranty period, then an SLA contract after that.
Why ERP projects fail
In our experience there are three main causes, and none of them are technical.
Scope wasn't written down at the start. The project begins with "we'll figure it out as we go," then a new requirement gets added every week. Each requirement adds two weeks. Add ten of them, and you've lost twenty weeks and doubled the budget.
No internal champion. The system gets rolled out from the top, employees see a new screen one day, and they push back. The fix is simple: bring one or two people from each department into the project. They give feedback and later train their colleagues.
Migration gets left for last. The new system is ready, but five years of data is still sitting in the old program. Migration is separate, dedicated work, and it's expensive. Plan it from the start and do a trial migration in the first month.
Employees aren't afraid of a new system — they're afraid of visibility, because now it's clear who does what. Don't introduce the system as a monitoring tool. Show each employee specifically what it makes easier for them.
Questions to ask before you choose
When you get a proposal, ask:
- Do you have a client in my industry, and can I talk to them?
- Can you show a live, working system — not a demo, with real users?
- Is data migration included in the price, and how long does it take?
- Who owns the code and the database?
- Is there support after handover, and what does it cost?
- What happens if employees don't end up using the system?
The last two questions matter most, and they're usually the ones people forget to ask.
The bottom line
ERP isn't software — it's a decision about your processes. It pays off starting around 30–50 employees, and it only works once your processes are stable.
Three steps to get started:
- Identify the two processes eating the most time, and calculate what they cost you per month
- Decide whether your process is standard or unique — that answers the "ready-made or from scratch" question
- Start with the Discovery stage. Two weeks and a concrete document, after which price and timeline stop being guesses and become numbers
We'll assess your project in 30 minutes
We'll look at your process, tell you whether a ready-made system or a from-scratch build fits, and you'll walk away with a rough timeline and budget.
Discuss your project
Shahbozbek Usmonov
Founder & CEO of ShahNur Software. Writes about ERP, automation, and building software that ships.
About the company
We will assess your project in 30 minutes
Discuss your projectContents