5 ERP implementation mistakes and how to avoid them

In this article14 sections

1. Why ERP implementation mistakes happen, and why it is rarely the software
2. Mistake 1: Buying before you know the problem you are solving
3. Mistake 2: Moving dirty data into a clean new system
4. Mistake 3: No owner, no champion, no decision maker
5. Mistake 4: Skipping training and ignoring how people feel
6. Mistake 5: Switching everything on in one big bang
7. Five more ERP implementation mistakes worth avoiding
8. How much does ERP implementation cost in Nigeria?
9. Your ERP implementation checklist for the first 90 days
10. How to choose an implementation partner you can trust
11. How to know your ERP is working after 90 days
12. Key takeaways
13. Frequently asked questions
14. Your next step: avoid these ERP mistakes from day one

Last year a distributor in Kano called me on a Sunday evening. He had paid ₦9 million for an ERP system eight months earlier. His storekeeper still wrote every delivery in a notebook, his accountant still worked in Excel, and the new system showed a stock figure that was wrong by almost a third. “Is the software bad?” he asked. It was not. The software was fine. He had fallen into some of the most common ERP implementation mistakes, and he had actually made three of them at once.

ERP stands for Enterprise Resource Planning. In plain words, it is one software system that holds your sales, stock, purchases, money and staff records in one place, so that everyone works from the same numbers. When it works, an owner stops chasing people for figures. When it fails, you have paid a lot of money for a very expensive notebook that nobody opens.

I have spent eight years building business software in Nigeria, including ERP, inventory and POS systems, and I am often called in to rescue projects that someone else started. In this guide I will walk through the five mistakes I see most often, explain why each one happens so easily in Nigeria, and show you what to do instead. You will also get cost ranges in naira, a 90 day checklist, and the exact questions I would ask any vendor before signing.

Why ERP implementation mistakes happen, and why it is rarely the software

When an ERP project goes wrong, the first reaction is to blame the product. “This software is not made for Nigeria.” Sometimes that is true, especially when a system cannot work offline. But in most rescue jobs I have done, the software could do everything the business needed. What failed was the way the business prepared for it, the way data went in, or the way people were brought along.

What failure really looks like

Failure is rarely dramatic. Nobody announces that the project is dead. It looks like this instead: cashiers sell in the system but the storekeeper still uses a book. The owner logs in during the first two weeks, then stops. Reports are printed but nobody trusts them, so the accountant rebuilds the same figures in Excel “to be safe”. Six months later the business is paying a monthly fee for a system that is used as an expensive receipt printer.

The five mistakes at a glance

Here is the short version of what this article covers. The table shows each mistake, how it shows up in a real Nigerian business, and the cheapest way to fix it.

Why this matters in Nigeria

Foreign ERP guides assume stable power, cheap fibre internet, staff who stay for ten years and vendors who answer emails within an hour. In Nigeria, NEPA takes light in the middle of stock taking, the network drops during a busy Saturday, a good storekeeper gets a japa visa, and the dollar price of your software plan changes overnight. Each of these turns a small mistake into an expensive one. That is why a Nigerian ERP project needs more planning, not less, than one in a country with steady infrastructure.

Mistake 1: Buying before you know the problem you are solving

This is the most expensive of all the ERP implementation mistakes because everything else grows out of it. A business owner sees a demo, loves the colourful dashboard, pays a deposit, and only then asks, “What exactly do I want this to fix?” By then the vendor has already decided the scope, and the scope is whatever their product happens to do.

The demo trap

Every vendor demo looks wonderful, because the demo data is perfect. Products have clean names, stock is always correct, and the internet never drops. Your business is not like that. Your Ibadan branch sells on credit to regular customers, your supplier in Lagos delivers half the order and promises the rest next week, and your accountant wants VAT shown a particular way. If none of that was tested in the demo, you have not seen the system yet. You have seen a film about it.

Write a one page problem statement

Before you speak to any vendor, write one page in plain language. Mine usually has four parts:

1. The pain in numbers. “We lose about ₦800,000 a month in unexplained stock differences.” “It takes my accountant six days to close the month.”
2. The processes involved. Selling, buying, stock transfers, credit sales, cash handling, payroll. Tick only those that truly hurt.
3. The people who will use it. How many cashiers, storekeepers, managers and accountants, and in which locations.
4. What success looks like in 90 days. “Every sale recorded with payment type. Stock count within 2 percent of the system. Month end closed in two days.”

This page becomes the yardstick for everything after. When a vendor says “we can also add a CRM and an HR module”, you can calmly ask whether it helps you reach the success line on your page.

Ask vendors for outcomes, not feature lists

A feature list tells you what the software can do in theory. An outcome tells you what will change in your shop. Instead of asking “Do you have stock transfer?”, ask “How will I know within one hour if 10 cartons sent from Wuse never arrive in Kubwa?” A vendor who has solved that problem before will answer with a story and a screen. A vendor who has not will start reading from a brochure. That difference is worth more than any discount they can offer.

Must have, nice to have, later

Split every feature request into three buckets. Must have means the business cannot run without it in the first phase, for example selling, stock per location and daily cash closing. Nice to have means it helps but can wait, for example loyalty points or a supplier portal. Later means it sounds exciting but has nothing to do with today’s pain. Most of the projects I rescue tried to deliver all three buckets in one go. I have written more about this way of thinking in my article on digital transformation for Nigerian SMEs, where the same rule applies: start with the problem, not the tool.

A school owner in Enugu once asked me for “everything”. Fees, results, stock for the tuck shop, payroll, and a parent app. After two long meetings we agreed on three must haves: fee collection, receipts and a debtors list. She went live in five weeks. The parent app came eight months later, when her staff were already comfortable. If we had started with everything, I honestly believe she would have abandoned the system by the second term.

Mistake 2: Moving dirty data into a clean new system

An ERP is only as honest as the data you feed it. If your product list has three spellings of the same item and your stock figures came from a guess, the system will faithfully show wrong reports, quickly and beautifully. Staff will see the wrong numbers, decide the system is useless, and go back to their notebooks. The software gets the blame for a data problem.

Your product list is probably a mess

Open your current product list, whether it is in Excel, an old POS or a notebook. In a supermarket I audited in Port Harcourt, “Indomie Chicken 70g” existed under five different names, two with different prices. Nobody had ever noticed, because the counter staff simply picked whichever one they saw first. Reports for that item were split across five lines, so the best selling product in the shop looked like five average ones.

Clean-up means one name format, one unit of measure (a carton is a carton, a piece is a piece), a category, a barcode where one exists, a cost price and a selling price. It is boring work. It is also the most valuable week of the whole project.

Customer and supplier balances

Nigerian businesses run on credit. Customers owe you, and you owe suppliers. If you load these balances loosely, disputes start in week one: “I already paid Aisha last month, why does it show I owe?” Every opening balance should be backed by a document, an invoice, a signed delivery note or a bank transfer record. Where there is no document, agree a figure with the other party and write it down before go-live.

Opening stock and the cut off date

The system needs a starting point: what was in every store on the last day of the old process. That means a real physical count, on a fixed date, by two people, with the shop closed or at least quiet. Do not export the “system stock” from your old POS and trust it. In almost every business I have counted, the old system figure was wrong for 15 to 30 percent of items.

These are the figures I use when planning a project. If a vendor tells you data migration takes “one afternoon”, ask what exactly they are migrating, because they are probably copying your mess from one place to another.

Mistake 3: No owner, no champion, no decision maker

If nobody with authority cares whether the ERP succeeds, it will not. I say this to every client at our first meeting, because it is the mistake that owners find hardest to hear. Many people believe that paying a vendor is the same as delegating the project. It is not. A vendor can install software. Only you can change how your business behaves.

The owner must be visible

In the first three months, the owner or managing director should open the dashboard every morning, and ask one or two questions about what they see. “Why did Lugbe branch void nine sales yesterday?” “Why is this item at zero when we bought 40 cartons on Tuesday?” You do not need to become a software expert. You only need staff to learn that the numbers on screen are the numbers you use to run the business. The day your managers realise you look at the system instead of the WhatsApp group, behaviour changes.

Choose a super user

A super user is a staff member who learns the system deeply and becomes the first person others ask for help. Pick someone curious and respected, not just the youngest person who is good with phones. Give them time in their week to do this work, and a small allowance or recognition. In one distribution company in Kaduna, our super user was a 45 year old storekeeper called Musa. He was faster with the system than any of the young graduates, and his colleagues trusted his advice because he had done their job for fifteen years.

Who does what on the project

Many owners tell me they are too busy to sit in the project. Fine, but the roles below must still be filled, even if one person wears two hats in a small business.

Decide who decides

Every ERP project produces small arguments. Should discounts need approval? Who can change a price? Can a branch manager void a sale after payment? If nobody has the authority to settle these questions, the project stalls or, worse, each department gets its own rules. Name one decision maker, usually the owner or a trusted general manager, and agree that questions are settled within 48 hours.

Mistake 4: Skipping training and ignoring how people feel

Training is the first item to be cut when the budget gets tight, and the first thing people blame when the project fails. A two hour session in the office, with everybody watching a projector, is not training. It is a presentation.

Train by role, with real tasks

A cashier needs to know how to sell, take split payments, print a receipt and close the till. A storekeeper needs to receive goods against a purchase order, do a stock transfer and count stock. A manager needs to approve discounts and read the daily report. Each group should practise their own tasks, on real screens, with your real products. I like to run a short “test day” where staff process yesterday’s transactions in the new system and we compare the result with the old records. Mistakes made on a test day cost nothing.

Understand why people resist

Some resistance is pure fear. Staff who have been quietly rounding up prices, borrowing from the till or moving stock without paperwork know that a proper ERP will make those habits visible. Others fear that the computer will show they are slow. And many honest staff worry, quite reasonably, about being blamed for shortages that are not their fault.

The way to handle this is to be open. Tell your team that the system is there to protect honest people, because every stock movement will have a name and a time on it. When a shortage appears, you can now find out where it began instead of blaming the whole shift. Then follow through: do not use the first month’s reports to punish small errors. Use them to teach.

Plan for the Saturday rush

The worst time to learn a new system is when a queue of ten customers is watching. Before go-live, let each cashier process at least thirty practice sales, including the awkward ones: a refund, a price override that needs approval, a customer who pays half by transfer and half in cash. When the first real Saturday comes, keep the super user beside the tills for the whole day, and keep a calm supervisor on the floor. Ten minutes of patient help on the first busy day will protect you from weeks of quiet resistance.

Set a paper ban date

Give staff a clear date after which paper sales books and side spreadsheets stop. Before that date, keep the old process available as a fallback, but only for one or two weeks. Without a date, “temporary” paper workarounds become permanent, and the system never contains the full picture.

Mistake 5: Switching everything on in one big bang

A “big bang” go-live means every department, every branch and every module starts on the same Monday morning. It looks bold and efficient on a project plan. In reality, it guarantees that when something breaks, and something always breaks, you are firefighting in five places at once.

Phase by module

I usually roll out in this order for a trading business: selling and stock first, purchasing second, cash and expenses third, reports and accounting links last. Selling and stock come first because staff do them all day, so they learn faster. Purchasing follows because it depends on a correct stock list. Accounting comes last because it depends on everything else being accurate. My guide on ERP for multi branch businesses shows how this works when you have several locations.

Pilot in one place first

Pick one branch, one warehouse or one department to go first. Choose the team with the most cooperative manager, not the biggest one. Let them run for two to four weeks, fix the small problems, and then use them as trainers for the next group. A pilot turns your first mistakes into cheap mistakes.

A restaurant owner in Warri once insisted on going live with her kitchen stock, waiters’ orders and supplier payments all on the same Friday, because “the vendor said it was ready”. By Saturday night, the kitchen printer was not matching the orders, two waiters had gone back to paper, and her cashier had entered the same delivery twice. We paused, went back to orders and cash only for three weeks, and switched the kitchen stock on after that. It worked because we stopped trying to be clever.

Plan the go-live weekend and a fallback

Choose a quiet time, not month end, not December, not the week before a school term begins. Agree what happens if the internet is down or the power goes out. Keep the old system available read-only for a week or two. Here is a timeline I often use for a company with one main location and a small store.

Twelve weeks may sound long. It is not. It is faster than restarting a failed project after eight months.

Five more ERP implementation mistakes worth avoiding

The five above are the big ones. These next five are quieter, but I have seen each of them sink a project in Nigerian conditions.

Ignoring power and internet

A system that needs a constant connection will fail in the middle of a busy Friday. Ask whether the sales screen works offline (it keeps selling without internet and uploads later), and insist on a live test where the Wi-Fi is switched off. Then plan the basics: a UPS (a battery box that keeps a computer running for a few minutes during a power cut) for each till and router, a generator changeover that does not shut everything down, and a backup data line, for example an MTN router as the main line and an Airtel or Glo MiFi as reserve. As at 2026, a small UPS costs roughly ₦45,000 to ₦120,000 depending on size and brand.

Customising everything

Customisation is tempting because you feel the software should follow your way of working. Sometimes it should. But every custom change costs money, complicates upgrades and creates something else that can break. A good rule: change your process first, and customise the software only for the rules that make your business different from everyone else’s, such as a special commission structure or a unique credit approval flow.

Choosing on price alone

The cheapest quote often leaves out training, data migration, hardware and support. It also often comes from someone who will not answer the phone after payment. I have seen ₦2 million projects turn into ₦6 million after all the missing items were added, and the owner was left with a system that nobody understood. Compare quotes line by line, and ask for a written scope of what is included.

Not owning your data

Your sales and stock history is one of the most valuable things your business owns. Ask three questions: Can I export everything to Excel at any time? Where is my data hosted? What happens if you close your company or I stop paying? If the answers are vague, walk away. Data ownership should be written in the contract, not promised on WhatsApp.

No support plan after go-live

Go-live day is the start, not the finish. New problems appear in week two, week six and at the first month end. Agree in writing who supports you, at what hours, how quickly they respond, and how much it costs. Retail businesses are busiest on Saturdays and public holidays, when many vendors are switched off.

How much does ERP implementation cost in Nigeria?

Owners always ask this early, and vendors often avoid answering. Here is an honest picture for 2026. Prices change with the exchange rate and hardware market, so treat these as planning ranges, not quotes.

The hidden costs that surprise owners

• Staff overtime for the opening stock count and test day.
• Lost sales if you close early to count and forget to warn customers.
• Extra data if branches use more internet than you planned. Budget ₦25,000 to ₦70,000 per location per month.
• Your own time. Expect two to three hours a week for the first two months.

What a failed project really costs

The visible loss is the money you paid. The bigger loss is the year you did not get the control you wanted. If your business turns over ₦40 million a month and 2 percent leaks through wrong stock, unrecorded discounts and slow month end closing, that is ₦800,000 a month or nearly ₦10 million a year. A failed ERP means paying for the system and continuing to lose that money. A well implemented one usually pays for itself inside a year from leakage alone. If you want to see how I think about numbers like this, my article on data analytics for SMEs explains what to measure.

Your ERP implementation checklist for the first 90 days

Print this and stick it on your office wall. It covers what I ask my own clients to do, in order.

Before you sign

1. Write your one page problem statement and share it with every vendor.
2. Ask each vendor to demonstrate your real scenarios, including a credit sale, a split payment and a stock transfer.
3. Test offline mode with the Wi-Fi switched off.
4. Get a written scope, price list, delivery timeline and support terms.
5. Speak to two existing clients of a similar size, by phone, not only by testimonial.

During setup

1. Clean the product, customer and supplier lists.
2. Count opening stock on a fixed date with two counters.
3. Name the owner, the super user and the decision maker.
4. Set roles and permissions: who can sell, discount, void, transfer, and see profit.
5. Run the test day with real transactions from the previous day.

After go-live

1. Owner opens the dashboard every morning for the first 90 days.
2. Do a small stock count every week on 20 to 30 fast moving items.
3. Compare the system’s daily sales with bank, PoS and cash received.
4. Hold a 30 minute review each Friday for the first month.
5. Enforce the paper ban date without exceptions.

The weekly stock count deserves a special mention. Software does not stop theft or mistakes by itself. It makes them visible, and you only see them if you compare the system with the shelf. My advice on supermarket stock control tips explains a simple way to run these counts without closing the shop.

How to choose an implementation partner you can trust

The person or company that sets up your ERP matters as much as the software. A strong partner will save a weak product. A weak partner will ruin a strong one. I wrote a fuller guide on how to choose a technology partner you can trust long term, but here is the short list.

Questions to ask

• Who will actually do the work, and will I meet them before I sign?
• How many projects like mine have you completed in Nigeria, and can I call two of those clients?
• What happens if the internet goes down at a branch?
• Who owns the data and the source code, and can I export everything at any time?
• What is your response time for a problem on a Saturday afternoon, and what does that support cost?

Red flags

Be careful when someone promises to go live in three days, refuses to give a written scope, will not let you speak to previous clients, or insists on full payment upfront with no milestones. A fair payment structure ties money to delivered work, for example a deposit, a payment after data migration and training, and a final payment after go-live and a short warranty period.

Local partner or global brand?

Global ERP names bring strong products and long feature lists, but they are often priced in dollars, and their local partners can be thin on support. A Nigerian partner is closer, understands your payment mix, NEPA and PoS habits, and can visit your shop in Abuja or Lagos. Neither choice is automatically better. What matters is the quality of the actual people who will support you next year.

How to know your ERP is working after 90 days

Do not judge the project by how many modules are switched on. Judge it by outcomes. At the 90 day review, I ask clients to check these points against the success line in their problem statement.

If it is not working, do not quit yet

A struggling project can usually be rescued in four to six weeks. Stop adding new features. Recount stock in one location. Retrain the staff who are struggling, this time with real tasks. Fix the biggest data errors. Re-launch with a fresh paper ban date. Most of the projects I have rescued were technically fine and only needed this reset. If yours needs more than that, or you suspect the software itself is the problem, get an independent opinion before you pay another naira. You can see the kind of systems I build on my projects page.

Let the system grow with you

After the first 90 days, add the next phase, whether that is accounting links, a customer loyalty scheme, or an online store. If your business is heading towards spreadsheets that no longer cope, my article on signs your business has outgrown spreadsheets will help you decide what to build next.

Key takeaways

• Most ERP implementation mistakes are about people, data and planning, not about the software itself.
• Write a one page problem statement with numbers before you watch a single demo.
• Clean your product list, balances and opening stock before go-live. Never load a mess into a clean system.
• The owner must use the dashboard daily, and one respected staff member should become the super user.
• Train by role with real tasks, be open about why the system exists, and set a firm paper ban date.
• Roll out in phases and pilot in one place. Avoid the big bang go-live.
• Plan for NEPA and internet failure with offline selling, a UPS and a backup data line.
• Budget realistically as at 2026: from ₦1.5 million to ₦6 million for a configured system, and more for custom, plus hardware, training and yearly support.
• Check the results at 90 days against your original success line, and reset instead of quitting if it is off track.

Frequently asked questions

Your next step: avoid these ERP mistakes from day one

If you are planning an ERP project, or you are stuck in one that is not delivering, I can help you fix it properly. I build and rescue ERP software for Nigerian businesses, including inventory management systems that work offline and school management software for owners who want fees and records under control. The easiest way to start is to book a free consultation, or message me on WhatsApp from the site. Tell me what your business does, how many locations you have, and what is going wrong today. I will tell you honestly what to fix first, and whether you need a new system at all. You can also read the general overview of ERP on Wikipedia if you want the background.