A founder of a 35-person trading company sat across from me last month and said, almost word for word, what I have heard at least fifty times: "Once we put in the ERP, everything will be sorted." I had to gently say no. The ERP will not sort anything by itself. The work that decides whether the project succeeds or stalls happens in the six to eight weeks before anyone touches the software. That is the uncomfortable truth most vendors will not tell an SME owner during a sales demo.
In my experience guiding SME ERP projects across trading, contracting, property and services, the pattern is consistent. The companies that prepare properly implement in roughly half the time, with far fewer change requests, and they actually use the system a year later. The ones that skip preparation end up in a familiar loop: blame the consultant, blame the software, then quietly go back to Excel. The difference is almost never the product. It is the readiness of the business before configuration starts.
- Most SME ERP failures are caused by weak preparation, not by the software itself.
- Master data should be cleaned before migration, not patched after go-live.
- Approval rules and exception handling must be agreed before configuration, in writing.
- Reporting requirements should connect to real management decisions, not vanity dashboards.
- SMEs that invest six to eight weeks in preparation typically implement in half the time.
Prepare Process Ownership
Before any software conversation, every important workflow in the business needs a name attached to it. Sales orders, purchase requests, supplier invoices, inventory movements, project claims, service jobs, payment releases and month-end closing all need an owner who is accountable when something goes wrong. The ERP can route a document, but it cannot decide who is responsible for it. That has to come from the business.
I once worked with an aircon contractor running four branches across the Klang Valley. Every branch manager assumed the head office finance team "owned" supplier invoices, while head office assumed the branches did. The result was a stack of unmatched delivery orders nobody felt accountable for. No ERP would have fixed that, because the gap was human, not technical.
Questions to answer before configuration
Sit down with your operations head and finance manager, and force yourselves to answer the following honestly, role by role:
- Who creates, reviews, approves, corrects, and closes each workflow?
- Who is allowed to approve an exception, and at what threshold?
- Who is permitted to change master data such as a customer's credit limit?
- Who actually reads the reports, and what decision do they make from them?
If you cannot answer those four questions for the top ten workflows in your business, you are not ready to configure an ERP. You are ready to start preparing for one. That is an important distinction.
Clean Master Data Early
This is the part SME owners hate hearing, because it is unglamorous and time-consuming. But duplicate customers, three different spellings of the same supplier, missing tax codes, vague item groups, ancient stock codes and unverified bank accounts will sabotage your go-live more reliably than any technical issue.
Why dirty data hurts more in an ERP than in Excel
In a spreadsheet, the human reading it can mentally merge "ABC Sdn Bhd", "A.B.C. Sdn. Bhd." and "ABC SB" into one customer. An ERP will treat them as three customers, generate three statements, age three sets of receivables, and confuse your salespeople. Migrating messy data is the single fastest way to lose user trust in a new system within the first month.
The cleaner the data you bring in, the less rework you will do after go-live. Here is the practical scope to review before any migration script runs:
Customer, supplier, item, chart of accounts, project, asset, employee, location, tax code, payment terms, and approval users. For example, a property facility company juggling 80 buildings I worked with discovered 1,400 active "assets" in their old system, of which roughly 300 had been decommissioned years ago. Cleaning that list before migration saved them weeks of post-go-live confusion on preventive maintenance schedules.
Treat this as a business exercise, not an IT one. The people who know whether a customer is still trading, or whether a supplier has been blacklisted, sit in sales and procurement, not in IT.
Define Approval Rules
ERP configuration lives or dies on approval logic. Purchases, payments, discounts, credit notes, stock adjustments, project claims and supplier changes all need clear rules by amount, by department, by role and by exception type. The mistake I see most often is SMEs describing approval rules verbally during a workshop, then changing their mind three times during user acceptance testing because nobody wrote it down.
The "below RM5,000" trap
A common SME pattern is: "Below RM5,000, the branch manager can approve. Above RM5,000, it goes to the director." Sounds simple. Then someone asks, what about a recurring monthly RM4,000 service contract that totals RM48,000 a year? What about a split PO of RM4,900 followed by another RM4,900 the next week? Real approval rules need to cover those cases before a developer encodes them, otherwise you end up rewriting the workflow twice.
Write the rules out on paper, get the managing director to initial them, and only then hand them to your implementation partner. That single document will save you weeks.
Clarify Reporting Needs
Resist the temptation to ask for "every report the old system had, plus dashboards." It is one of the most expensive habits in SME ERP projects. Start instead from the management decisions you actually make each week and each month, and work backwards to the report that supports each one.
The reports that genuinely matter for most SMEs are surprisingly few: margin by project, overdue receivables, stock ageing, payment exposure to top suppliers, service SLA performance, unresolved claims, approval bottlenecks and a rolling cash forecast. If a report does not change a decision someone is going to make, it is decoration. Decoration is fine later, but not in phase one.
Decide What Not to Customise
Every customisation adds cost today, testing time next month, training effort at go-live, and upgrade risk every year for the next decade. The discipline of saying no is one of the most valuable things an SME owner can bring to an ERP project.
The "we have always done it this way" test
Before approving any customisation request, ask one question: is this a real business requirement, or is it a workaround we invented because the old system could not do it properly? Roughly half the customisation requests I see from SMEs fall into the second category. The aircon contractor I mentioned earlier had a beloved "manual service report adjustment" screen that everyone insisted was essential. It turned out the screen only existed because their old system rounded labour hours wrongly. The new ERP handled it correctly out of the box. We saved them about RM18,000 in custom development by asking why, not how.
Prepare Users for Change
ERP changes daily behaviour, and that is the part SMEs consistently underestimate. Storekeepers may need to scan goods in on receipt instead of writing them on a clipboard. Site supervisors may need to upload photos to a job before claiming progress. Finance staff may need to give up the personal Excel files they have nursed for a decade. None of that is a software problem. It is a people problem dressed up as one.
Training that actually sticks
Training should explain the business reason first, and the button clicks second. If a storekeeper understands that scanning at receipt prevents the company from paying twice for the same delivery, they will scan. If they only know "click here, then click there", they will revert to the clipboard the first week the system feels slow. Invest the time to tell people why, and adoption looks after itself.
A 6-Week SME Preparation Playbook
To make this concrete, here is the structure I typically run with an SME before any software configuration begins. It assumes a company of 30 to 100 staff with one or two operating entities. Adjust the pace, not the sequence.
Week 1 — Scope and sponsorship
The managing director nominates a project sponsor, an internal project lead, and module champions for finance, sales, procurement, inventory and operations. We agree the in-scope entities, modules and go-live target, and write a one-page charter everyone signs. No software discussion yet.
Week 2 — Process mapping
Each module champion walks through their current workflow in their own words. We map it on a whiteboard, note the pain points, and tag every step with an owner. By end of week, the top ten workflows are documented in plain language, not in software screens.
Week 3 — Master data audit
Export every master data set from the existing systems and spreadsheets. Sales and procurement review customers and suppliers for duplicates and dormant records. Operations review the item and asset master. Finance reviews the chart of accounts. By end of week, you should have a target "clean list" target count for each data set.
Week 4 — Approval rules and exceptions
The managing director, finance manager and operations head sit in one room and agree the approval matrix for purchases, payments, discounts, credit notes and stock adjustments. The output is a single signed document. This is usually the most uncomfortable week, and the most valuable.
Week 5 — Reporting and must-have vs nice-to-have
Every requested report and customisation is sorted into three columns: must-have for go-live, useful in phase two, and decoration. The managing director makes the final call on disputed items. Anything in column three is parked, not killed, which makes the conversation easier.
Week 6 — Change readiness and partner briefing
Communicate the project to all staff, explain why the change is happening, and identify the users who will need the most hand-holding. Hand the prepared package — process maps, clean master data, signed approval matrix, prioritised report list — to the implementation partner. Configuration starts the following Monday, and it starts on solid ground.
The Honest Close
The hardest part of SME ERP work is not the configuration. It is convincing a busy owner that six weeks of unglamorous preparation is more valuable than six weeks of additional software features. I understand the temptation to skip ahead. Preparation does not feel like progress, and there is no demo to show the board. But every shortcut taken in these six weeks is paid back, with interest, somewhere between user acceptance testing and the third month after go-live.
If you are an SME owner about to start an ERP project, the single most useful thing you can do this week is not call another vendor. It is to sit down with your operations head, pick the top ten workflows in your business, and try to name an owner for each one. If that conversation is hard, you have just identified where your real work begins. The software comes later, and it works much better when it does.