ERP Business Analysis: What Is Decided Before Configuration
ERP budget overruns are decided in analysis, not configuration. See the process maps, requirement logs, data owners and integration lists to demand.

AI-generated header image for: ERP Business Analysis: What Is Decided Before Configuration
ERP budgets are rarely lost in configuration screens. They are lost earlier, when process maps are vague, requirements are undocumented, and no one owns master data. That is why serious buyers treat business analysis as the first control point, not a formality before the kickoff. Before a single setting is configured, the analysis phase should make your operating model explicit: how work flows today, which requirements are standard, which are gaps, and which are not applicable. It should name an owner for every master data set and inventory every integration that leaves or enters the ERP. This post explains why the analysis phase, not configuration, decides your ERP budget. It walks through process maps, requirement classification, data ownership, integration inventories, and the evidence a buyer can verify before signing. It also highlights red flags that signal thin analysis and shows what a documented analysis phase protects. The Standish Group's long-cited findings, including high cancellation rates and cost overruns, reinforce the lesson: if your partner cannot show maps, a requirements log, and a gap log with an owner per item, your signed estimate is a guess, not a price.
Why the Analysis Phase, Not Configuration, Decides Your ERP Budget
Scope, price, and risk in an ERP implementation are largely settled during business analysis, before anyone opens a configuration screen. Configuration executes decisions analysis already made. It cannot fix what analysis never decided.
That matters because the evidence is old and often misused. The Standish Group's 1995 CHAOS report, a study of IT projects generally and not current ERP data, found that 31.1% of projects were cancelled before completion, 52.7% cost 189% of original estimates, and only 16.2% finished on time and on budget, falling to 9% at large companies.
The part that still applies is subtler than the headline. The report warned that a design frozen in full detail, with no room for changes in business practice, becomes its own cause of failure. Requirements must be decided early and then defended through change control, not quietly renegotiated at go-live. It also noted that industry failures are "covered up, ignored, and/or rationalized," which is why buyers should ask for written artifacts rather than accept verbal reassurance. Poor scope discipline stays visible long after kickoff in the common reasons ERP implementations fail.
Our position as an ERP consultancy, not a research finding, is this: an estimate built without a process map, a requirements log, and a gap log with an owner per line is a guess with a price attached. Before you sign, review what to know before choosing ERP software.
Process Maps: The Operating Model on Paper Before It Becomes Screens
A process map is not a flowchart of what usually happens. It is the operating model written down before a screen exists. A usable map names the trigger, the sequence of steps, decision points, the role performing each step, the systems touched, inputs and outputs, and exception paths. A happy-path-only map cannot be configured or tested against, because exceptions are where configuration decisions actually live.
Coverage must reach order to cash, procure to pay, inventory and warehouse movement, production or project delivery, and the finance close. Map each to the depth where handoffs and approvals happen, not to departmental summaries. The notation is not improvised: business process diagrams follow a formal standard, ISO/IEC 19510 for BPMN, which defines process elements and their semantics.
That same map becomes the foundation for configuration decisions, test scripts, user acceptance testing, and training material, so any gap in it propagates through every subsequent phase.
Use this depth test in the room: if the map does not name who approves, who enters, and what happens when a step fails, the analysis is not finished. The ownership rule follows. Business process owners sign the map, not IT alone. That signature is where business process re-engineering begins.
Classifying Requirements: Standard, Gap, or Not Applicable
Two-way fit/gap leaves a third category unrecorded, and a requirement that will not be met is still a decision. Classify every line in the requirements log as standard, gap, or not applicable, with a named business owner on each row. A worked template, with illustrative rows:
Requirement | Classification | Business owner | Decision and justification | Impact if unmet | Linked test scenario |
|---|---|---|---|---|---|
Credit limit check before order release | Standard | Credit Manager | Configuration only | Over-limit orders ship | TC-014 |
Zone-based picking across warehouses | Gap | Warehouse Manager | Accept extension; effort 15-25 days | Manual picking, slower fulfillment | TC-022 |
Fixed-asset tax report format | Not applicable | Finance Director | Out of scope; tax software handles it | Report produced outside ERP | None |
Standard, configuration-only delivery, is generally the lowest-cost and lowest-risk path, though the right choice depends on the specific requirement.
Gap means customization, an extension, an ISV add-on, or a changed process. Every gap needs a named owner, an effort range, and an explicit accept-or-reject decision so it cannot resurface later as a surprise change request.
Not applicable means deliberately out of scope. Write the reason and the agreed workaround so it stays out. In practice, this classification is the one most commonly omitted, and an unrecorded "no" is indistinguishable from an oversight until go-live.
Gaps and standards both depend on business ownership, because ERP is a business project first. Practitioners in the field have argued for designing the process around the system rather than the reverse, a decision that belongs in this table.
One honest caveat on wording: these three labels are our working definitions. There is no single authoritative Arabic mapping for them, so agree the terminology with your partner at kickoff and keep one vocabulary across every document.
Naming a Data Owner for Every Master Data Set
Once requirements are classified, the next decision is who owns master data. Every set needs one owner: chart of accounts, customers, vendors, items and item categories, units of measure, price lists, warehouses and locations, employees, fixed assets, and opening balances. List each by job title and named person, add a deputy, and set a cut-off date by which the data must be cleaned and frozen.
Unnamed ownership delays go-live more than missing functionality. Cleansing, de-duplication, and validation sit with the business, not the partner, and nobody does it when nobody owns it. Missing features can be scoped and priced; dirty data cannot be configured around. ERP failure analyses have long listed the absence of a single owner among the leading causes of implementation breakdown.
Agree data quality rules during analysis: mandatory fields, numbering and coding conventions, duplicate rules, tax and regulatory fields required in your jurisdiction, and currency and unit conventions.
The analysis phase must produce a data migration plan covering sources, extraction method, validation approach, and reconciliation checks against opening balances. Freezing that plan is the gate that lets configuration begin, and it is where data debt becomes visible before it becomes a budget line, as we describe in our note on legacy ERP migration risks.
With owners named and data rules agreed, the next artifact is the integration inventory.
The Integration Inventory: What Leaves and Enters Your ERP
Master data ownership closes one gap. Interfaces close the next, because every interface is a contract between systems, not a favor between vendors.
Every line in the integration inventory must carry: source system, target system, direction, trigger (event, batch, or manual), frequency, data volume, format, and a named business owner on each side. Across Lebanon and the KSA, the common set runs from e-commerce and POS, CRM, banking and payment files, payroll, WMS or logistics, manufacturing shop-floor or weighbridge systems, and reporting or BI layers built on the Power Platform. Where warehouse movement is handled inside Business Central, the interface burden shrinks; warehouse management built into your ERP covers that case.
Decide failure handling before go-live, not during it: partial failure rules, who gets alerted, whether the transaction retries or queues, and the manual fallback. Answer the operational questions in writing: API limits, middleware, authentication, environment separation, and who supports the endpoint when the other vendor changes something.
The consequence is plain. An interface missing from the inventory surfaces during user acceptance testing or after go-live, when it is most expensive to build.
How a Buyer Verifies the Analysis Is Real Before Signing
With the integration inventory in hand, you can audit the rest of the analysis. Bring this yes/no table to the next meeting.
Verify before signing | Yes/No |
|---|---|
Process maps delivered and signed by process owners | |
Requirements log with classification and owner per line | |
Gap log with named owner and effort range per line | |
Master data owner list with dates | |
Integration inventory including failure handling | |
Every requirement traced to a test scenario |
Requirements are the blueprint for design, testing, and implementation, so missing traceability is not cosmetic.
The fastest test: if the same document could go to another company in another industry with only the logo changed, it is not analysis. Ask three questions: who wrote this, did that person sit with the people doing the work, and what remains open on the follow-up list?
Demand in writing the assumptions and exclusions list, plus how changes after sign-off are priced and governed. For a concrete view of the deliverables and decision points that should appear, see what a Business Central implementation actually looks like.
The earlier framing holds: no process map, requirements log, and gap log with owners means the number you are signing is a guess.
Red Flags That the Analysis Is Thin
The checklist tells you what should exist on your desk. These patterns tell you it will not.
Verbal assurance instead of documents. "We will handle it in configuration" is a promise, not a decision. If it is not written down with an owner and a date, it returns later as a change request.
Gap lists with no owner, no effort range, and no accept-or-reject decision. Every line quietly converts into a future change request, priced after your negotiating leverage is gone.
No named data owners and no freeze date. Cleansing, de-duplication, and validation belong to the business. With nobody accountable and no cut-off date, master data problems surface at go-live when they are most expensive to fix.
"We will integrate," with no list. No systems, directions, volumes, or failure handling. Regional regulatory and channel flows, including tax-reporting interfaces and POS-to-ERP integration, are usually the first casualties. Our breakdown of تحديات متاجر السعودية shows how much of that load sits outside the core system.
Analysis compressed into a week of workshops to protect the timeline. A fixed price then follows, built on assumptions nobody documented. U.S. GAO reporting on large ERP programs has long tied implementation risk to unfinished requirements and data conversion work.
One flag means ask for the document that closes it. Two or more means treat the quoted number as provisional.
What a Documented Analysis Phase Protects
The red flags are easier to spot in hindsight. What protects you is what you physically hold before you sign.
Five artifacts, by name
Holding all five artifacts before signing is the operational conclusion of every section above.
Locked vocabulary, agreed at kickoff, ensures every subsequent document uses the same three-way classification.
Data ownership, assigned by name and date during analysis, prevents cleansing debt from becoming a go-live delay, as covered above.
Keep the assumptions and exclusions list beside the contract and re-read it before approving every change request. That discipline also pays off when the platform itself changes, so file a Business Central 2026 Release Wave 2 readiness checklist alongside your analysis artifacts.
Index of Solutions works with organizations across the Middle East on Microsoft Dynamics 365 Business Central, LS Central, and the Power Platform, from analysis and design through customization, integration, training, and post-go-live support.
Conclusion
Your ERP budget is largely decided before configuration begins. The five artifacts, process maps, requirements log, gap log, master data owner list, and integration inventory, are the evidence that turns a quoted price into a committed scope. If you want that rigor, ask Index of Solutions to guide your analysis and design for Microsoft Dynamics 365 Business Central, LS Central, or the Power Platform, with support through customization, integration, training, and post-live care. Start with the artifacts. Then sign with confidence.
