Technical Lab · 0043

AS-IS to TO-BE process mapping — the complete practitioner's guide.

The single most misunderstood phase of every ERP project. What it actually means, how to execute it step by step, how to build the gap analysis that drives your scope, and every mistake I have seen kill Bangladesh implementations. A guide you can apply to your next project tomorrow.

Why This Guide Exists

Every consulting firm and ERP vendor will tell you that process mapping is important. Almost none of them will tell you how to do it, in detail, with the specificity that lets you actually run the sessions, build the documents, and defend the scope to a skeptical Managing Director.

This guide is the one I wish I had at the start of my first ERP engagement. It covers the full methodology — not just the theory. It includes the exact questions I ask in discovery interviews, the gap register structure I use on every project, real examples drawn from Bangladesh manufacturing and trading companies, and a full account of the mistakes I have personally seen derail implementations.

By the end, you will be able to run an AS-IS to TO-BE mapping engagement from first session to signed gap analysis document. This is not a 10-minute overview. It is a complete practitioner's reference.

Prerequisites

This guide assumes you understand what an ERP is and why companies implement them. If you're new to the ERP discovery process, read that article first. If you want BPMN diagram templates for Bangladesh manufacturing contexts, see Lab 0010.

01 — The Conceptual Framework

What AS-IS and TO-BE Really Mean

The terms are simple. The concepts are not. Most failed projects fail not because the team didn't know the terms, but because they applied them incorrectly. Let me be precise.

The AS-IS State

The AS-IS is the unfiltered, honest, ground-level reality of how your business operates today. Not how management thinks it operates. Not how the SOP manual says it should operate. How it actually operates — the WhatsApp approvals, the Excel workarounds, the paper registers, the verbal agreements, and the shadow systems that exist because the official process breaks down.

  • Includes every manual step and workaround
  • Includes exceptions as well as the standard path
  • Includes unofficial tools and communication channels
  • Includes pain points, delays, and error-prone handoffs
  • Sourced from operators, not managers

The TO-BE State

The TO-BE is your designed future workflow — how the business will operate inside the ERP after go-live. It is not a screenshot of Odoo's default behavior. It is a deliberate design that maps your business logic onto the ERP's capabilities, eliminating identified waste and automating identified manual steps.

  • Eliminates waste identified in the AS-IS
  • Maps your business rules onto ERP capabilities
  • Accounts for your industry's specific compliance needs
  • Is achievable within the ERP (native, configured, or customized)
  • Approved by process owners, not just management

The relationship between the two states is where the real insight lives. The delta between AS-IS and TO-BE is the project scope. Every item in that delta is a task: a configuration to build, a process to change, or a custom module to develop. If your gap analysis is complete, your statement of work is complete. If it is not, your statement of work is a guess.

The gap between AS-IS and TO-BE is not a problem to minimize. It is the project.

A second common mistake: treating TO-BE as "the Odoo standard." Odoo has default workflows. But "default" was designed for a generic business. Your business is not generic. A Bangladesh RMG company with a 5-tier approval structure for overseas LCs cannot simply adopt Odoo's out-of-the-box purchase approval. The TO-BE must be your process, designed by you, enabled by Odoo — not the other way around.

02 — Preparation

Before Phase 1: Setting Up for Success

A discovery engagement that starts without preparation produces bad data. You show up to the first interview and the department head has no idea what you need. You ask for documents and spend a week chasing them. The session produces wish lists instead of process descriptions. Protect against this with a structured preparation week.

Assemble the Right Team

Process mapping is a team sport. You need:

Business Analyst (you)Designs and runs the sessions, draws the maps, identifies gaps, owns the output documentation. Project SponsorThe senior stakeholder who has made the ERP decision. Their involvement is critical for gaining access, securing meeting time, and making "change the business" decisions later. Department LeadsOne per functional area (Procurement, Sales, Production, Finance, etc.). They know the intended process and the exceptions that management tolerates. Floor OperatorsAt least one per department. They know what actually happens. Do not skip this — it is the most important group. IT / Systems LeadKnows what other systems are in use and what integrations currently exist. Often overlooked and always critical.

Document Request List

Send this list at least one week before discovery starts. Receiving the documents in advance — even partially — dramatically improves session quality because you can ask informed questions instead of starting from zero.

Process documentsAny existing SOPs, process manuals, workflow diagrams — even if outdated. Outdated SOPs are useful: they tell you what management intended the process to be. Forms and templatesEvery paper form, Excel template, or printed register used in the process. Purchase requisitions, delivery challans, goods received notes, production job cards, inspection checklists. Current system accessRead-only access to the current system (Tally, SAP, legacy ERP, or custom software). Walk through it — don't just ask about it. Org chartWho reports to whom. This reveals approval hierarchies that might not match the described process. Volume dataNumber of purchase orders per month, number of sales invoices, number of production orders, number of employees by department. Critical for understanding complexity and sizing the implementation. Exception logAny recent audit findings, escalation emails, or "this always breaks" conversations. Gold.

Stakeholder Briefing

Before the first session, brief the project sponsor and all department leads on what discovery is and what it is not. Left unbriefed, department heads will treat the session as a product demo request and produce wish lists. Set these expectations explicitly:

What discovery is NOT

It is not a feature request session. It is not a demo of the new system. It is not a chance to redesign the entire company. It is not the place to debate whether you should have bought the software.

What discovery IS

It is a structured interview to understand how work currently flows through your department — step by step, including every exception, workaround, and approval. The output is a process map, not a feature list.

03 — Phase 1

Phase 1: The AS-IS Discovery

This is the most important phase of the project. The fidelity of your AS-IS map determines the quality of everything that follows: the TO-BE design, the gap analysis, the configuration spec, and the UAT script. Compress it, and the project pays the cost downstream — at UAT, at go-live, or both.

Step 1: The Gemba Walk

"Gemba" is a Japanese manufacturing term for "the actual place." A Gemba Walk means physically walking the process — following a single transaction from start to finish, in real life, on the factory floor or in the office, watching it happen.

This is not optional. It is the single most effective discovery technique I know. Every interview I have ever conducted was improved by the Gemba walk I did before it. Here is why: people describe the process they think they run. Walking the floor shows you the process they actually run — and the gap between the two is often enormous.

Bangladesh Context

In Bangladesh manufacturing, the Gemba walk often reveals critical hidden steps: delivery challans being counter-signed by a store manager who is not in the org chart, production job cards being maintained in a physical register that no one mentioned, or LC documents being processed through a WhatsApp group that includes three external parties. None of these will appear in an interview unless you saw them first.

Pick a single representative transaction for each major process area and follow it end-to-end. For procurement, start with a purchase requisition and follow it until the supplier invoice is posted. For production, start with a sales order and follow it until the finished goods are dispatched. Note every handoff, every document created or received, every system touched, every person involved.

Step 2: The Discovery Interview — Format and Facilitation

The interview session is where the map is built. Run one session per department. Do not combine departments — the dynamics change and cross-departmental tensions produce politically edited answers, not honest ones.

Session length90 minutes per department. Two hours maximum — concentration drops and answers become vague. AttendeesYou (BA) + department head + one floor-level operator. Maximum 4 people in the room. More and it becomes a committee meeting. Room setupWhiteboard or large paper on the wall. Draw the process in real time, in front of the attendees. This keeps them engaged and surfaces corrections immediately. No demosDo not show Odoo during discovery. The moment you show the system, the session becomes a feature evaluation, not a process interview. Screen stays closed. RecordingAsk permission, then record audio. You cannot write, draw, and listen simultaneously. Transcript review fills gaps you miss in the session. Total sessionsOne per department. For a 5–8 department company, 5–8 sessions across 1.5–2 weeks.

The floor-level operator is the most important person in the room. Every department head will describe the intended process. The operator will describe the actual one. Direct the questions to the operator when you need to understand exceptions and workarounds. Direct them to the department head when you need to understand approval authority and business rules.

Step 3: The 7 Discovery Questions

These questions work across every function — procurement, sales, production, finance, HR. Ask them in sequence for each major process step. The follow-ups are as important as the questions themselves.

Q1: "Walk me through what happens when [trigger event] occurs. What is the very first step?"

Why: Gets the process started. Forces specificity. "We process the order" is not an answer. "The sales executive receives the email and enters it in the register" is.

Q2: "Who does this step? What is their exact role and title? Who else can do it?"

Why: Maps role-to-task. Reveals single points of failure (only one person can approve). Critical for access rights configuration later.

Q3: "What information do you need to complete this step? Where does that information come from?"

Why: Identifies data inputs and their sources. Reveals hidden dependencies between departments. Uncovers shadow systems (WhatsApp messages, phone calls, separate spreadsheets).

Q4: "What do you produce at the end of this step? A document? A record? A notification? Who receives it?"

Why: Identifies outputs and handoffs. Every handoff is a potential automation point in the ERP. Every paper document is a data entry burden to eliminate.

Q5: "How long does this step normally take? What causes it to take longer?"

Why: Quantifies the pain. A step that "normally takes 2 days but sometimes 2 weeks" is a critical gap candidate. Volume and duration data also drives ROI calculations later.

Q6: "What goes wrong here? What are the most common errors or exceptions?"

Why: Exceptions are not footnotes — they are where the real process lives. Every exception is a business rule that must be captured in the TO-BE design. The standard path is easy; exceptions are where ERP implementations collapse during UAT.

Q7: "Is there anything in this process that you think is unnecessary, broken, or just done 'because we've always done it'?"

Why: Opens the door to process improvement. Sometimes operators know exactly which steps are waste but have never been asked. This seeds the TO-BE workshop.

Rule

Never accept a vague answer and move on. "We get approval from management" is not specific enough. Who? Which level? How long does it take? What happens if they are traveling? What is the minimum amount that requires approval vs the amount that doesn't? Every vague answer is a gap that will surface during UAT at the worst possible time.

Step 4: Document Artifact Collection

After each session, collect every physical artifact associated with the process. Walk to the person's desk and ask them to show you. This step is routinely skipped and routinely regretted.

Paper Forms
Purchase requisitions, delivery challans, gate passes, production job cards, inspection reports, quality check sheets. Photograph or scan each one. Note the fields — they tell you what data the business actually tracks (vs. what the system should track).
Map to ERP fields
Excel Templates
The master production planning sheet. The stock reconciliation workbook. The daily dispatches register. Each Excel file is a future ERP report or automation. Understand the calculation logic — not just the output.
Map to ERP reports
Registers & Books
Physical registers (store register, vehicle log, inward register) are common in Bangladesh factories. These must be captured — if the ERP doesn't replace them, they become a parallel data source and a compliance risk.
Digitization scope
WhatsApp / Email Trails
Ask to see how approvals actually happen. In most Bangladesh SMEs, the procurement manager sends a screenshot to the MD's WhatsApp for approval on any order above a threshold. This is a real approval step and must be replicated (or replaced) in the ERP.
Approval workflow
Current System Screens
Ask the operator to show you the current system (Tally, legacy ERP, custom software) in action. Screenshot every relevant screen. The field names and data entry flow often contain process logic that isn't in any document.
Data migration source

Step 5: Drawing the AS-IS Map — BPMN Basics

Your AS-IS process map must be a formal diagram, not a handwritten sketch. It needs to be readable by your client, your configuration team, and a third-party auditor. The industry standard for this is BPMN 2.0 (Business Process Model and Notation). You do not need to master every BPMN symbol. You need the five core elements:

Event
Start or end of a process. A trigger (PO received) or outcome (invoice posted).
Task
A discrete unit of work. One action, one actor, one output. Keep each task atomic.
Gateway
A decision point. Yes/No branches, parallel splits, conditions. Gateways capture exceptions.
Sequence Flow
Arrows showing order. One direction only. Loops are valid (approval rejected → revise → resubmit).
Swim Lane
A horizontal band for each role or department. Tasks sit inside the lane of who performs them.

Use swim lanes religiously. The swim lane view instantly makes visible every handoff across department boundaries — and handoffs are where delays, errors, and duplicate work accumulate. In a typical Bangladesh manufacturing procurement process, you will find 6–9 handoffs between Purchase, Accounts, Store, and Management. Each one is a candidate for automation.

Recommended tool: Draw.io (diagrams.net). It is free, browser-based, requires no installation, has BPMN shape libraries built in, and produces files that can be shared without a paid license. For most Bangladesh projects, this is the right choice. If the client has a Lucidchart or Microsoft Visio license, use those instead — the principle is the same.

Draw the map collaboratively. Project it on a screen during the interview session and draw in real time. When participants see the map taking shape, they immediately correct errors and add detail. A map drawn after the meeting from notes alone misses 20–30% of the nuance.

Step 6: AS-IS Validation and Sign-Off

Once the AS-IS maps are drafted, run a dedicated validation session with each department. Show them the map and ask one question: "Is this accurate?" Do not ask if they are happy with the process. Do not ask if they want to change it. Ask if the map reflects reality.

Common corrections at validation:

After all departments have validated their maps, get formal written sign-off (even an email) from the department head and the project sponsor. This is the AS-IS baseline. Any future claim that "the process has always worked this way" can be measured against it. The AS-IS sign-off also freezes the scope of what is being changed — without it, the discovery phase never truly ends.

04 — Phase 2

Phase 2: Designing the TO-BE State

The TO-BE design phase is where the Business Analyst shifts from recorder to architect. You are no longer just documenting — you are designing. The discipline required is different: you must challenge assumptions, eliminate waste, and hold your ground when people want to rebuild the old process in a new system.

Lean Principles Applied to Office and Factory Processes

Lean manufacturing identifies seven forms of waste (Muda). Every one of them applies to office and administrative processes — not just the factory floor. Use this framework to evaluate every step in your AS-IS map:

WASTE 01

Overproduction

Generating reports nobody reads. Printing documents that go straight into a drawer. Creating purchase orders for items already in stock.

WASTE 02

Waiting

A purchase order sitting in an inbox waiting for approval. An invoice on hold because the GRN isn't done yet. A production order delayed because the BOM hasn't been updated.

WASTE 03

Transport

Walking a paper form from one desk to another. Emailing a file for someone to print, sign, scan, and email back. Physical handoffs that add no value.

WASTE 04

Over-processing

Five-layer approval for a $50 purchase order. Three copies of the delivery challan when one would suffice. Data entered in three separate systems.

WASTE 05

Inventory

In process terms: information queued up and waiting. Unprocessed purchase requests. Unreviewed expense claims. Unposted journal entries. WIP that doesn't move.

WASTE 06

Motion

Searching for information that should be centrally available. Re-entering data that already exists in another system. Looking up customer credit terms that should surface automatically.

WASTE 07

Defects

Manual data entry errors. Invoices posted with wrong amounts. Stock adjustments covering counting mistakes. Every defect has a cost — find them in the AS-IS before they migrate to the TO-BE.

For each AS-IS step, ask: which waste category does this step exhibit? Steps with no waste are kept. Steps with waste are redesigned in the TO-BE — either automated, eliminated, simplified, or consolidated.

The TO-BE Design Workshop

Do not design the TO-BE alone. Run a structured workshop with the department leads, the project sponsor, and the lead configurator (the Odoo consultant who will build the system). The BA facilitates; the others contribute.

Session formatHalf-day workshop per process area. Full-day for complex areas like Finance or Manufacturing. Input materialThe validated AS-IS map for the area. The waste analysis. A summary of Odoo's native capabilities in this area (from the lead configurator). MethodStart with the AS-IS map on screen. Walk through each step. For each step, ask: does this step survive into the TO-BE? If yes, how? If no, what replaces it? Decision authorityProcess owners decide on business changes. The lead configurator decides on system capability. The BA decides on scope and documentation. The project sponsor resolves deadlocks. OutputA TO-BE process map (BPMN format) for each process area, with a list of decisions made and outstanding questions flagged.

The most important facilitation skill is pushing back on "we need it to work exactly like it does today." This phrase will appear in every session. It is sometimes legitimate (a compliance requirement with no flexibility) and sometimes pure inertia ("it's how we've always done it"). Your job is to tell the difference.

When someone says the process must stay as-is, ask: "If we didn't have this step, what would break?" If they can name a specific consequence — a regulatory requirement, a contractual obligation, an audit finding — keep the step. If they cannot name a consequence, the step is a candidate for elimination or redesign.

TO-BE Design by Process Area

The specific design questions vary by function. Here are the key questions for each major area in a typical Bangladesh manufacturing or trading company:

Procurement (Purchase-to-Pay)
Can purchase requests be raised directly in Odoo rather than on paper or in Excel? What approval levels should survive, and at what BDT/USD thresholds? Can 3-way matching (PO → GRN → Invoice) replace manual reconciliation? How should partial deliveries be handled? What happens to open POs at financial year-end?
Odoo: Purchase → Inventory → Accounting
Sales (Order-to-Cash)
Who can create a quotation? At what point does it become a confirmed order? Is credit limit enforcement required before order confirmation? How are partial shipments handled — one invoice or multiple? How are export orders (with LC) handled differently from domestic orders?
Odoo: Sales → Inventory → Invoicing
Manufacturing (Plan-to-Produce)
Is production driven by sales orders (MTO) or by a forecast/MPS (MTS)? Are Bills of Materials stable or frequently revised? Are work centers tracked individually or by batch? Is scrap and rework tracked at the operation level? How is quality inspection integrated — before or after production?
Odoo: Manufacturing → Inventory → QC
Inventory (Receive-to-Issue)
How many warehouses and locations must be tracked? Are lot or serial numbers required for traceability? What triggers a replenishment — min/max rules, reorder points, or MRP? How are physical counts conducted and reconciled? Is landed cost allocation required on incoming shipments?
Odoo: Inventory → Purchase/Manufacturing
Finance (Record-to-Report)
What is the chart of accounts structure? Are multiple currencies used and at which exchange rate (PO date, invoice date, or payment date)? Is Mushak VAT tracking required? How are bank reconciliations done currently? What reports does the MD review monthly and at what frequency?
Odoo: Accounting → Reporting

What TO-BE Is NOT

Three specific things that will corrupt a TO-BE design if left unchecked:

05 — The Gap Analysis

The Gap Analysis: Turning Mapping Into Scope

The gap analysis is the document where discovery meets execution. It is the formal record of every difference between the TO-BE design and the ERP's native capabilities, along with a decision on how each gap will be closed. Without it, scope is undefined. Without defined scope, the project has no boundary, and scope creep is inevitable.

Building the Gap Register

The gap register is a structured spreadsheet. Every gap identified during TO-BE design is logged as a row. The columns capture everything needed to understand, decide, and track the gap through implementation.

Minimum columns for a usable gap register:

Column What to Capture Example
Gap ID Sequential reference (G-001, G-002). Used for tracking and cross-referencing with configuration specs. G-017
Process Area Which functional area the gap belongs to. Procurement
AS-IS Description How the business does it today. Concrete and specific. LC documents tracked in a separate Excel workbook maintained by the Accounts team
TO-BE Requirement What the ERP needs to do. Written as a requirement, not a feature request. All LC documents must be attached to the corresponding purchase order and tracked with expiry date alerts
Odoo Capability What Odoo can do natively in this area. Document attachment on PO is native. LC-specific fields and expiry alerts are not.
Gap Type Category of the gap (see below). Configuration + Development
Decision How the gap will be closed: change business / workaround / develop. Develop: add LC tracking fields on PO + email alert for expiry <30 days
Effort Estimate Development hours, configuration effort, or business change effort. 8–12 dev hours
Priority Must-have for go-live / Nice-to-have / Future phase. Must-have
Owner Who is responsible for resolving this gap — client or consultant. Consultant (development)
Status Open / In progress / Resolved / Deferred. Open

Gap Categories

Every gap falls into one of four categories. The category determines the type of work required to close it — and different categories carry very different risks and costs:

STANDARD GAP

Odoo can do this natively — if configured

The feature exists in Odoo but has not been enabled or is set to a different default. Three-way matching, multi-currency, approval workflows, reordering rules — these are all standard capabilities. Configuration gaps are the lowest risk and lowest cost to close. They require no code, only correct settings.

CONFIGURATION GAP

Odoo can do this with significant setup work

Not just enabling a feature — requires setting up complex chart of accounts, building approval rule trees, configuring multi-company intercompany flows, or creating elaborate pricing structures. The capability is native; the implementation effort is non-trivial. Underestimating configuration gaps is a common cause of project overruns.

DEVELOPMENT GAP

Odoo cannot do this natively — custom code required

The requirement cannot be met with any combination of standard features and configuration. A custom Python module or XML view override is required. Development gaps carry the highest risk: custom code must be tested, maintained, and upgraded with every Odoo version. Never accept a development gap without fully understanding the ongoing maintenance implication.

BUSINESS CHANGE GAP

Odoo can do this — but the business process must change

The ERP has a good native way to handle this, but the current business process is incompatible with it. Instead of forcing custom code, the business should adapt its process to the ERP standard. This is almost always the cheapest, most maintainable solution — but it requires change management and often the most political resistance. This is where your relationship with the project sponsor matters most.

The 3 Decisions for Every Gap

When you present a gap to the project team, you must also present a recommended resolution. There are three choices — and they are always the same three choices, regardless of the gap:

Decision A
Recommended first
Change the Business. Adapt the business process to work within Odoo's native capability. This is the cheapest option to implement and the cheapest to maintain long-term. It eliminates custom code that must be updated with every Odoo version. It is often resisted by stakeholders but is frequently the correct answer — especially when the AS-IS process is demonstrably inefficient.
Cost: Low
Risk: Low
Effort: Change mgmt
Decision B
Workaround. Use a creative combination of native Odoo features to achieve the TO-BE requirement without writing custom code. For example: using the "notes" field and a custom stage in a kanban view to simulate a lightweight LC tracking system, rather than developing a full custom module. Workarounds are faster and cheaper than development but require more user discipline and are more fragile.
Cost: Medium
Risk: Medium
Effort: Design
Decision C
Last resort
Develop. Write a custom Python module to force Odoo to meet the TO-BE requirement exactly. This is appropriate when the requirement is genuinely non-negotiable (regulatory compliance, contractual obligation) and no workaround is adequate. Always document: who will maintain this code? What happens when Odoo 19 is released? Every custom module is a long-term liability that must be factored into the total cost of ownership.
Cost: High
Risk: High
Effort: Dev + QA

Custom code is a long-term debt. Only take it on when the business requirement is genuinely non-negotiable.

06 — Bangladesh Case Examples

Real Bangladesh Examples: AS-IS to TO-BE in Practice

Theory solidifies through examples. Here are three complete AS-IS to TO-BE transformations drawn from real Bangladesh implementations, with the key gaps and decisions for each.

Case 01

RMG Company — Purchase-to-Pay Process

A ready-made garments manufacturer with 1,800 workers, 3 buyers, and a procurement spend of approximately BDT 40 crore/year. Pre-implementation process was almost entirely paper-based.

AS-IS (how it actually worked):

Requisition
Production supervisor writes a requisition slip by hand. Gives it to the store manager. Store manager checks physical register to see if it is in stock. If not, writes a purchase memo on a preprinted form.
Paper register
No system
Approval
Purchase memo sent to MD via the Manager. MD approves via WhatsApp message ("ok") for amounts under BDT 50,000. For larger amounts, MD must physically sign the form — which can take 3–5 days if MD is traveling.
WhatsApp approval
No audit trail
Ordering
Procurement manager calls 2–3 suppliers verbally for rates. Selects supplier. Raises a handwritten purchase order or WhatsApp message. No formal PO number system — suppliers reference by date.
No PO system
Verbal rate comparison
Receiving
Supplier delivers with a DC (delivery challan). Store manager checks quantity manually. Enters into physical stock register. Forwards DC to Accounts. Accounts has no visibility of what was ordered — no matching is done.
No 3-way match
Manual entry
Payment
Supplier sends invoice (often a week or more after delivery). Accounts matches against the DC physically. Raises payment voucher. MD signs cheque. No payment terms enforced — suppliers are paid when the MD has time to sign.
No payment terms
Manual matching

Key gaps identified:

Case 02

Bicycle Manufacturer — Production Order Flow

A bicycle and cycle parts manufacturer exporting to Europe and the Middle East. Multi-level BOMs, component imports with LC, and a complex quality inspection regime.

Most complex AS-IS gap found: The production schedule was maintained in a large Excel workbook shared via Google Drive. The production manager updated it daily. Every department (Purchase, Store, Assembly, QC, Dispatch) maintained their own copy, leading to version conflicts and production stops when component availability data was stale.

TO-BE design: A single source of truth in Odoo Manufacturing. Material Requirements Planning (MRP) drives purchase orders automatically from confirmed production orders. All departments see live stock positions in Odoo. The Excel workbook is eliminated entirely.

Critical gap — multi-level BOM: The bicycle has assemblies within assemblies (frame sub-assembly, wheel sub-assembly, gear sub-assembly, final assembly). Odoo Manufacturing supports multi-level BOMs natively. The gap was not the software — it was the data. The existing BOMs had not been formally maintained for 3 years and contained obsolete components. The gap was a business change gap: the client needed to clean and validate all BOMs before go-live. This became a 6-week data migration workstream that had not been anticipated in the original project plan — precisely because the AS-IS phase had been rushed.

Case 03

Trading Company — Multi-Warehouse Inventory

An importer and distributor of industrial goods with 3 warehouses across Dhaka and Chattogram. Inventory valuation disputes between the MD and Accounts were the primary driver of the ERP decision.

AS-IS finding: The root cause of the valuation disputes was discovered during the Gemba walk. Each warehouse was maintaining inventory in a separate Excel file at a different cost basis. Dhaka warehouse used last-purchase cost. Chattogram warehouse used average cost. The head office rolled them up manually once a month — with adjustments made "by judgment" when the numbers looked wrong. There was no single, consistent inventory valuation method in the company.

TO-BE decision: Enforce Average Cost (AVCO) valuation across all three warehouses in Odoo, configured as a single company with multiple warehouse locations. This was a business change gap — the decision required the MD to accept AVCO as the standard, which changed the reported gross margin on certain product lines. The gap analysis made this explicit. It was debated and resolved before configuration started, not during UAT.

Lesson

In all three cases, the most valuable output of the AS-IS phase was not the process map — it was the question it forced. The RMG company had never asked whether WhatsApp approvals created an audit risk. The bicycle manufacturer had never asked who owned the BOM data. The trading company had never asked why their inventory valuation was inconsistent. A rigorous AS-IS phase surfaces questions that no demo, no RFP, and no vendor presentation will ever ask.

07 — Anti-Patterns

8 Mistakes That Derail AS-IS to TO-BE Mapping Projects

These are not hypothetical. Each one is drawn from a real implementation that ran into serious difficulty — some of which I was called in to rescue after the fact.

M01

Interviewing only managers

Department heads describe the process as it should work. Floor operators describe it as it does work. When you only interview managers, your AS-IS map documents the intended process — which is often dramatically different from the actual one. The most valuable 20 minutes of any discovery session is spent with the person at the lowest level who touches the process daily.

M02

Starting configuration before AS-IS sign-off

Vendor pressure to "show progress" leads to configuration starting while discovery is still underway. The result: the configuration team makes assumptions that contradict the AS-IS findings. Rework is expensive — a configuration built on wrong assumptions costs 3–5x the original effort to fix, because you are now fixing a running system rather than a blank one.

M03

Designing TO-BE as "Odoo standard"

The TO-BE is presented as a walkthrough of Odoo's default workflows without any adaptation to the client's business. The client approves because they don't know what they're approving. Then UAT arrives and every tester says "this is not how we work." A TO-BE must be the client's process inside the ERP, not the ERP's process imposed on the client.

M04

Skipping the exception paths

The standard process is easy to map. The exception paths are where implementations fail. What happens when a supplier delivers 80% of the order? What happens when the MD rejects a purchase approval at midnight before a critical production run? What happens when a customer returns half of a fully-invoiced order? Every exception is a test case waiting to fail during UAT. Map them in the AS-IS. Design them in the TO-BE.

M05

No formal gap register

Gaps are discussed in meetings and noted verbally. The project proceeds. By UAT, nobody can agree on what was in scope. The client insists that a feature was promised; the consultant insists it was never scoped. Without a written, signed gap register, both parties are guessing. The gap register is the contract. It must be formal, written, and signed before configuration starts.

M06

Treating every gap as a development requirement

Some consultants — especially those without strong Odoo configuration depth — default to custom development for gaps that could be solved with configuration or business change. Custom code is expensive to build, expensive to test, and most expensive of all to maintain across Odoo version upgrades. Every development gap should be challenged with the question: "Can we change the business process instead?"

M07

No stakeholder sign-off at each phase

The AS-IS is completed but never formally approved. The TO-BE is designed but never formally validated. The gap analysis is shared but never signed. Without formal sign-off at each phase, every subsequent decision is contested. Get written approval (even email) at: AS-IS completion, TO-BE validation, and gap analysis freeze. These three sign-offs protect both sides.

M08

Confusing data migration with process migration

The AS-IS mapping reveals process complexity. It also reveals data complexity — but they are not the same problem. Process migration (changing how people work) is addressed by the TO-BE design. Data migration (moving historical records into the new system) is a separate workstream with its own timeline, data owner, and quality gate. Conflating them in the gap analysis creates confusion about what is a process gap and what is a data gap. Track them separately.

08 — Tools

Tools for Process Mapping

The tool matters less than the discipline. A well-facilitated session captured in Draw.io will produce better output than a poorly facilitated session captured in the most expensive enterprise process mapping software. That said, here are the tools I recommend and when to use each:

Tool Best For Cost Bangladesh Fit
Draw.io (diagrams.net) BPMN process maps, AS-IS and TO-BE diagrams. Browser-based, no install, exports to PNG/PDF/XML. BPMN shape library built in. Free Excellent — no license cost, works on any browser, files sharable without tool access.
Lucidchart BPMN and flowcharts with more polish. Better for client-facing deliverables where presentation quality matters. $9–15/user/month Good — but adds cost for the client team if they need to edit.
Microsoft Visio Enterprise-grade process documentation. Full BPMN 2.0 support. Standard in larger corporates. $5–15/user/month Good for clients already in the Microsoft ecosystem.
Miro / FigJam Collaborative workshop tools. Real-time co-editing. Ideal for the TO-BE design workshop where multiple people need to contribute simultaneously on a virtual whiteboard. Free–$10/month Good for remote workshops. Less ideal for formal BPMN output.
Microsoft Excel The gap register. A well-structured Excel workbook with consistent columns and dropdown validation is more usable than many dedicated tools for this purpose. M365 subscription Excellent — universally understood, no learning curve, easy to share and filter.
Notion / Confluence Living documentation of the gap analysis, decision log, and project assumptions. Best when the consultant and client are collaborating on the same platform. Free–$8/user/month Growing adoption in Dhaka corporate sector. Requires client buy-in.
09 — The Bigger Picture

Connecting AS-IS/TO-BE to the Broader Implementation

Process mapping is not a standalone phase. It is the foundation that every subsequent project phase builds on. If the foundation is weak, every downstream phase is more expensive to execute:

→ Configuration
The gap register is the configuration team's primary input. Each Standard and Configuration gap becomes a configuration task. The TO-BE map is the reference for every setup decision. Without a complete TO-BE, configurators make assumptions — and assumption-based configuration fails at UAT.
Input: Gap register
Output: Configured system
→ Data Migration
The AS-IS artifact collection identifies every master data source: the supplier list in Tally, the product list in Excel, the customer database in the legacy CRM. The data migration plan starts here. The AS-IS also reveals data quality issues — duplicate suppliers, inconsistent product codes, missing pricing — that must be resolved before migration.
Input: AS-IS artifacts
Output: Migration plan
→ UAT Scripts
Each TO-BE process flow becomes a UAT test scenario. Each exception path becomes a UAT edge case. A well-written TO-BE map produces a UAT script almost automatically. For a detailed guide on UAT scripting, see Lab 0009.
Input: TO-BE maps
Output: UAT scripts
→ Training
The TO-BE process maps are the training material. Instead of training users on "Odoo" generically, you train them on "how to raise a purchase request in Odoo" — which means showing them the TO-BE flow they helped design, in the system they helped configure. Role-specific training maps directly to the swim lanes in the TO-BE diagram.
Input: TO-BE by role
Output: Training plan
→ Change Management
The delta between AS-IS and TO-BE is the change that end users must absorb. The bigger the delta, the larger the change management effort. Visualizing the change with side-by-side AS-IS and TO-BE maps during town halls and department briefings reduces anxiety and makes the upcoming change concrete rather than abstract. See Lab 0007 for the change management framework.
Input: AS-IS + TO-BE delta
Output: Change plan
The Bottom Line

AS-IS to TO-BE process mapping is not a deliverable you produce to satisfy a project manager's checklist. It is the intellectual work of understanding a business and designing its future. Done well, it makes every phase that follows faster, cheaper, and more predictable. Done poorly — or skipped — it guarantees the most expensive kind of project failure: the kind that is discovered at go-live, in front of 200 waiting users. If you want to build this skill systematically, my Requirements Gathering course walks through discovery interviews and process mapping step by step. Need a BA to run this for your next ERP project? Let's talk →

Frequently asked questions

What is AS-IS process mapping?

AS-IS process mapping is the act of documenting exactly how a business operates today — including every manual step, workaround, exception, and unofficial approval. It is not an idealized description of the intended process. It is an honest capture of the actual process as performed by the people doing the work. The key distinction: managers describe the process they intend; operators describe the process they execute. A good AS-IS map reflects the operator's reality, verified and signed off by the manager.

What is TO-BE process mapping?

TO-BE process mapping designs how the business will operate after the ERP is implemented. It is not a description of how the ERP works by default. It is a blueprint of how your business will run inside the ERP, with waste eliminated, approvals streamlined, and integration points explicitly designed. The TO-BE is authored by the Business Analyst and process owners together — not delivered by the software vendor as a product demo.

How long does AS-IS to TO-BE process mapping take?

For a mid-size Bangladesh company with 5–8 departments, expect 3–5 weeks: 2 weeks for AS-IS discovery and validation, 1 week for TO-BE design workshops, and 1–2 weeks for gap analysis documentation and sign-off. For a larger or more complex company (15+ departments, multi-site, multiple entities), plan 6–10 weeks. Rushing this phase is the single most expensive shortcut in ERP implementation — every week saved in discovery typically costs 3–5 weeks in UAT rework.

What is a gap analysis in ERP implementation?

A gap analysis identifies every difference between your TO-BE process design and what the ERP can do natively. Each gap is then categorized — Standard, Configuration, Development, or Business Change — and a decision is made on how to close it: adapt the business process, build a workaround, or write custom code. The gap analysis directly produces the configuration scope, the development backlog, and the change management agenda. It is the document that turns a process design into a project plan.

Do I need BPMN to map processes? Can I use simple flowcharts?

You do not need to master BPMN to map processes effectively — a clear, consistent flowchart with swim lanes serves most purposes. However, BPMN notation is worth learning for two reasons: (1) it is internationally standardized, so any BA or configurator on the team can read it without a key; and (2) its specific symbols for gateways, events, and parallel flows capture nuance that generic flowcharts miss. The minimum set to learn: circles for start/end events, rectangles for tasks, diamonds for decisions/gateways, and horizontal swim lanes per role. That covers 90% of what you need on a real project.

What tools are used for process mapping?

Draw.io (diagrams.net) is the best starting point — it is free, browser-based, and has a full BPMN shape library. For workshops, Miro or FigJam allow real-time collaborative editing when multiple people need to contribute simultaneously. For the gap register, a well-structured Excel workbook remains the most universally accessible format. For documentation, Confluence or Notion work well if both sides are willing to adopt a shared platform. The tool is secondary to the discipline; a great map in Paint is more valuable than a mediocre map in Visio.

What is the biggest mistake in AS-IS to TO-BE process mapping?

The most consistently damaging mistake is interviewing only managers and not floor-level operators. Managers describe the process as it should work. Operators describe it as it does work. These two descriptions diverge significantly in every organization — and the gap between them is exactly where ERP implementations run into trouble during UAT. The second-biggest mistake is starting ERP configuration before the AS-IS is formally validated and the gap analysis is signed off. Both mistakes have the same consequence: expensive rework discovered too late in the project lifecycle.

When should I recommend custom development for a gap vs changing the business process?

Recommend custom development only when: (1) the requirement is genuinely non-negotiable — a regulatory compliance requirement, a contractual obligation, or a technical constraint with no workaround; and (2) no native configuration or creative workaround can meet the requirement adequately. Before recommending development, ask the client: "If you had to choose between changing this process to work with Odoo natively, or paying for development and maintenance of a custom module, which would you prefer?" Often, framing it as an ongoing cost changes the preference. Custom code should be treated as a liability on the project balance sheet — taken on only when the value unambiguously exceeds the long-term cost.