A second branch turns a business system into an operating-model question. Two tills can run the same software and still report different stock. A price change can reach one outlet and miss another. Cash can be collected in three places and reconciled in none. Buying software before deciding how branches should operate produces an expensive way to disagree.
What changes when you operate more than one location
Distance exposes every weak rule. Stock is no longer in one room, so it can be in transit, borrowed, damaged or counted twice. Cash is banked by different people on different days. Local managers make promises about delivery, credit and price that the head office discovers later.
- Authority: who can change a price, approve a refund, open credit or order stock.
- Stock: transfer requests, in-transit quantities, receiving confirmation and cycle counts.
- Cash: daily close, bank deposit, petty cash and variance approval.
- Reporting: one cut-off time so branch totals can be compared.
- Connectivity: what happens when a branch loses internet during trading hours.
Decide the branch operating model first
There are three broad models. A centralised model keeps pricing, purchasing and credit at head office. A federated model gives branch managers defined local authority within a shared framework. A franchise model separates brand and system standards from local ownership.
Whichever model you choose, write a decision-rights matrix before evaluating software. List price changes, local offers, purchasing, stock transfers, refunds, credit limits and hiring. For each, name the decision maker, the approver and the evidence. Software can enforce a decision; it cannot make one. A branch system that assumes central control will be bypassed in a federated business, and a system built for local freedom will fail in a franchise network that needs brand consistency.
Requirements that apply to every branch
- One product and price master: every outlet sells the same SKU with a controlled local price rule.
- Stock visibility and transfer: available, reserved, in transit and received quantities are separate states.
- Cash and banking: shift close, deposit, variance and approval are recorded against the branch.
- Purchasing and receiving: head office and branch orders follow the same approval limits.
- Sales and returns: a return can be traced to the original sale, even at another branch.
- Staff and permissions: a cashier, supervisor and branch manager see different things.
- Daily close and reporting: one cut-off, one definition of sales and one exception report.
- Customer records: consent and contact history travel with the customer, not the till.
- Offline operation: sales continue during an outage and reconcile when the connection returns.
System approaches and their trade-offs
A central ERP suite gives one data model and one reporting layer. It is strongest when the branches follow similar processes and the company can absorb a long implementation.
Best-of-breed systems with an integration layer fit each function well but move the complexity into integrations. Every connector needs an owner, a failure path and a reconciliation check.
A branch POS with nightly synchronisation is resilient to poor connectivity and cheap to run locally. Reporting is delayed, and conflicts such as a price change made in two places must be resolved by a clear rule.
A custom operations portal can model unusual transfer, credit or commission rules. It brings a maintenance obligation, so require documentation and an internal owner from the start.
A hybrid with managed spreadsheets can be a sensible first step for a small network. It becomes dangerous when the spreadsheet is treated as the source of truth for stock or cash.
A shortlisting scorecard for branches
| Criterion | Weight | Evidence to request |
|---|---|---|
| Multi-branch data model | 20% | Demo with two branches and a transfer |
| Stock accuracy and transfer control | 15% | In-transit, receiving and count workflow |
| Offline and recovery behaviour | 15% | Test with the connection disabled |
| Roles and approvals | 10% | Role matrix for cashier, supervisor and manager |
| Reporting and cut-off | 10% | Sample daily and branch comparison report |
| Integration fit | 10% | Named connectors and failure handling |
| Rollout and training | 10% | Phased plan with acceptance gates |
| Total cost of ownership | 10% | Three-year cost including hardware and support |
Pricing questions before you sign
- Is the price per branch, per user, per transaction or per module?
- What does implementation include: configuration, data migration, hardware setup and training?
- Are integrations quoted separately, and who maintains them when an API changes?
- What happens when a new branch is added: a fixed onboarding fee or a repeat of the full implementation?
- What support is included out of hours, and what is the response time for a trading-hours outage?
- What is the exit assistance and data export commitment?
Pilot in two contrasting branches
Choose one high-volume urban branch and one branch with weaker connectivity or a different customer mix. Run the same scenarios in both: a price change, a stock transfer, a return from another branch, an offline sale, a shift close and a cash variance. Compare the result with the current process, including the staff time needed to correct differences. A pilot that only runs in the best-connected branch tests the demo, not the model.
Rollout sequence and acceptance
- Fix the operating rules and freeze them for the pilot.
- Run the two-branch pilot for a full stock and cash cycle.
- Resolve transfer, pricing and permission failures before adding a third branch.
- Turn the corrected configuration into a branch template.
- Roll out in waves, with a freeze list and a hypercare window for each.
Acceptance should be evidence-based: a transfer reconciles at both ends, the daily close matches the bank deposit, permissions block the wrong role, and an offline sale appears once after reconnection. If those four tests fail, the software is not ready for more branches.
What the contract must settle
The contract should name the owner of the branch configuration, the data and the integrations. Require a documented export of products, stock, prices, customers and transaction history, plus assistance during a transition. Confirm what happens at termination, how long data is retained and who removes access. A rollout that cannot be reversed is a lock-in with a project plan.
Red flags in a branch rollout
- The demo uses a single branch with clean stock data.
- The vendor agrees to unlimited customisation without a change process.
- There is no offline plan or no conflict rule for synchronisation.
- Stock transfer rules are left to branch staff to interpret.
- Branch managers see the system for the first time during training.
- Module pricing hides integration, hardware or reporting costs.
- No exit clause, export format or documentation commitment.
Where this fits in a rollout
Our Business Automation division designs branch workflows, approval rules and reporting around the operating model you choose. The pricing page explains how a pilot and a multi-branch rollout are scoped. For the stock side of the same decision, read what inventory synchronisation costs in Bangladesh.
Sources
- ISO 9001:2015, Quality management systems, accessed 17 November 2026.
- COSO, Guidance on Internal Control, accessed 17 November 2026.
- Balanced Scorecard Institute, Balanced Scorecard basics, accessed 17 November 2026.
- ISO 22301:2019, Business continuity management systems, accessed 17 November 2026.
Responses (0)
No comments yet. Add the first one below.
Leave a response
Your email address will not be published. Required fields are marked *.