Manufacturing & custom software · Bengaluru · Whitefish, USA
Guides

Software for Multi-Branch Businesses: Designing Data, Permissions, Stock and Reporting

The design decisions that make business software work across several branches, cities or countries — separating branches from legal entities, shared and local data, permissions, stock transfers, reporting, currencies, time zones and phased rollout.

By Corposol Team··11 min read
Commercial buildings and shops along a city street
Photo: A wide street scene in an urban area with low-rise commercial buildings by PattayaPatrol, CC BY-SA 4.0 (cropped)

Software that works beautifully in one location often starts to crack at the second. Someone adds a "Branch" dropdown, and for a while that seems to be enough. Then the new branch needs its own invoice numbers. A manager in one city can suddenly see another city's margins. Stock sent between locations disappears from one count and takes three days to appear in the other. Head office asks for a group sales report, and three people spend two days combining spreadsheets because the same customer exists under four spellings.

Multi-branch capability is not a field you add to a form. It is a set of design decisions that reach into almost every table, screen, permission and report. They cost relatively little to make at the start and a great deal to retrofit. This guide works through ten questions that a business should answer before building or commissioning software for several branches. For the operational side of running branches, see our guide to multi-branch service operations; here the focus is on how the system itself should be designed.

1. What does "branch" actually mean in your business?

The word often hides several different concepts, and the software needs to separate them:

ConceptQuestion it answersExample
Legal entityWhich company issues the invoice and files the taxes?Company registered in each country
Tax registrationWhich registration applies to this transaction?Separate registration by state
BranchWhich unit serves the customer and owns the result?Pune service centre
Stock locationWhere are the goods physically?Pune warehouse, van 12, rack B
Reporting regionHow does management group the results?West region

These do not map one-to-one. One legal entity can run ten branches. One branch can hold four stock locations. A restructure can move a branch to another region without changing anything else. If the software treats all of this as one "Branch" list, every future change becomes a data problem. Separate them from the start, even if only one or two are used in the first version.

2. Which data is shared and which belongs to a branch?

This is the most consequential question. Go through each type of data and decide, in a table like the one below, who owns it. The answers below are a common pattern, not a rule.

DataCommon approachReason
Product and service catalogueShared; controlled centrallyConsistent codes make group reports possible
CustomersShared, with branch relationshipsOne customer can use several branches
PricesGroup default with local overridesLocal markets differ, within limits
StaffShared, with roles by branchPeople transfer and cover other locations
Orders, jobs and invoicesOwned by one branchClear responsibility and accounting
Stock balancesOwned by a stock locationGoods can only be in one place
Numbering seriesSet by entity or branchLegal and audit requirements

Shared customers need extra thought. Where branches create customers separately, duplicates are inevitable unless the system checks for a matching phone number, email or tax ID at the moment of entry. A clean shared customer record lets a national client be served consistently, with a single view of their outstanding balance. But be clear about which branch owns the relationship, or two teams may chase the same payment.

3. One database or several?

For most single-owner businesses, the simplest strong design is one application and one database, with a branch recorded on each transaction. Group reports, transfers and shared customers are then straightforward, and an update reaches everyone at once.

The main risk is an information leak: a query that forgets to filter by branch shows one branch's data to another. A good team deals with this centrally, by applying branch restrictions in one shared layer of the application rather than rewriting them on every screen, and by testing the restrictions automatically.

Separate databases can be appropriate when branches are truly independent businesses, when customer contracts or local law restrict where data may be held, or when a franchisee owns its own data and may leave. The cost is real: updates must be applied to each copy, consolidated reporting requires an extra data pipeline, and shared master data has to be synchronised. Choose it for a specific reason, not because it sounds safer.

4. Who can see and do what, and where?

A plain role such as "Manager" is not enough. Multi-branch permissions have three dimensions: the action (view, create, approve), the type of data (stock, customers, margins) and the scope (own records, own branch, region, group). Written as a matrix, the rules become easy to review:

RoleStock availabilityCustomersMargins and profitDiscount approval
Sales executiveAll branchesOwn and assignedNoneUp to their limit
Branch managerAll branchesOwn branchOwn branchUp to branch limit
Regional managerAll branchesOwn regionOwn regionUp to regional limit
Finance teamAll branchesAll branchesAll branchesNone

Also allow for people who work across branches. A user should hold a role at a branch, with the option of holding different roles at different branches. The same person might be branch manager in Nashik and relief supervisor in Pune. Temporary cover, with an end date, avoids the common habit of sharing a colleague's login during leave. Audit logs should record both the user and the branch context of important actions.

5. How does stock move between branches?

Stock transfers are the most common source of multi-branch errors. The usual mistake is to subtract from one location and add to another in a single step, as if the goods teleported. Consider an illustrative transfer of 20 units from Pune to Nashik, shipped on Monday and delivered on Wednesday, of which only 19 arrive in good condition:

EventPune on handIn transitNashik on handPending investigation
Start1000300
Transfer requested and approved (20 reserved)1000300
Dispatched on Monday8020300
19 received on Wednesday800491
Shortage resolved (written off with approval)800490

In every row, physical units are accounted for, so the group total never silently changes. The one missing unit is visible instead of being lost in a later stock count, and someone has to make a decision about it. "Available to sell" can also differ from "on hand": Pune has 100 on hand after approval, but only 80 can be promised to its own customers.

Decide, too, whether the transfer is simply an internal movement or has a financial side. Transfers between branches of the same entity may require only internal documents, while movements between legal entities or tax registrations can be treated as taxable sales. Valuation can also be affected if branches carry goods at different costs. Your accountant should confirm these rules for each country and state you operate in.

6. How are money, numbering and tax handled?

  • Document numbering. Many jurisdictions require sequential, unique invoice numbers, sometimes by registration or financial year. Build configurable series, and allocate the number at the moment a document is finalised, so that draft orders do not leave gaps.
  • Tax. Rates and rules should be data rather than code. They may depend on the seller's registration, the customer's location and the item. Save the applied tax on each line, so a later rate change cannot alter an old invoice.
  • Currency. Each transaction needs its original currency and exchange rate, and each entity needs its own base currency. Group reports then convert using a clearly stated rate policy. Mixing these up produces totals that no one can later explain.
  • Inter-branch balances. When one branch delivers work sold by another, decide up front how revenue and cost are split, or managers will spend each month arguing over the numbers.

7. What about time zones and local working hours?

Store timestamps in a standard form, typically UTC, and display them in the relevant local time. But be careful about which "local" you mean. A service deadline should use the customer site's time zone. A shift report uses the branch's. A group report needs one stated convention. A branch's "Monday sales" may be different from head office's "Monday sales" unless the reporting day is defined. Business hours, public holidays and weekends should be configured per branch, particularly if your system calculates service-level deadlines or attendance. The same applies to language, date formats and document templates in different countries.

8. Which local variations will you allow?

Every branch believes it is different. Some are. If the software allows everything, you lose standardisation and comparable reporting. If it allows nothing, staff return to WhatsApp and spreadsheets for whatever the system doesn't handle. A useful test sorts each request into one of three groups:

  • Fixed across the group: codes, status definitions, approval principles, audit rules and financial controls.
  • Configurable by branch: price lists, working hours, teams, service areas, document footers and local tax settings.
  • Handled by exception: a genuinely unique requirement, reviewed centrally and either generalised for the group or declined.

Prefer configuration over custom code branches. Five conditional rules saying "if branch = Chennai" are an early warning that a real concept is missing from the design. Perhaps it should be a setting such as "requires a second approval above this value", which every branch can then use.

9. Can the reports be trusted and compared?

Group reporting is usually the reason a business invests in shared software, but comparison depends on definitions. If one branch closes a job when the technician leaves and another when the invoice is paid, a report showing "average completion time" is meaningless. Define status meanings before designing dashboards.

Comparisons also need context. Raw revenue favours the largest branch, so include measures relative to scale, such as jobs per technician, gross margin percentage, debtor days or repeat-visit rate. Let managers drill down from the group total to a region, a branch and the individual documents behind each figure. If a number can't be traced to its source, branch managers will rarely accept it. Decide how shared costs are allocated, and show results both before and after the allocation, so managers are measured on what they control.

10. How will you roll it out, and open the next branch?

Big-bang launches across many branches concentrate every risk into one week. An illustrative rollout for eight branches might look like this:

  1. Prepare the shared data. Clean the catalogue, remove duplicate customers and agree the group codes. This is usually more work than anyone expects.
  2. Pilot one representative branch, with a capable manager but ordinary conditions. Avoid choosing only your best branch, or you won't uncover typical problems.
  3. Stabilise for a few weeks. Fix issues, refine the training and record the questions people actually ask.
  4. Roll out in waves of two or three branches, using staff from the pilot as local champions.
  5. Close the old system branch by branch, with a stated end date and a validated opening stock count and customer balances.

Transfers between a migrated branch and one that has not migrated need an interim process, so plan it deliberately. If the branches are too interdependent for that, consider migrating them in logical clusters.

Finally, make opening a new branch a planned function of the system. A short setup process should create the branch, its stock locations, numbering series, tax settings, users, price overrides and reporting region, ideally from a template. If adding a branch needs a developer, the software will eventually slow down the very growth it was built to support.

Warning signs in an existing system

If you already run several branches, these symptoms suggest the design needs attention:

  • Staff have multiple logins in order to work at different branches.
  • Reports are exported and combined by hand before management can read them.
  • The same customer exists several times with different balances.
  • Stock differences appear after transfers, and no one can trace them.
  • Staff can see data they shouldn't, or keep private spreadsheets for data they can't.
  • Code contains branch-specific behaviour that only one developer understands.
  • Opening a new branch triggers a small software project.

Using these questions in your project

You do not need to build every part of this at once. A first phase might support two branches, shared customers, basic transfers and one group report. But the underlying structure — separating entities from branches, recording branch ownership, scoping permissions and tracking stock in transit — should be in place even if it's lightly used at first. Adding screens later is ordinary work. Reworking the data model after years of live transactions, across a business that can't stop operating, is not.

Frequently asked questions

Should each branch have its own database?

For a single business, usually not. One shared database with a branch recorded against each transaction makes group reports, shared customers and transfers much simpler. Separate databases can be justified by strict legal, data-location or independence requirements, but they make consolidated reporting harder.

Can a branch see another branch's stock without seeing its financial data?

Yes, if permissions are designed by action and by data type rather than simply by branch. A sales user might see stock availability across the group, yet see customers, margins and reports only for their own branch.

How do we handle a user who works at more than one branch?

Assign the user a role at each branch instead of tying them to one home branch. They can then switch working branch or view a combined list, and every action they perform is recorded against the correct branch.

Can different branches use different prices or tax settings?

Yes. Design a group default with controlled branch or regional overrides. Each order should keep the price and tax values used at the time, so later changes to the rules do not alter historical documents.

Should we roll out software to all branches at once?

Usually not. A pilot in a representative branch, followed by waves of branches, lets you fix problems while the impact is small. Simultaneous launches are normally reserved for situations where branches are too interdependent to run on separate systems.

  • Multi-branch software
  • Multi-location business
  • Custom software development
  • Permissions
  • Inventory transfers

Let's design software for your branches

Tell us about your process or idea. You'll get a free consultation and a clear, fixed quote — no obligation.

  • Free consultation
  • Prototype before you commit
  • Fixed, itemised quote
  • You own the software

Get a free quote

We reply within one working day. Your details are never shared.