Moving Excel sheets, Access files and legacy software into one shared system, reconciled to the trial balance at every stage
Data migration and spreadsheet replacement covers auditing existing files, cleaning and deduplicating records, loading them into a new system and reconciling the loaded result against known totals. Both systems then run side by side before cutover. Originals stay untouched.
- Originals stay untouched, work runs on copies
- Each load is checked against a total the accounts team already uses, such as the trial balance, the ageing report or the last stock count
- Both systems run in parallel before cutover
The work covers stock and party sheets at distributors, appointment registers at clinics, site-wise cost workbooks at builders and legacy software whose vendor no longer answers the phone. Some projects are one clean import. Others carry several years of history across dozens of files, where the main job is deciding which spelling of a customer name is correct.
Where spreadsheet data becomes hard to trust
Multiple Copies of One Sheet
Several people keep their own copy of the same sheet and email it around, so each version holds edits the others lack. Which file is current becomes the argument.
One Customer Under Several Names
One customer can appear as a full name, a short name, a name with a town and an abbreviation, each carrying its own balance. Totals appear only after merging.
Pricing Logic Held in a Formula
A workbook may hold dealer pricing in nested IF conditions and VLOOKUPs written by someone who has since left, and the formulas work while nobody can explain them. Rates change. The formulas do not.
A Stock File Held by One Person
The stock file sits on one desktop, so when that person is absent the warehouse works from memory and sales may quote items that were sold yesterday.
Legacy Software Without Support
The application still runs on an ageing PC, but it has no export menu, no manual and no vendor who answers. Everyone fears touching the data.
Sort and Paste Errors
Sorting one column without its neighbours moves rates against the wrong items, and nobody notices until an invoice carries a price that was never quoted.
Month-End Assembled by Hand
Branch sheets, the collection register and the accounting software export are opened side by side and combined by hand into a report. By then the month has closed.
Spreadsheets record how a business actually works, built by people who needed something usable on a Tuesday afternoon, and that logic deserves to survive the move. The job is to carry it into a system several people can open at once, with every row accounted for as loaded, merged or held back with a reason.
Files read, mapped, loaded and checked against known totals
The work starts with a review of what exists: every file, who edits it, which copy is the master and what each formula does. Fields are then mapped, cleaned, deduplicated and loaded into a staging copy, never straight into the live system, and totals are reconciled against reports the business already trusts. Live use waits for agreement.
Business Rules Carried Across
Pricing logic held in a workbook is business knowledge. It is read, written out in plain language, confirmed with the person who uses it and rebuilt as a rule the new system applies to every quotation.
Reconciliation Against Trusted Reports
Party balances are compared with the trial balance, stock value with the last physical count and receivables with the ageing report. A load is accepted only when those totals match.
Parallel Run Before Cutover
Sheets and the new system both run for an agreed period, with numbers compared day by day on the same invoices, receipts and stock movements. Cutover then becomes a planned decision.
What changes after the move
What the business gets, before any talk of technology.
One Current Record Instead of Competing Files
Sales, accounts and stores open one record. A rate change reaches everyone at once, including a salesperson on a phone in a customer’s office, who no longer quotes from last month’s file.
One Ledger for Each Customer
Several spellings of one customer merge into a single party with its full history attached, so outstanding amounts and credit limits apply to the whole account, not one entry.
Opening Balances Tied to the Books
Receivables, payables and stock value in the new system are tied to the trial balance and the physical count, giving the accountant a firm basis for month-one review.
Formula Logic Documented
Logic that lived inside one workbook becomes a written rule that anyone can review, change on purpose and keep when staff leave, not a formula one person understood.
Role-Based Access to Sensitive Columns
Purchase rates, salaries and margins can be limited by role, so access no longer depends on who happens to open the shared folder or who forwards a file.
Older Years Kept as a Searchable Archive
Old years stay searchable. A question about an invoice from several years back no longer means opening a folder of dated files and guessing which copy is the final one.
What data migration and spreadsheet replacement includes
Everything from the first discovery session to production and long-term upkeep.
Spreadsheet and System Audit
Every file, database and application is listed with its owner, its editors, how often it changes, what it feeds and which copy the business treats as correct.
Business Rule Extraction
Pricing slabs, commission logic, discount rules and validation held in formulas and macros are read out and written in plain language. Users confirm each one.
Data Profiling and Quality Assessment
Files are counted for blank fields, invalid dates, text sitting in number columns, malformed phone numbers, tax ID numbers that fail the checksum and contradictory rows. It shows what is there.
Field Mapping and Target Design
Each source column is mapped to a field in the new system, and the map also lists values with no destination yet and fields nobody was recording.
Deduplication and Master Data Cleanup
Customers, vendors and items are matched across spellings, abbreviations and phone numbers, and every merge is proposed for approval instead of being applied silently by a script.
Data Cleansing and Standardisation
Names, addresses, units of measure, dates and codes are converted to one standard form, and each edit goes into a log beside the original value.
Extraction From Legacy Applications
Data is pulled from old software. Methods include reading the database directly, exporting through the report screens, parsing printed formats as a last resort and re-keying only what nothing else can reach.
Opening Balance and Accounting Reconciliation
Party balances, stock value, receivables and payables are carried in and matched against the trial balance or the last audited figures. The accountant takes part in that check.
Historical Data and Archiving
Cut-off year is agreed. Everything older goes into a searchable archive, so that nothing is thrown away and nothing stays behind on an old desktop nobody wants to switch on.
Test Loads and Reconciliation Reports
Loads are repeated into a staging copy, and each run produces a written report listing rows read, rows loaded and rows held back with reasons. Totals are compared with the business’s own reports.
Parallel Run and Cutover Planning
Both systems run for a set period. The plan covers a freeze window, a final load of changes made during the freeze and a written way back if the switch fails.
Post-Migration Support and Corrections
Support continues through the first weeks after cutover. It covers rows that look wrong, reports that do not tie and fields someone only thought of after real use began.
Where migration from spreadsheets is used
The processes this service is most often asked to set up and automate.
Distributor With Separate Branch Stock Sheets
Every branch maintains a stock sheet that head office totals only occasionally, and item codes vary from branch to branch, so one product shows up under several names. The work builds a single item master, matches it to the last physical count and carries party balances across against the trial balance.
Clinic Group With Registers Kept at Each Branch
Patient names are spelt differently at each branch, ages are stored as text and repeat patients are created afresh at every visit. Matching runs on phone number and date of birth. Doubtful pairs go to the front-desk supervisor.
Builder With Site Costing in Separate Workbooks
Material, labour, subcontractor and plant costs sit in separate files for each project, and none of them agree with the ledger. Cost heads map to one structure. Each site is matched to its ledger entries.
Manufacturer Running Old Software With No Export
The application runs on one desktop. The vendor cannot be reached. Recovering the item master, past job history and the parts lists means reading the database directly, and recovered totals are checked against printed reports the office kept.
Trading Firm With Rate Slabs in One Workbook
Dealer pricing comes from a formula that calls several hidden sheets, and its author has since left the company, taking the explanation with them. The logic is traced and written out as slab rules, confirmed by the sales head, then rebuilt as rules applied to each quotation.
Service Company With Sheets That Still Suit the Work
A firm with two coordinators and modest job volume is considering business management software, but its sheets are tidy, properly shared and understood by both users. The right step may be a shared online workbook with validation and access control, with a system revisited when volume grows.
Is this the right service for you?
It suits businesses that have outgrown manual processes and want a system they can rely on.
Businesses Where Several People Edit the Same Files
When two people cannot open a sheet at once without one losing work, the file has become a queue and the lost edits are never found. A shared system removes that limit.
Companies Adopting an ERP, CRM or New Application
The software is chosen, and years of customers, items and balances must now load correctly before go-live, so migration is scoped apart from configuration.
Businesses on Software Without Support
The old application still runs, but the vendor is silent, the machine is ageing and nobody knows how to get the data out if it stops. Extraction can be done before that day arrives.
Owners Who Need a Reliable Figure
When a question such as receivables owed or stock value takes days and several people to answer, the data is scattered. One system answers it from one place.
Teams Merging Branches or Companies
Two item masters, two customer lists and two sets of codes must become one, and people who know the records settle each difference. Load order never decides the outcome.
Businesses Where a Spreadsheet Still Fits
A sheet kept by one or two people, understood, backed up and shared properly, is working software and should stay as it is. Budgeting, one-off analysis and quick modelling suit Excel. If the only problem is a file on one desktop, a shared online workbook with validation and access control may be enough. The assessment says so where that applies.
What it can do, in plain terms.
Source Inventory
A register of every file, database and application, with its owner, update frequency and whether the business treats it as authoritative.
Field-Level Mapping Document
A map from source column to target field, with the transformation applied to each, written so an accountant can read it as easily as a developer.
Data Profiling Reports
Blanks, outliers, invalid formats, broken references and contradictory rows are counted and tabulated per column before any plan is agreed or any date promised.
Fuzzy Matching and Merge Review
Likely duplicates are found across spellings and phone numbers, and each pair appears with its evidence side by side so that a person makes the call.
Repeatable Load Scripts
The migration runs as code, so a correction means a re-run of the load scripts against the staging copy, not a week of redoing work by hand.
Staging Environment
Loads land in a copy first. The team opens real records in it and raises objections before anything touches the live system, which stays closed until the totals agree.
Row-Level Exception Handling
Rows that fail validation are held and listed with a reason, then returned for a decision by whoever owns that record in the business. Nothing is dropped silently.
Reconciliation Against Known Totals
Counts and values are compared with the business’s own reports, such as the trial balance, the ageing report, the stock valuation and last month’s sales register.
Full Audit Trail
Every conversion step is logged with its rule, which lets anyone trace a figure in the new system to the cell it came from.
Legacy Database Extraction
Old MySQL, SQL Server, Access, FoxPro and DBF files are read directly where the application offers no export, with the original database file copied first.
Archive Access to Old Years
History sits in a read-only archive. Live tables stay smaller, and a query about an old year can still be answered from the archive without restoring a backup or hunting for the old machine.
Rollback Plan
A written route back to the previous position, with the original files untouched, in case the first days of use go badly.
How the work runs, step by step.
From the first conversation to a release your team uses every day. Each step says what you provide and what you get back.
- 01
File Audit and Master Identification
Every file is collected and its editors are listed. The client provides the real files, not cleaned-up versions, and says which copy the business relies on when two disagree.
- 02
Profiling and Rule Extraction
Quality is measured and formulas are read. Macros are read too, to recover business rules. The client gives time with the people who use each sheet, because a rule counts as confirmed only when the person applying it agrees with the reading.
- 03
Mapping, Cleansing and Merge Decisions
The mapping document and the duplicate list are produced. The client decides which record survives a merge, which years come across and how to treat rows with no valid value.
- 04
Test Loads Into Staging
Data lands in a copy. A reconciliation report follows each run. The client provides two or three staff to open real records and flag anything wrong, plus the accountant to check balances.
- 05
Parallel Run
Both systems run for an agreed period. Results are compared day by day. The client enters work in both places. It is tiring but catches misread rules.
- 06
Cutover, final load and support
The sequence is a freeze, a final load of changes made during it and the switch, on a date the client picks, often a quiet weekend after month-end close. A named person stays reachable that first morning.
What you receive when the work is done.
Everything needed to run, maintain and extend the software without depending on us.
It connects to the systems you already use.
Tools that work stay. The new system is joined to them rather than replacing them.
Spreadsheets and Files
- Microsoft Excel
- Google Sheets
- CSV and tab-delimited exports
- Access databases
- PDF and printed reports
Accounting Systems
- QuickBooks
- Xero
- Zoho Books
- Sage
- Other desktop accounting software
Business Applications
- Zoho CRM
- Salesforce
- Odoo
- SAP Business One
- Custom and legacy in-house software
Databases
- MySQL and MariaDB
- PostgreSQL
- Microsoft SQL Server
- FoxPro and DBF files
- SQLite
Storage and Documents
- Google Drive and Shared Drives
- Microsoft OneDrive and SharePoint
- Network file shares
- Scanned document folders
Chosen for reliability and a long life.
Picked to match your scale and to stay maintainable for years.
Extraction and Loading
Matching and Cleansing
Staging and Storage
Document Extraction
Verification
Why owners choose GullySystem.
Formulas Are Read Before They Are Replaced
Nested conditions in a workbook often hold years of pricing logic, and that knowledge is recovered and written down before the file is retired. Nothing is lost with the file.
Failed Rows Are Listed, Not Dropped
Failed rows are held and listed. A migration that reports success while quietly removing rows is worse than one that reports its failures, because the missing customer surfaces only when someone chases a payment.
The Accountant Reviews the Balances
Opening balances are reconciled against the trial balance with the person who files the returns taking part, because a developer reading a sheet cannot assert them.
Parallel Running Is Part of the Plan
A parallel period is built into the project. It is repetitive work, yet it separates a quiet go-live from one that needs corrections for weeks while the team works around wrong figures.
Original Files Stay Unchanged
Work is done on copies, and the sources are kept exactly as they were and handed back at the end. The existing record survives whatever happens to the project.
A Spreadsheet May Be the Right Answer
Some businesses fit one shared workbook with validation and access control better than a new system, and the assessment states this where it applies.
How the engagement can be shaped.
Each model says what it suits and how the work is scoped. Pick the one that fits, or ask which one we would suggest.
Migration Assessment
Useful for a business holding files and an ageing system without a clear picture of the move. The file audit, quality report and mapping plan are one deliverable set, invoiced together. A data analyst reads the sources, and the sheet’s owner explains what each is for. No software is chosen or built at this stage. Additional sources get an estimate before they are opened. All documents stay with the client.
One-Time Migration Project
Fits a business whose destination system is already chosen. The fee is set after the sources are read and covers extraction, cleansing, test loads and reconciliation. A data engineer and a tester carry it out, and one finance or operations person signs off the reconciled totals. Altering a mapping rule triggers a fresh estimate. The project closes at cutover. Scripts, logs and source copies are handed to the client.
Legacy Data Rescue
For data stranded in software whose vendor no longer replies. Scope is the extraction of named databases or files into readable formats, invoiced as one task. A data engineer works from files and backups, and a long-serving user on the client side explains the old screens and reports. Whether a new system follows is decided later. More sources mean another estimate. Readable data and mapping notes stay with the client.
Migration Plus Build
Suits a business replacing spreadsheets with software while moving its history in the same project. Fees run by phase, with the sequence and dates set in the proposal. The team combines a project lead, developers, one data specialist and a tester, and the sheet’s long-time owner answers questions about the old rules. Small changes are absorbed in the sprint. Larger ones are priced first. The client takes ownership of every repository and every record at handover.
Questions owners ask first.
Plain answers to the questions asked before work starts.
How much of our history should we bring across?
It depends on how often the business looks back. Live records need current masters and open balances: customers, vendors, items, receivables, payables, open orders and stock, with a few years of transactions kept searchable in the new system. Older years sit in an archive. Copying every year multiplies the cleaning effort and carries over records no one opens, so the owner picks a cut-off year before the work begins.
Will we lose data during the migration?
No, provided the process described here is followed. Work runs on copies. The original files are never altered. Each load produces a report of rows read, rows loaded and rows held back with the reason, and totals are reconciled against trusted reports before the load is accepted. Failed rows return for a decision. After cutover, the previous position stays available.
How do you handle duplicate customers and items?
Matching is automatic, and the decision is human. Matching runs across spellings, abbreviations, phone numbers, tax IDs and addresses, and each likely pair appears with both records side by side. Someone from the client’s team decides which one survives and what happens to its history, and no merge is applied without review.
Our old software has no export option. Can the data still be recovered?
Usually yes. Most applications keep data in a database that can be read directly, such as MySQL, SQL Server, Access, FoxPro or DBF files. Where the database is locked, reports can be exported and parsed, and scanned documents can be read with OCR, but recoverability is checked before a quote is given. Anything that must be re-entered is stated up front.
Do opening balances have to match our accounting software exactly?
Yes, to the last digit or with a written explanation. Party balances, receivables, payables and stock value in the new system should tie back to the trial balance or the last audited figures. The reconciliation runs with the accountant present. The migration counts as complete when every difference is either zero or explained in writing, and an unexplained one keeps the old sheets running.
What is a parallel run, and can we skip it?
A parallel run means entering the same work in the old sheets and the new system for an agreed period, then comparing what each produces on the same day. It doubles the effort. Its value is catching a misread rule or a wrongly mapped rate slab while a working fallback still exists and the owner can still compare against the old sheets. A small, simple move can skip it. Anything involving money or stock should not.
How long will our team be tied up in this?
It is more than many owners expect. The team decides which record survives a merge, confirms that recovered rules match practice, checks sample records after each test load and enters work twice during the parallel run. The load clusters around decision points. It involves two or three people plus the accountant. Hours are confirmed after scoping.
Can you keep the Excel reports our owner is used to reading?
Yes. The new system can produce the same report in the same layout and export it to Excel on a schedule. Dashboards can arrive gradually. Nobody relearns the report format.
What if the person who built our spreadsheets has left?
The work can still proceed. Formulas can be traced, hidden sheets opened and macros read, and the logic written out in plain language for the people who use the output to check. A file does not record intent. Any rule that nobody can confirm is flagged and listed for the owner, and it is never carried across as an assumption.
What does a data migration cost?
Cost follows data condition, not row count. The factors are the number of separate sources, how consistent they are, how much is free text rather than codes and how many duplicates need decisions. Others are whether a legacy database must be reverse-engineered, how many years stay live and how long the parallel run lasts. The price is confirmed after the audit.
When should the cutover happen?
Usually just after a month-end or year-end close, at the start of a quiet week, when the books are settled and little is in progress. Peak seasons and filing deadlines are better avoided. The freeze window and the final load are planned around the client’s calendar, with the accountant and the warehouse lead both consulted first.
Can we move off spreadsheets gradually instead of all at once?
Often yes. Migration can begin with the area causing trouble, then move to the next, for example sales and stock first, then purchase, then costing. The sequence has to respect dependencies. Moving billing before the item master is ready only creates a second problem on top of the first, so the order is planned with the client. Some sheets run beside the system for a while.
Should we replace our spreadsheets at all?
It depends on how the sheets are used. If one or two people maintain a few sheets that are understood, shared online and backed up, replacement adds little. A system fits when several people work at once, when one figure exists in several places and disagrees, or when one cell error can produce a wrong invoice. It also fits when access control matters. Volume beyond one person’s tracking fits too. Share your requirements and the actual files can be reviewed to show which of these conditions apply, so the decision rests on your sheets.
Read before you decide.
Practical notes on the questions this service raises, not product marketing.
Shared Online Workbook or Business Software: Which Fits Your Team
When a shared online spreadsheet with validation is enough, when a proper system earns its place, and the questions that tell the two apart for a team.
What a Parallel Run Involves Before You Switch Off Spreadsheets
What running old sheets and a new system side by side means, what to compare each day, who carries the double entry and when the old files can stop.
How to Check Your Spreadsheet Data Before Migrating It
Find the master copy, count blanks and bad formats, spot one customer under several names and read the formulas before your sheets move to new software.
When Should Your Business Move from Spreadsheets to Software?
A spreadsheet works until several people, rules and records depend on it. Five tests that show when a business should move to software, and when to wait.
More guides on this page’s topic
- Data migration checklist: before, during and after cutover
- Duplicate customers in a spreadsheet: why it happens and how to merge
- How to set opening balances in new software from your old books
- Data reconciliation after migration: proving the totals match
- Inherited a spreadsheet nobody understands? How to read its formulas
- Old software with no export option: ways to get your data out
- How much history to migrate and how much to archive
- Choosing a cutover date: month end, quarter end or year start
- Phased or all at once: two ways to move off spreadsheets
- What a field mapping document is and what goes in it
Let’s Build the Right Software for Your Business
Tell us about your current operational challenge, spreadsheet bottleneck, or software requirement. An experienced engineer will review your workflow and reply within one business day.
- No obligation consultation
- Senior engineer reviews your brief
- Your operational details stay 100% confidential
Start with your requirements
Seven short steps that tell us what you need. A senior engineer reads them before anyone gets in touch.
Share your requirementsRelated services.
ERP, CRM and Business Software Implementation
The system the data moves into, selected, configured around your process and rolled out to the people who have to use it.
Custom Software and Product Engineering
When the logic in your workbooks is the business and no packaged product fits it, a system built around those rules.
Data Engineering and Business Intelligence
Once the data sits in one place, the pipelines and dashboards that turn it into the numbers your Monday meeting runs on.
Training and Adoption
The part that decides whether the sheets actually get retired: role-based training, help at go-live and adoption measured afterwards, so that the old files are genuinely switched off.