Migrating From an Old System: The Data Clean-Up Nobody Warns You About
When Wekesa decided to move his wholesale business from an old desktop accounting program to a modern cloud system, he imagined the hard part would be learning the new software. He was wrong. The hard part was discovering that his customer list contained "Mama Njeri", "Mama Njery", "Mama Njeri Kawangware" and "Njeri (DO NOT GIVE CREDIT)". All four were the same person. She owed money under two of the names and had paid under the other two. Nobody mentions this part when you plan a data migration in Kenya, or anywhere else.
This is the data clean-up nobody warns you about. Every system you have ever used has been quietly collecting mistakes, duplicates and abandoned records for years. Moving to the cloud does not fix them. It just moves them somewhere shinier, unless you deal with them first.
Why old data is messy
Data gets messy for perfectly human reasons. Different people entered information in different ways over many years. Old systems often had no rules about required fields or formats. Staff came up with workarounds, like typing notes into the phone number field because there was nowhere else to put them. Nobody ever deleted anything, just in case.
Common problems you will find:
Duplicate customers, suppliers and products with slightly different spellings
Phone numbers in mixed formats, some with country codes and some without
Records for customers who closed shop years ago
Notes, warnings and gossip typed into fields meant for addresses
Products that no longer exist but still have stock balances
Opening balances nobody can explain
Do not migrate the mess
It is tempting to move everything as it is and promise to tidy up later. Later never comes. Once dirty data is in the new system, it starts feeding reports, invoices and reminders. Suddenly Mama Njeri gets four statements, two of them insisting she owes money she already paid. She will not be amused, and she knows where your shop is.
Clean before you move. It is slower at the start but saves months of confusion afterwards.
A practical data migration clean-up plan
You do not need special tools to begin. A spreadsheet export and a patient person will get you most of the way.
Export everything from the old system into spreadsheets, one file per type of record
Decide what you actually need to move, since you may not need ten years of closed invoices
Sort by name to spot duplicates, then merge them carefully, keeping the correct balances
Standardise formats for phone numbers, dates and addresses
Remove or archive records that are clearly dead
Agree on opening balances with your accountant before you import anything
Assign a named person to own each set of data. Customers belong to sales, stock belongs to the store, balances belong to accounts. When everyone owns the data, nobody does.
Keep the old system around for a while
Do not switch off the old system the day the new one goes live. Keep it available, read-only if possible, for a few months. You will need it to check historical details, answer the auditor and settle the occasional argument about who paid what in 2019. Make sure you also keep a full export stored safely, in case the old machine finally gives up during the next blackout.
Test with real work
Before going live, run a trial with a sample of real transactions. Raise a few invoices, receive a few payments and print a statement. Compare the results to the old system. Achieng in accounts will spot problems in five minutes that the project team missed in five weeks, because she actually uses the numbers every day.
The reward
Data clean-up is unglamorous work. Nobody throws a party when the duplicate suppliers are merged. But clean data is what makes a new cloud system genuinely useful. Reports become trustworthy. Reminders go to the right people. Your team stops second-guessing every number.
Wekesa spent three weeks cleaning his data before migrating. It was tedious. But the first statement the new system sent to Mama Njeri was correct, addressed to the right name, and showed the right balance. She paid the same afternoon via M-Pesa, which Wekesa insists was the best return on investment of his entire career.
Moving off an old system soon? Start with one export and one afternoon of sorting this week. If the mess turns out bigger than expected, talk to Philsoft about planning the migration, so your new system starts life clean.