Operations
Migrating from spreadsheets to a core: a sequence that does not destroy PAR
A go-live sequence for Ghanaian lenders still running loans or hire purchase in Excel — what to migrate first, what to freeze, and how to run parallel.
Michael Addo, Implementation Director · 20 May 2026 · 16 min read
Spreadsheet books fail in predictable ways: multiple versions, broken SUMIFs, and a hero who is on leave. Migrating them fails in predictable ways too: too much history, cashiers untrained, and a parallel run with no freeze date. This sequence is the one we use on OrbitMFI go-lives for HP and MFI books.
Week 0 — freeze the story
Name the products, branches, and live accounts. Agree what 'outstanding' means. Photograph the current statements. If two Excel files disagree, pick a source of truth in writing before you import anything.
Week 1 — clients and KYC
Deduplicate on Ghana Card and phone. Do not import obvious ghosts. You can always add a client; you cannot easily unmerge a mess after disbursement.
Week 2 — live schedules only
Import remaining instalments, not 2018 receipts. Back-solve if you must. Train cashiers on posting with dummy receipts on day one of parallel. Credit origination training comes after cash is safe.
Parallel for two cycles
Compare PAR, cash, and a sample of twenty statements. When they match, freeze Excel. Announce a date after which the spreadsheet is archive-only. If someone keeps a shadow file, you will never know which book is true.
History later, if at all
Closed loans can wait. Examiners care about the live book and the trail going forward. Importing a decade of noisy Excel is how go-lives slip three months.
