Profile before you promise
Start by measuring what you have: how many customers, items, open transactions, and years of history; how many duplicates; which fields are blank or inconsistent; and which custom tables carry real business logic. Profiling gives you facts to size the effort and expose risks early. It also reveals who understands the data, which is often a person near retirement or a long-term consultant, and their knowledge needs capturing before they leave.
- Count records by type and by age.
- Identify duplicates and orphaned records.
- List custom fields and tables with their business meaning.
Decide what stays, what moves and what is archived
Not every record needs to move. Active master data and open transactions move; closed history older than a chosen horizon is archived in a read-only form for audit and customer inquiries. Defining the horizon with finance, legal, and operations avoids debates later. Keep an index so someone can find an archived transaction when a customer or auditor asks.
Implementation and integration services →Cleanse and transform with documented rules
Cleansing is the work of correcting and standardizing, and transformation is converting from the old structure to the new. Write every rule down: how duplicates are resolved, how addresses are standardized, how codes map to NetSuite lists, and how units of measure convert. Automated rules are repeatable across trial loads, while manual fixes are not, so push as much as possible into scripts that can be rerun.
Keep a log of exceptions and decisions. When someone asks later why a customer was merged, you can answer.
NetSuite integration services →Trial loads and reconciliation
Plan at least two or three full trial loads into a test account. After each, reconcile: totals by account, subledger balances against the ledger, inventory quantities and values by location, and counts of customers, vendors, and items. Differences are investigated and fixed in the source, the rules, or the mapping, and then the load is repeated. By the final rehearsal, the process should be boring.
Cutover strategy: one step or phases
A single cutover moves everything at once, which is simpler to manage but riskier. A phased approach moves by entity, region, or module, which reduces risk but means running two systems and managing interfaces between them for a time. Choose based on how interconnected your operations are and how much temporary complexity your team can handle. Whichever you choose, define the fallback criteria before you start.
Our NetSuite ERP practice →Decisions to settle before you start
- History horizon. Set the age beyond which records are archived and the format of the archive.
- Data owners. Name the person who approves each data set before it loads.
- Freeze period. Agree when changes in the old system stop, and how late transactions are handled.
- Fallback. Define the conditions under which you would delay go-live or revert.
A realistic first 90 days
- Days 1 to 30. Profile the data, capture undocumented rules from the people who know them, and define the archive horizon.
- Days 31 to 60. Build the transformation scripts and run the first trial load with reconciliation.
- Days 61 to 90. Run a second trial, resolve exceptions, and plan the cutover weekend with roles and checkpoints.
Pitfalls to avoid
- Discovering data problems in the last month. Profile on day one, not when the load fails.
- Manual cleansing. Hand edits cannot be repeated for the next trial load. Script them.
- Losing the tribal knowledge. Interview long-time users while they are still available and record what they know.
- No archive access. If nobody can find old transactions, the first customer dispute becomes a crisis.

