ERP Migration & SelectionFree worksheetPDF, Excel, Word exportJob Shop Machining

Job Shop ERP Data Migration Worksheet

Data migration is where most small shop ERP projects lose time. This worksheet lists each data object, how many records it has, how clean it is and who owns it, and estimates the effort so the plan is based on something more than a guess.

Download:
Data objectSourceRecord countBase hours per 1,000 recordsData clean percentEstimated effort hoursOwnerStatus
6.63
3.024
58.24
29.16
34.8
2.31
28.5
3.276
Total records
12,970
Total effort hours
165.94
Average data cleanliness
70.6%
Weeks for one person at 20 hrs per week
8.3 weeks
Labor cost at 70 per hour
$11,616

The effort formula adds up to double the base hours when data is fully dirty (0 percent clean) and the base hours when it is fully clean. Base hours per 1,000 records are planning assumptions: replace them with figures from your own trial loads.

Review your migration plan

Optional: send us the worksheet and we will comment on sequencing and risk areas.

Optional. The tool and every export work without this. No spam.

How to use the migration worksheet

List every data object you must carry into the new ERP: customers, vendors, parts, bills of material, routings, work centers, open orders, inventory and open balances. For each, enter the source system, record count, a base hours per 1,000 records assumption and an honest percent clean. The worksheet calculates effort hours, then totals them and converts the total into weeks for one person. Update status as each object moves from extraction through cleaning, test load and validation.

How the effort estimate works

Effort equals records divided by 1,000, times base hours per 1,000, times a dirtiness factor of two minus the clean fraction. Perfectly clean data costs the base hours. Fully dirty data costs double. The factor is a planning simplification, not a measured law, so calibrate it: load a sample of each object into a test environment, time the cleanup, and replace the assumptions. Complex objects such as bills of material and routings carry higher base hours than customers or vendors.

  • Parts, BOMs and routings usually dominate effort
  • Inventory needs a physical count before load
  • Open AR and AP must tie to the general ledger
  • Decide what not to convert and archive it instead

Sequencing the work

Load in dependency order: customers, vendors, work centers and units first, then parts, then bills of material and routings, then open orders and inventory, and finally balances. A part cannot be loaded cleanly if its units of measure or vendors do not exist yet. Plan at least two trial loads before the final one, and reconcile each against source counts and values.

Common mistakes

Shops overrate how clean their data is, forget to assign owners, migrate years of inactive parts, and discover missing fields during the final load. Take a hard look at the part master first. It is almost always the largest and messiest object.

Frequently Asked Questions

How many trial loads should I plan?

Plan at least two full trial loads plus the final cutover load. The first finds structural problems like missing fields and mapping errors, the second proves fixes, and the final should be a rehearsed repeat with no surprises.

Should I migrate inactive parts?

Usually not. Migrate parts used in the last two or three years or on open orders, and archive the rest. Fewer, cleaner records make the new system faster and the project cheaper.

What does clean percent mean?

It is your estimate of how many records in that object could load without fixes: no duplicates, no missing required fields, correct units. Sample a hundred records and count the ones needing edits to get an honest number.

Can this worksheet be used for any ERP?

Yes. The objects and sequencing are generic to job shop and discrete manufacturing ERPs, whether you are moving between JobBOSS, E2, Global Shop Solutions, ProShop, Fulcrum or something else.

If you want a second set of eyes on migration scope, or tooling to profile and compare source and target data on your own network, we can help. Data never needs to leave your environment.