ERP Migration & SelectionFree calculatorPDF, Excel, Word exportAcumaticaSMB Manufacturing

Acumatica Data Load Window Calculator

Cutover weekends fail when nobody timed the load. This calculator turns row counts and a measured throughput into elapsed hours, including the reloading of rejected records and your validation time, and tells you whether the window is big enough.

Your numbers

Master records (customers, vendors, items)

records

Open transactions (orders, invoices, balances)

records

BOM and routing lines

lines

Load throughput per stream

records per hour

Measure it in a test load; API or import-scenario throughput varies with validation, workflows and customisations.

Parallel load streams

3 streams

Rejected records on first pass

4 %

Cutover window available

hours

Validation and sign-off time

hours

Your results

Total elapsed time with validation

19 hrs

Slack in the window (negative means overrun)

17 hrs

Total records

77,000

First pass load time

10 hrs

Rework load time

1 hrs

Rejected records are reloaded with a 1.5x factor for investigation and correction overhead.

Throughput is the key input and must come from a rehearsal load in a copy of your own environment.

Free, optional. Reopen it on any device.

Save results:

Elapsed hours by number of parallel streams

Sweeps Parallel load streams from 1 to 10, holding your other inputs fixed.

Migration rehearsal review

Share your rehearsal results and we will help identify the throughput bottleneck.

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

Was this tool useful?

How to use it

Count master records, open transactions and BOM and routing lines in the extract. Take throughput from a rehearsal load, not from a guess, because validations, workflows and custom code change it dramatically. Add the number of parallel streams you can run without lock contention, the expected error rate and your validation time. Compare the elapsed total with the window you can negotiate with the business.

The model

First pass time is total records divided by throughput x streams. Rework adds the rejected share loaded again at 1.5 times cost, and validation is added as a flat period. The chart shows diminishing returns as streams increase; in practice dependencies such as items before BOMs limit parallelism, and the database will eventually saturate.

Load in dependency order: master data, then open items.
Run at least two full rehearsals and update the throughput.
Reduce volume by archiving closed history instead of migrating it.

Common mistakes

Teams assume the rehearsal error rate will hold on production data, forget cutover-day freeze and extract time, and schedule no buffer. Keep a rollback decision point in the runbook. If the model shows negative slack, cut scope or add rehearsal-proven streams, not hope.

Frequently Asked Questions

Where do I get throughput?

From a rehearsal load in a sandbox with production-like data and customisations. Vendor numbers on empty systems will be far higher than your real figure.