From spreadsheets to an ATS: a migration checklist

The migration itself is a day of work. The decisions that make it succeed or fail are made before anyone exports anything.

Most companies run hiring on a spreadsheet for longer than they should, and the reason is rarely the software cost. It is that migration feels like a project, and nobody has time for a project.

It is not a project. The export takes an afternoon. What takes thought is deciding what to bring, and that decision is where migrations go wrong.

Before you export anything

1. Decide what you are actually migrating

Three categories, and only one of them is straightforward:

  • Active candidates in live processes. Migrate all of them. This is non-negotiable — losing someone mid-process is the one unrecoverable error.
  • Open roles. Recreate rather than import. Roles carry stage definitions and criteria that a spreadsheet does not contain, and recreating them is where you actually design the process.
  • Historical candidates. The genuine decision, and the answer is usually less than you think.

2. Answer the retention question first

This is the step that gets skipped and later hurts. Importing three years of historical CVs into a new system means importing three years of retention exposure — and now it is in a system that applies policies, so the problem becomes visible.

Ask two things before importing any historical record: do we have a basis to hold this, and would we contact this person again? For most spreadsheet rows over eighteen months old, the honest answers are no and no. Migration is the cleanest opportunity you will get to not carry that forward. The framework is here.

3. Design the stages before importing

A spreadsheet usually has a status column with values that accumulated organically: "sent", "waiting", "maybe", "chase". Do not reproduce those. Define the stages you want — typically four to six — and map the old values onto them. Anything that does not map cleanly was probably not a stage.

4. Agree the evaluation criteria

Migration is the moment to introduce structure, because everyone is already changing how they work. Four to six criteria per role family, with anchors: scorecards. Introducing this later, as a separate change, meets far more resistance.

The migration itself

5. Clean the export

One row per candidate per role. Consistent email addresses — these are your deduplication key. Dates in one format. Names split into first and last if the system expects it.

6. Import in two passes

Active candidates first, verified by hand. Then historical, if you decided to bring any. Do not import everything and sort it out afterwards; a mixed import that goes wrong is difficult to unpick.

7. Check duplicates before you finish

Spreadsheet hiring produces duplicates reliably — the same person applied twice, or was added by two people. Merge before recruiters start working, not after.

8. Migrate the notes, or accept losing them

Interview notes usually live in a comment column or a separate document, and they are the part of the migration most likely to be dropped for convenience. Decide deliberately. If a hiring decision is questioned later, the notes are the record.

The first two weeks

9. Stop using the spreadsheet completely

Running both is the most common failure and the most predictable. Within a week the spreadsheet is authoritative again, because it is familiar, and the migration has failed without anyone deciding to abandon it. Pick a date, archive the file read-only, and hold the line.

10. Move one requisition first

One live role, one recruiter, end to end. Every problem you will hit appears in that role, at one twentieth of the cost of hitting it across all of them.

11. Set up the two things people notice

Calendar integration and application confirmation emails. These deliver visible relief in week one and buy goodwill for the rest of the change. See which integrations matter.

12. Book a thirty-day review

Stages that are not being used, fields nobody fills in, steps people work around. The first configuration is always wrong somewhere, and fixing it at thirty days is easy. Fixing it at six months means unpicking habits.

What to expect

Reporting is the change people notice most. Questions that were unanswerable — where candidates drop out, which source produces hires, how long each stage takes — become routine. Most teams are surprised by at least one of the answers, usually the amount of elapsed time spent waiting rather than working.

The second change is that hiring stops depending on one person memory. That is the durable benefit, and it is the one that is invisible until the person who held it all is on holiday.

For what the first roles look like in practice, see pipelines; for cost, pricing.

Get hiring insights in your inbox

One email a week — the best playbooks, benchmarks and product updates. No spam.

More from Resources

Everything you need to hire better — in one place.