All posts
Guides7 min readUpdated November 17, 2026By

From Excel to a CRM: moving over without losing data or your team

The short answer: start with a data model everyone can read, show the team a clickable prototype before any code, move the data with a mapping someone signs off in advance, and run the new system next to the spreadsheet for about 2 weeks before you freeze the sheet. Migrations that fail usually fail in one of two places: wrong data on day one, or a team that drifts back to the spreadsheet.

Everything here comes from systems we've built for clients moving off spreadsheets, and from internal systems we built for ourselves.

Last updated: November 17, 2026

When is it time to replace spreadsheets with a CRM?

It's time when the spreadsheet stops being the place where the truth lives. You'll know the signs: three copies of the same file with names like "final_v2", an "owner" column that gets updated once a month, and questions like "what happened with this customer?" whose answer sits in someone's WhatsApp.

Managing customers in Excel works fine until the number of people writing to the sheet passes the number of people who understand it. From there, every structural change breaks someone else's formula, and there's no way to tell who changed what.

If you're still deciding between building a system and buying a ready-made tool, start with our guide to building or buying a CRM. This one picks up from the moment you've decided to move.

Where does an Excel to CRM migration start?

It starts with the data model, drawn as a map of subject areas plus one detailed view per area. One giant diagram with every table on it looks impressive, and the people on the client's side can't read it.

One system we designed came to 27 tables. We showed it as a single map of 6 areas (contacts, organisations, activity, tasks and so on), with its own detailed view for each. When we talked about contacts, everyone was looking at a handful of tables they knew from their own work, and the conversation stayed on the work.

Your spreadsheet is the best raw material for the model. Every column is a field, every tab is probably a table, and every cell where someone typed "see note" is a field that's missing. We go through the sheet with the people who use it daily and ask, for each column, what it means and who fills it in.

Why build a prototype before writing code?

People react to screens far better than to a spec. A clickable prototype shows the team what their day will look like, and the reactions it gets save weeks of fixes after launch.

For one of our internal systems we built a 28-screen prototype before the first line of code. People clicked, got stuck and asked. A question like "where do I see my customers who haven't been updated this week?" is a ten-minute fix when it comes up in front of a screen. After launch it can be a week.

It's enough for the prototype to look like the system, with data that looks like yours, and to let people click through it. The logic comes later.

How do you migrate data from Excel without losing it?

Map every column to a field, merge the duplicates, show the mapping before anything loads, and produce a report of what was merged, what was skipped and why. A CRM that opens on day one with wrong data never recovers, because the team concludes it can't be trusted.

Duplicates are most of the work. The same company spelled 3 ways ("Cohen & Sons", "Cohen and Sons Ltd", "Cohen Sons"), the same contact twice with the phone number in a different format. Merging needs rules that someone on the team approves, plus a list of every borderline case for a person to decide.

StageWhat happensWhat you check
Column mappingEvery column in the sheet gets a field in the system, or is marked as not movingNo column disappeared quietly
CleanupPhones, dates, city names and statuses made consistentValues that didn't fit any format
Duplicate mergeDuplicate companies and contacts combined by approved rulesThe list of borderline cases, each with a person's decision
Preview20 real records shown as they'll look in the systemThe people who use the sheet recognise their own customers
LoadThe data goes inRecord counts against the sheet, after merging
Migration reportWhat was merged, what was skipped, and whyEvery skipped row has a reason

The report at the end is what builds trust. When someone asks "where's that customer?", you open the report and see they were merged with another record, and which one.

How long should you run the CRM alongside the spreadsheet?

About 2 weeks, then you freeze the spreadsheet as read-only. Shorter, and you haven't caught the things that happen once a week. Longer, and the team gets used to updating in two places, or worse, only in the sheet.

During those 2 weeks every new update goes into the system. The sheet stays open for checking and comparing. At the end it turns read-only and goes to the archive, so everything that was in it is kept.

What keeps the team in the CRM instead of the spreadsheet?

A system that behaves the way people expect in the small places. People abandon a new system over ten irritating moments a day long before they abandon it over a missing feature.

These lessons come from client apps we've shipped, and each one is something a client saw on a real screen:

  • A failed load shows an error. One app showed "no data" while the database held 837 rows. A user who sees an empty list assumes the data was deleted, and goes back to Excel that same day.
  • Dates in Israeli order and 24-hour time. 27.10 and 14:30, the way everyone here writes them. It sounds small, and a US-order date confuses everyone who looks at it.
  • Phone numbers and other numbers isolated inside right-to-left text. Without that, a phone number inside a Hebrew sentence flips and reads wrong.
  • On a phone, every table becomes a card per row. Field staff won't scroll a table sideways.
  • Forms say which field is wrong, under that field. A general "check the form" message leaves the user guessing.
  • Editing a record saves only the fields the form shows. In one system, a form that lacked an approved-amount field wiped that amount on every save. That's the kind of bug that makes people keep a copy in Excel "just in case".
  • A feedback button inside the app. Staff send a request with a screenshot from inside the system, instead of a WhatsApp message that gets lost.

That last one matters more than it looks. In the first weeks you'll get dozens of small requests. When they land in one place with a screenshot, most can be fixed quickly, and the team sees someone's listening.

Who on your team should own the move?

One person who uses the spreadsheet every day and has the authority to say "this is how we do it now". A manager who signs off from a distance can't answer the questions that come up, and a junior who answers them can't make them stick.

That person decides what each column means, approves the merge rules, looks at the preview and recognises the customers in it, and is the first to stop opening the sheet. In practice they spend a few hours a week with us during the build and more during the parallel run. When the move goes well, it's usually because this person was in the room for every decision about the data.

They also become the first stop for questions from the rest of the team. Someone who knew the old sheet inside out and now knows the system is worth more in the first month than any training session we could run.

How long does it take to move from Excel to a CRM?

The first usable screen usually arrives within about the first 2 weeks of the build. The full move depends on how many areas the system covers and the state of the data, and the parallel run adds about 2 more weeks.

We'd rather ship one area that works (contacts and tasks, say) and let the team use it before building the next. That way the data moves in pieces and the team learns the system gradually. We lay out the whole project, stage by stage, in what an automation project looks like.

If you want the system to do some of the work itself (summarising calls, updating statuses, reminders), see our guide to AI agents and CRM. And if the spreadsheet holds sensitive data that only some people should see, read how to decide who sees what in a CRM before you load anything. When you're ready to build, that's our custom CRM development work.

FAQ

How long does an Excel to CRM migration take?

The first usable screen usually arrives within about 2 weeks, followed by about 2 weeks of running alongside the spreadsheet. A system that covers several areas goes live in pieces, one area at a time.

Can we keep using Excel for reports?

Yes. Exporting to Excel from the system is fine, and most teams keep doing analysis in a spreadsheet. What changes is the direction: the system is the source, and the spreadsheet is a read-only copy. Updates go back only through the system.

What about the history in our inboxes and WhatsApp?

Move what drives decisions: the latest status, the date of last contact, and a short note on what was agreed. Most history isn't worth moving by hand. If some of it is worth keeping in full, you can connect the mailbox to the system so every new email is logged from now on.

Who cleans the data?

Both sides. We write the cleanup and merge rules and run them, and someone on your team who knows the customers approves the rules and decides the borderline cases. Without that person, the system will merge two different companies with similar names.

What if people don't adopt it?

Usually it points to one of a few things: the data in it is wrong, something basic about it is annoying, or the spreadsheet is still open for editing. Check each, listen to what comes in through the feedback button, and freeze the sheet when the parallel run ends.

Next step

If your business runs on a spreadsheet and you want to know what it would take to move it into a system, talk to us. We'll go through the sheet with you and tell you what moves as it is and what needs cleaning first.

More from the blog

We use cookies to make your experience smoother, faster, and a little more fun. If you stick around, we'll take that as a thumbs-up. Learn more