How to move from spreadsheets to a CRM without losing track

1. What should you move into the CRM?

Begin with the work the CRM needs to support, not every spreadsheet the business has ever created. List your current enquiries, active customers, open opportunities and outstanding actions. Decide whether past customers, old quotes and archived projects are needed for daily work or should remain in a separate, controlled archive.

Keep the first version focused on one journey, such as enquiry to accepted quote. Your accounting package can remain the source of invoice and payment information. Importing everything creates more fields to maintain and more opportunities to confuse records that serve different purposes.

• Identify the spreadsheet or system that is authoritative for each type of information
• Choose an owner who can resolve conflicting records and approve the migration
• Define your sales stages in plain language, including what qualifies a lead to move forward
• List the reports and reminders the team must be able to use on the first day

2. How do you clean customer data before importing it?

Make a dated, access-controlled backup before changing anything. Work on a copy, not the only version of the customer list. Standardise obvious inconsistencies such as company names, telephone formats and sales-stage labels, then flag uncertain values for a person to review.

Do not assume two rows with the same name are duplicates. A shared office email address might represent several people, while one customer may have multiple properties or projects. Establish what identifies a contact, an organisation and an opportunity before merging them.

Check for hidden tabs, filtered-out rows and formulas. Exporting a sheet can lose context or change formatting. Telephone numbers should normally be treated as text so leading zeroes survive, and ambiguous dates such as 04/05/2026 need an explicit day-month-year interpretation.

3. What does mapping spreadsheet columns mean?

Field mapping means deciding where each spreadsheet column belongs in the CRM. A column called Client may need to become an organisation name, a contact name or two separate fields. Next action should usually become a task with an owner and due date, not a note nobody is prompted to read.

Create a simple mapping document that records the source column, destination field, expected format and any transformation. Keep the old record identifier where possible. It makes reconciliation easier and helps prevent duplicate creation if an import must be repeated.

• Customer name: separate the organisation from the person where relevant
• Quote status: translate old labels into an agreed set of sales stages
• Follow-up date: create a dated task assigned to the responsible team member
• Notes and files: confirm whether these need a separate import or manual transfer
• Marketing preferences: preserve the original status and supporting records rather than treating every contact as subscribed

4. How should you test a CRM import?

Use a small, representative batch before importing the full list. Include an ordinary enquiry, a customer with several contacts, a missing email address, an unusual telephone format and an opportunity with a pending task. A test containing only perfect records will not expose the problems that matter.

Disable automated emails, reminders and downstream integrations during the import unless you have deliberately configured and tested their behaviour. Otherwise, creating old opportunities can accidentally send welcome messages or trigger tasks that were never intended.

Check more than whether the import says successful. Open individual records, search for them, inspect linked opportunities and confirm that another authorised team member can see the right information. Also check that a user with restricted access cannot see records they should not.

A worked example: moving a service business into a CRM

Consider a fictional five-person service business with 800 spreadsheet rows. These are not necessarily 800 customers: some rows represent repeat jobs, some are new enquiries and others are duplicate entries. The team first separates contacts from opportunities and gives each open opportunity a named owner.

They test 25 varied records, including a customer with two jobs and an enquiry without an email address. The test reveals that follow-up dates were imported as text rather than tasks. They correct the mapping before moving the remaining records.

These figures are illustrative, not an AVS client result or a recommended batch size. The useful lesson is that a small test can uncover a process error that a clean-looking contact list would hide.

5. How do you check nothing important was lost?

Compare the source and destination using the same definitions. If one spreadsheet row becomes a contact plus an opportunity, a simple row-count comparison will mislead you. Reconcile contacts, organisations, open opportunities and outstanding tasks separately.

Review rejected rows and import warnings rather than accepting a partial success. Spot-check valuable or time-sensitive records, compare pipeline totals where relevant, and confirm that files and notes still belong to the correct customer. Record which exceptions remain unresolved and who will fix them.

• Account for every source record as imported, intentionally excluded or awaiting correction
• Check that opportunity values, currencies and expected dates have not changed
• Confirm that every active enquiry has an owner and a clear next action
• Test exports and the recovery process before relying on the CRM as the only working system

6. How do you switch over without creating two versions of the truth?

Choose a clear cutover time and tell the team where new work must be recorded from that point onwards. Pause edits to the old sheets while the final import runs, or maintain a controlled list of changes that must be transferred afterwards.

Avoid leaving both systems open for unrestricted editing indefinitely. That creates disagreements the CRM was meant to solve. Keep the old files read-only for reconciliation and recovery, with access and retention managed appropriately.

Agree a rollback plan before the switch. Define what failure would justify returning to the previous process and how you would preserve records created since cutover. A backup alone is not enough if nobody knows how to resume work safely.

7. What should your team learn first?

Train people on their actual daily tasks: record an enquiry, find a conversation, assign a follow-up and update an opportunity. Use realistic examples rather than a tour of every feature. Explain which fields are required and why they matter to another person's work.

During the first week, review a handful of live records together. Look for missing owners, unclear stages and overdue tasks. Treat these as opportunities to improve the process, not evidence that staff need more mandatory fields. A simpler system used consistently is more useful than a sophisticated one people avoid.

What about UK GDPR and customer information?

Moving information into a CRM does not automatically make it compliant with UK data protection law. Review what personal data you need, the purpose and lawful basis for using it, who can access it and how long it should be retained. Check supplier terms, security arrangements and any relevant international transfers.

Do not upload sensitive customer records into public AI tools to clean or summarise them without an appropriate assessment and safeguards. Preserve marketing preferences during migration and remember that storing a contact is not the same as having permission to send marketing. Use the Information Commissioner's Office guidance or qualified advice for your circumstances.

Do you need a custom CRM to leave spreadsheets?

No. A ready-made CRM is often the right answer for standard contact management and sales follow-up. A custom system deserves consideration when essential workflows, permissions or connections cannot be handled cleanly by an established product.

Whichever route you choose, the migration principles stay the same: define the process, preserve the source, test the mapping and verify the result. The goal is not simply to stop using spreadsheets. It is to make customer information dependable enough that the whole team knows what to do next.