
01
The short answer
Do not upload your entire workbook to the CRM in one click. Keep an untouched copy, decide how to identify each account and contact, import a small test batch, then compare the results before continuing. This lets you spot duplicates, incorrectly mapped fields and missing relationships without putting your original files at risk.
An import does not automatically transfer everything a spreadsheet contains. Comments, attachments, formulas, tabs, older versions and relationships between sheets each need to be checked. Keep Excel as your backup copy until you have verified the data and its relationships in the CRM.
02
What should you do before working on the Excel file?
Start by listing the files and the people who use them. A client list may be spread across several workbooks, tabs or copies sent by email. Decide who will resolve differences when the same information varies from one file to another. For example, a contact’s address may have been updated in one tab but not the others. Do not merge these values automatically without a rule.
- 1List the workbooks, tabs, owners and most recent update dates.
- 2Make a copy of each original file and keep it read-only. Work from preparation copies.
- 3Note the columns in use, important formulas, dropdown lists, and any comments or attachments to keep.
- 4Decide which data to carry over, which to archive and which is obsolete.
- 5Assign one person to make corrections and another to verify the import.
Also prepare a short column dictionary: does “Office Phone” mean the primary phone number or a secondary field? Should “Active Client” become a status, a stage or a tag? An Excel field name does not always show how the CRM will use it. In HubSpot, for example, required fields and some identifiers depend on the type of record being imported. In Dynamics 365, data must also be organized according to the type of record being targeted. Check the requirements of your chosen CRM before building the final file.
03
How do you know if the spreadsheet is ready to import?
A clean file is not necessarily a safe file to import. Each row must also represent the right type of record, the columns must match the CRM fields, and an identifier must let you recognize a record that already exists. If several rows have the same email, identifier or variations of the same name, the CRM cannot reliably tell whether they are duplicates or two different people.
- One row, one record
- Separate contacts, companies, opportunities and activities according to the CRM’s objects. Avoid title rows, subtotals and merged cells.
- A stable identifier
- Keep a customer number or source identifier if one exists. Otherwise, decide on the best available matching key. Email can help identify a contact, but it may be missing or change.
- Consistent values
- Standardize dates, phone numbers, countries, statuses and list values. Also check for extra spaces and incorrectly formatted email addresses.
- No hidden formulas to lose
- Identify calculated columns and decide whether the CRM should receive the result, recreate the calculation or leave that field out.
- Relationships are explicit
- Each contact associated with a company must be linkable to it by an identifier or reference value that does not point to multiple companies.
- Information to keep is identified
- Identify comments, notes, activity dates and attachments separately. A column containing a file name does not mean the file will be imported.
Avoid using only a person’s name as a key: two people can have the same name, and one person can appear under different versions of it. The documentation for HubSpot, Zoho CRM, Salesforce and Dynamics 365 describes different ways to match records, such as a record ID, a unique field or, in some cases, an email address. Available rules and how they work depend on the CRM and its configuration. Validate them before importing, especially if some records already exist in the system.
04
How do you preserve links between companies, contacts and history?
A spreadsheet can show a relationship without actually representing it. For example, a tab may repeat a company name on every contact row; the CRM, however, often expects the contact record to be associated with a specific company record. A name written in different ways—“ABC Group,” “ABC Inc.” or “ABC”—can prevent the association or create several companies.
- 1Import primary records, such as companies, first, or prepare identifiers in the same file that will let you recognize them.
- 2Then import contacts with a column that clearly identifies their company using a stable identifier or an unambiguous matching value.
- 3Check a few companies with several contacts, as well as special cases: contacts linked to multiple companies, subsidiaries or companies without a web domain.
- 4Import activities and notes separately if needed, keeping a reference to the record each one belongs to.
- 5Open records in the CRM and confirm that links and history appear in the right place.
Notes and activities are not always equivalent to comments in an Excel cell. A “Last Call” field does not, by itself, recreate every conversation. If you need to keep comments, dates or attachments, inventory them and test their transfer separately. Some import features support activities or notes, but required fields, possible associations and transfer methods vary by product. For an attachment, check not only that the file is present, but also that it is linked to the right record.
05
How do you prevent duplicates and overwrites?
Before importing, choose the mode that matches your goal: create new records, update existing ones or do both. If you only want to change two fields, do not reimport every column from your old export. Keep the record identifier and include only the fields you want to change. A file prepared to create new records is not necessarily safe to use to update existing ones.
- Use the right identifier
- For records already in the CRM, an internal ID exported from that CRM is often more precise than a name. An external identifier can also work if it is unique and kept unchanged.
- Find duplicates in the file
- Look for repeated identifiers and review similar emails or names. Decide which row is correct before importing, rather than letting the result depend on row order.
- Check blank cells
- Behaviour varies by software and by the option selected. HubSpot notes that blank cells are ignored during import; in Zoho CRM, options are also available to prevent some existing values from being replaced by blank cells. Confirm the setting that applies to your import.
- Check duplicate rules
- Depending on the CRM and its configuration, matching rules may flag or block records. They do not replace a reliable key or a review of ambiguous cases.
- Keep a record of the batch
- Record the file, date, import mode, mapped fields and row count. A temporary tag or source field can make checking easier, if your configuration allows it.
Be especially careful when data has been exported from the CRM, edited in Excel and then imported again. In that case, keep the CRM identifier unchanged and confirm whether the import will update existing records. Unnecessary columns or old values can bring outdated information back in. If you are unsure of the outcome, stop before the full import and have the CRM administrator review the mapping.
06
How do you test the import before retiring Excel?
A test is not just about confirming that the software accepts the file. You need to confirm that the right records were created or updated, relationships work, notes are in the right place and important values were not changed by mistake. If your CRM offers a test environment that suits your subscription and configuration, use it. Otherwise, prepare a small, representative batch and check the possible effects before importing it into your working environment.
- 1Choose representative rows: an existing contact, a new contact, a possible duplicate, a company with several contacts and a record with a note or attachment.
- 2Prepare the mapping column by column. Check required fields, date formats and whether the import will create or update records.
- 3Review any automations, notifications and rules that might run during the import. Decide with the administrator how to test them without unwanted effects.
- 4Run the test batch and review the import report: rows that succeeded, were skipped, failed or were only partially processed.
- 5Compare CRM records with the source rows, then correct the file or mapping before the next batch.
Do not check only the total number of contacts. Also compare totals by record type and review a sample of important values: email, phone, status, owner, associated company, date and most recent activity. For errors, correct only the affected rows and keep the report. Dynamics 365, for example, documents how to review successful, failed or partial imports; other CRMs offer their own import history. Use the reports and screens provided by your software rather than assuming each row was handled as expected.
Do not delete Excel as soon as the first import finishes. Validate one batch, then the next; keep the originals and an export of the CRM after each important stage. Once checks are complete, you can make the old files read-only and add a note stating which system is now the source of truth and from what date. This transition prevents two versions from continuing to be edited in parallel.
07
Do you have to choose a standard CRM?
Moving from Excel to a CRM is also an opportunity to check whether the tool fits the way you work. If your needs are standard—tracking contacts, companies, sales and tasks—an existing CRM can structure the data and reduce manual work. You can then add AI to find information, summarize exchanges or automate certain tasks, depending on the features actually available and your organization’s rules.
If your spreadsheets mainly manage a very specific process, the right choice may not be to recreate every tab in a CRM. It may make sense to keep the tool and add automation or AI, or to explore a custom application connected to your data. Language models can reduce the work involved in designing some applications, but they do not automatically solve data quality, security or maintenance. At TECHMA, we help compare these paths without assuming that replacement is always the best option.
To get started, choose one process to move, define what the migration should improve and inventory the data that supports it. A discovery call with TECHMA can help clarify the approach and options. The goal is not to migrate as many rows as possible: it is to make information reliable and usable without losing the relationships or history your team needs.
Sources
The pages the facts in this guide come from. Each link was checked when the guide was written.

