Moving Spreadsheet Leads Into a CRM: A PH Checklist

The checklist: keep a source copy, agree the fields, settle duplicate rules, map ownership, test a small batch and reconcile the results. Keep automatic customer messages out of the first import until you have reviewed how the selected CRM handles them.

This guide covers an operational move after you have chosen a CRM. The small-business CRM page covers the wider question of whether a CRM fits your team.

Name the person who can approve the mapping and resolve unclear records. That may be the owner or a sales manager who understands the current sheet.

Define the scope. Are you moving open enquiries only, or customers, closed deals and history too? A contact list and a history of sales activity are different datasets. Write down which records and attachments belong in this move, and which will remain in an archive.

Keep the process narrow enough to check. If your sheet also tracks inventory, delivery and accounting, separate those needs before trying to put every column in a sales pipeline. The CRM automation page explains how capture and follow-up can connect once the records and stages make sense.

Save an untouched export and make a working copy. Give both clear names and record which sheet and date each came from. Restrict access to the people doing the work.

Note the number of source rows and any filters used. A filtered export can quietly leave out records, while an empty row can inflate a count. Record which rows you exclude and why.

Zoho's current import guidance recommends changing a copy of the file and starting with a small subset before a larger import. This is a useful principle even if you choose another CRM. Follow that product's current import rules for the actual procedure. Zoho CRM import guidance.

Create a mapping sheet before uploading anything. Give each source column a destination field, format and rule for blanks.

Do not create a new field merely because a column exists. Ask how staff will use it. Keep information that supports the defined process, and avoid moving unrelated personal details into a wider-access system.

Agree what a blank means. It might mean unknown, not applicable or deliberately cleared. Those meanings are not interchangeable when an import updates an existing record.

Check the exported file itself. A cell that looks right in a workbook may change when another tool reads it.

Store phone numbers as text while preparing the file. Microsoft explains that Excel can remove leading zeros and recommends applying Text format before typing or pasting values that need those zeros. Formatting a column later does not necessarily restore data already lost. Microsoft's leading-zero guidance.

Keep the original phone value beside any cleaned value until you have checked it. Confirm the country code rather than adding one to every number by assumption.

Choose an unambiguous date format that the destination CRM accepts. Test dates where the day and month could be confused. Also confirm the account time zone for any follow-up times. HubSpot's import documentation specifies formats for dates, phone values and other property types; use the equivalent rules for your chosen system. HubSpot import-file requirements.

Define the match rule before the import. A repeated email, phone number or name can mean the same person, a shared business contact or a different enquiry. Do not merge records solely because two names look similar.

Decide whether the import should create records, update existing records or do both. Record how it handles blank fields and conflicting values. Review uncertain matches with the person who owns the source data.

In HubSpot, updates and deduplication rely on the appropriate unique identifiers, such as Record ID or email for contacts. Other systems and record types can use different rules. A spreadsheet's row number is not automatically a reliable customer identifier. HubSpot's unique-identifier requirements.

Keep a log of merges and exclusions. That gives the team a way to answer why a source row did not become a separate CRM record.

Choose records that cover the awkward cases, not just the clean ones. Include a missing phone, a duplicate, a long note, an unfamiliar character, a closed enquiry and a record with an absent owner.

Review the chosen CRM's import and workflow settings before the test. Establish whether an imported record can trigger messages, tasks, assignments or integrations. Use a test environment where available, and agree how you will identify and remove test records through the product's supported process.

Inspect the result as a staff member would. Can that person find the enquiry, understand the stage and see the next action? Check the original values against the imported record. A successful upload message does not answer those questions.

Account for each source row: imported, merged, excluded or failed. The totals should tell a complete story. Then check a sample of the actual fields, owners and dates.

Read the error report and keep it with the mapping record. Decide which failed rows need correction and which belong outside the agreed scope. When you retry, use the CRM's match rules so that the retry does not create another set of records.

Decide where staff should work after approval. Keeping two editable sources without a clear rule can bring back conflicting notes and missed follow-ups. Retain the old copy as a controlled reference if the business needs it.

Give the business the mapping, import results, exception log and staff instructions. Identify who can change fields, who maintains lead-source connections and who checks unassigned records.

Confirm that the business has a lawful basis for holding and using the contact data. The NPC's AI guidance also applies when AI systems process personal data during development, testing or deployment. Review that processing separately if AI will read or summarise CRM records. NPC Advisory No. 2024-04.

The first useful result is a customer list your team can explain and work from. Add wider automation only after the records and responsibilities pass review.

Suppose a source sheet contains eight enquiry rows: five clean enquiries, one repeat of an already listed enquiry, one genuine second enquiry from an existing customer and one row with no usable matching information. In this teaching example, import six distinct enquiries, account for one repeated enquiry and hold one unresolved row for review. Six plus one plus one reconciles to eight source rows.

Do not assume six enquiries means six contacts. The genuine second enquiry should reuse the correct contact while retaining its separate request. The actual result depends on the selected CRM's data model and matching rules. This is a reasoning exercise, not a performed vendor import.

Try the returning-customer case in the handover teaching model. Then use the software comparison scorecard to design an authorised trial in your chosen CRM.

Discuss your CRM import and handover with AI Automation Services.