CRM setup is the work of configuring a new CRM to match how your business actually operates: mapping your process, designing pipeline stages and custom fields, connecting your integrations, and preparing your data before any of it gets imported. It happens after you’ve chosen a platform and before your team starts using it.
Almost every guide on this subject is about migration, which is the data half of the job. The configuration half decides whether anyone uses the thing. In our experience the businesses that abandon a CRM within a year rarely lost their data, they just ended up with a system that didn’t match how they sell, so the spreadsheet quietly came back.
This guide covers the setup sequence in the order the decisions need making. Where migration detail matters, it points at our CRM migration guide rather than repeating it.
In this guide
- Why does CRM setup go wrong?
- Step 1: Map your workflows before configuring anything
- Step 2: Connect your integrations before the data
- Step 3: Get your data ready to import
- Step 4: Test the import, then run it properly
- What to leave switched off for now
- TL;DR
Why does CRM setup go wrong?
Most small business CRM projects fail not because of the software. They fail because the data going in is a mess, the system gets configured to mirror old habits rather than improve them, and the team never gets shown how to use it properly.
The result is a CRM that becomes an expensive contact list nobody trusts. Within six months, someone suggests going back to spreadsheets.
Getting the setup right upfront, in the correct order, avoids all of that. It takes longer to do properly, but you only have to do it once. If you want to know what typically goes wrong before you start, our companion guide covers the five mistakes that cause the most damage.
Step 1: Map your workflows before configuring anything
Before you open the CRM settings, sit down with the people who’ll use it and document how your sales and follow-up processes actually work. What happens when a new lead comes in? What stages does it move through? What triggers a follow-up? Who owns what?
This matters because every CRM lets you customise pipelines, fields, and automations. If you configure those based on assumptions rather than your real process, the system will feel wrong to your team and they’ll work around it.
A few things to map before you configure:
- Your sales pipeline stages, named as your team actually uses them
- The custom fields you need that the CRM doesn’t include by default
- Which tasks or follow-ups you want the CRM to automate
- Which contacts get tagged or segmented and how
The CRM should be configured to reflect how your team works, not the other way around. If you’d rather have this done for you, what to ask a CRM consultant covers what a proper setup and handover should include, and our UK CRM setup pricing sets out the going project rates. If you’re unsure how to structure your tags and custom fields before migration, our guide to CRM tags and custom fields covers the decision rules and a practical starter set.
Step 2: Connect your integrations before the data
Before you import your data, get your integrations working. This means connecting the CRM to your email, your web forms, your email marketing tool, and any other software your team uses daily.
Setting integrations up early matters for two reasons. First, it’s much easier to test them on a clean system than after a full data import. Second, some integrations affect how data is structured in the CRM, and you want those structures in place before records start coming in.
Common integrations worth prioritising:
- Email inbox (Gmail or Outlook) for logging conversations automatically
- Web forms so new enquiries land directly in the CRM as leads
- Email marketing tool if you’re sending newsletters or sequences
- Calendar for scheduling and tracking meetings
If you’re using Zapier to connect tools that don’t have a native integration, set those up and test them before the data migration. Our guide to using Zapier with your CRM walks through how to approach that.
Step 3: Get your data ready to import
Only now, with the configuration decided, is it worth touching your data. Preparing it last rather than first sounds backwards, but you can’t map a spreadsheet column to a custom field that doesn’t exist yet, and every import you run before the fields are built is one you’ll run again. Once you’re live, keeping records clean becomes an ongoing job, which is where governing your CRM data comes in.
Before exporting anything from your existing spreadsheets or old CRM, go through your contacts and remove duplicates, fix formatting inconsistencies, delete contacts that are years out of date, and fill in gaps where you can. Pay particular attention to:
- Duplicate records for the same person or company
- Contacts with no email address or phone number
- Inconsistent naming conventions (Ltd vs Limited, for instance)
- Fields being used for different purposes by different people
Assign one person to own this process. Data cleaning done by a committee produces inconsistent results.
The table below runs through every common data type you’ll likely import, the issues that typically show up in each, and how to fix them before the import runs.
| Data type | What it includes | Common issues before import | How to clean it up |
|---|---|---|---|
| People (contacts) | First name, last name, job title, personal email, mobile/direct phone | Names crammed into one column, “Mr/Dr” prefixes, inconsistent capitalisation (JOHN SMITH vs john smith), missing first or last name | Split full names into separate First Name and Last Name columns. Strip titles into their own column. Apply proper case. Every person record needs at least a first or last name, otherwise most CRMs will reject it. |
| Organisations / companies | Company name, website, main phone, industry, employee count, parent company | Same company spelled five different ways (“Acme Ltd”, “ACME Limited”, “Acme Co.”), trading names mixed with legal names | Standardise to one canonical version per company. Decide upfront: legal name or trading name. Remove Ltd/Inc/LLC inconsistencies. Deduplicate first. Most CRMs treat organisation names as unique, so duplicates will fail on import. |
| Email addresses | Work email, personal email, generic inboxes (info@, sales@) | Missing @ signs, trailing spaces, multiple emails stuffed in one cell, role-based inboxes mixed with personal ones | Validate every email contains an @ and a domain. Trim whitespace. Split multi-email cells into separate columns (Email 1, Email 2). Flag generic inboxes separately. They shouldn’t go against a named person. |
| Phone numbers | Mobile, office, direct dial, fax | Mixed formats (07700 900123, +44 7700 900123, (0)7700-900-123), no country code, extensions glued onto the number | Pick one format. E.164 (e.g. +447700900123) is safest and universally accepted. Move extensions into a separate column. Remove brackets, dots, and spaces if your CRM is fussy about formatting. |
| Postal addresses | Street, city, region/county, postcode, country | Multi-line addresses pasted into one cell, missing country, postcode in the wrong column | Use one column per address component (Street, City, County, Postcode, Country). Always populate Country. Never assume it. |
| Tags / segments | Lead source, customer type, lifecycle stage, interests | Free-text tagging gone wild (“hot lead”, “Hot Lead”, “HOT”, one cell containing 15 tags with no consistent separator) | Agree a finite tag vocabulary before import. Lowercase everything. Use a single separator, either comma or semicolon, whichever your CRM prefers. |
| Custom fields | Anything industry-specific: contract renewal date, account number, preferred contact method | Mixed data types (dates as text, numbers with currency symbols, dropdown values that don’t match the picklist) | Match each column to a planned custom field type in the CRM. For dates pick one format for the whole sheet. For checkboxes use Y/N or TRUE/FALSE. For dropdowns, every value must match the picklist exactly, with no typos and no extra spaces. |
| Dates | Birthdays, contract dates, last contacted, first signed up | UK vs US format mixed (is 01/02/2025 January or February?), Excel auto-converting dates to serial numbers, blank cells entered as “N/A” | Pick one supported format (yyyy-mm-dd is unambiguous and universally accepted). Reformat the entire column at once. Replace “N/A”, “unknown”, and “TBC” with blank cells. |
| Notes and history | Meeting notes, call logs, previous comments | One giant blob with no dates, multiple notes mashed together, notes attached to the wrong contact | Use the separator convention your CRM expects. Be aware: Excel truncates cells at around 32,000 characters, so heavy note histories may need an API import rather than a CSV. |
| Deals / opportunities | Deal name, value, stage, expected close date, owner | Won/lost deals mixed with open ones, currency symbols in value columns, stages that don’t exist in the new CRM yet | Filter to what you need (open deals plus the last 12 months of won/lost is usually enough). Strip currency symbols. Value should be a pure number. Map old pipeline stages to new ones before importing. |
| Marketing status / consent | Subscribed, unsubscribed, bounced, GDPR consent date, consent source | Missing consent records, blank or ambiguous status, unsubscribes not flagged | Never import unsubscribes as active contacts. That’s bad practice and likely illegal under UK GDPR. Confirm consent source and date for each contact. Map statuses to the values your CRM uses (typically Active, Inactive, Unsubscribed, Denylisted). |
| Record ownership | Which salesperson or user owns the contact or deal | Owner column has full names that don’t match CRM usernames, owners who have left the company | Use the CRM’s exact username format (often an email address or handle). Reassign records belonging to ex-staff to a current user or a shared account before importing. |
| File encoding | The CSV file itself | Special characters (accented letters, pound signs, smart quotes) showing as ? or gibberish after import | In Excel, save as CSV UTF-8 specifically, not the default CSV option. File, Save As, CSV UTF-8. In Notepad or TextEdit, save with Unicode (UTF-8) encoding. |
Step 4: Test the import, then run it properly
Don’t attempt to migrate all your data in one go without testing first. Export a small, representative sample of around 50 to 100 contacts and import that into the CRM to see what happens.
A test migration tells you whether your field mapping is correct, whether the data formats are compatible, and whether anything is getting lost or garbled in the import. It’s far easier to fix a problem on 100 records than to untangle it across 5,000.
Check the imported records manually after the test. Look at a mix of contact types and confirm that every field has landed where it should. If something looks wrong, fix the source data or adjust your field mapping before proceeding.
Then run the full import
Once the test import lands cleanly, bring the rest across. Most CRMs accept CSV imports, and many offer direct import from common platforms like HubSpot, Salesforce, or Pipedrive if you’re switching from one of those.
After the full import, validate the results before you let your team in. This doesn’t mean checking every record; it means checking a statistically representative sample across different contact types and making sure:
- Total record count matches what you imported
- No fields are blank that should have data
- Custom fields and tags have come through correctly
- Any company or deal associations are intact
Validation isn’t optional. It’s the point at which you confirm the import worked, rather than assuming it did.
What to leave switched off for now
The most common setup mistake we see after a rushed rollout isn’t too little configuration, it’s too much. Automations firing on rules nobody agreed, six pipelines where two would do, and custom fields created because the CRM offered them rather than because anyone needed the data.
Leave these until the CRM has been in real use for a month:
- Automation rules beyond the obvious. Auto-assignment and a follow-up task on a won deal are fine at launch. Scoring models and multi-step sequences need real usage data behind them.
- Extra pipelines. Start with one, or two where quoting and delivery genuinely differ. Splitting further is easy later. Merging isn’t.
- Reporting dashboards. Build these once a month of real data exists. A dashboard built on the import alone measures your old spreadsheet.
- Optional fields. Every field you add is a field someone has to skip. If nobody can name the report it feeds, don’t create it.
A month in, you’ll know which of these you want, and the answers will be different from the ones you’d have guessed at launch. If you’d rather not work through this alone, our CRM setup and configuration service does exactly this sequence with you.
TL;DR
- Map your actual workflows before touching CRM settings. Configure the system to fit your process, not the default template.
- Connect integrations before the data arrives. They’re easier to test on an empty system, and some of them change how records are structured.
- Prepare your data last, not first. You can’t map a column to a custom field that doesn’t exist yet.
- Always run a test import of 50 to 100 records and check them by hand before committing to the full one.
- Validate a representative sample before you let anyone log in. Record counts, custom fields, tags, and associations.
- Leave automations, extra pipelines, dashboards and optional fields switched off for the first month. Over-configuration kills adoption faster than under-configuration.
Configuring a CRM from scratch?
The decisions in week one are the ones that stick, and they’re much cheaper to get right than to unpick. We can map your process and configure the platform around it before your team ever logs in.