A customer calls support about an implementation promised during a sale. The support agent finds the account, but the deal notes belong to a different company record. The salesperson logged the conversation under the parent organization, billing uses the subsidiary’s legal name, and the customer recognizes neither of those names as the product workspace they use.
Choosing a CRM platform starts with making those relationships legible. A customer relationship management system holds the information people rely on to understand a customer and act on that understanding. A pipeline view may help sellers see deals, and a service view may help agents resolve issues. Both depend on getting the underlying record right. The strongest evaluation follows one customer across teams, tests an exception, and checks who can see or change the result.
Draw the customer before comparing products
Write down what your organization means by customer. In a business-to-business sale, the customer might include a parent company, several purchasing entities, a reseller, and individual users. A consumer business might need to represent one person with several orders, contact channels, or subscriptions. A membership organization may care about relationships among people, employers, and chapters. Each example asks different things of a CRM.
HubSpot’s CRM documentation distinguishes objects, records, and properties and describes associations between records. Those terms are useful for thinking through the design even if you choose another system. An object is a type of thing, such as a company or a deal. A record is an individual instance of that type. A property stores an attribute, while an association describes a connection. Ask the provider to demonstrate the specific relationships you need and identify any limits that apply to the plan under consideration. HubSpot’s CRM database guide.
Consider a buyer who works for one company and advises a second. Creating two contact records can split the interaction history and introduce two versions of the person’s preferences. Linking one contact to several organizations may serve the workflow better, provided the CRM can distinguish the person’s role in each relationship. In the demo, make the buyer an approver in one deal and an adviser in another. Then ask someone unfamiliar with the setup to find both roles.
Define who owns the facts
A shared customer record can display information drawn from several systems. Give each important field a source and an owner. Sales operations might own the opportunity stage. A billing system might own the legal customer identifier and invoice status. Customer success might own the renewal risk assessment. Ownership clarifies where a correction should happen and prevents an integration from sending an outdated value back to the source.
Use a small field inventory focused on real decisions. For each field, record who writes it, who reads it, what happens when it is missing, and whether you need the history of changes. A field used in a service commitment or financial export deserves a clearer definition than a private note. Where several groups use the same label differently, resolve the meaning before migrating. A field called “active” could refer to a current contract, a user with access, or an account that still receives outreach.
Discuss mismatches explicitly. If a customer changes its legal name, should sales update the display name or should finance confirm the legal entity first? What happens to open quotes when the parent company changes? Identity changes can be routine for one team and consequential for another. Make sure the software gives people a clear way to update records while retaining history needed for reporting and service.
Compare operating models
| CRM approach | A plausible fit | What to demonstrate |
|---|---|---|
| Lightweight sales CRM | One team needs reliable contact history and pipeline management | Deduplicate an import, reassign a deal, and export complete history |
| Cross-functional CRM suite | Sales, marketing, and service need a shared record and defined handoffs | Transfer a closed deal to service and confirm visibility by role |
| Extensible enterprise CRM | Multiple divisions need complex relationships, permissions, or workflows | Model subsidiaries, regional access, and an exception without duplicate records |
| CRM connected to specialist tools | Existing service, billing, or marketing systems already serve their teams well | Recover a failed sync and reconcile record counts and field ownership |
The CRM itself may cost less than the work around it. A suite can bring groups onto a shared platform, but some modules may have separate licensing and administrative requirements. An extensible platform can represent unusual relationships, but its flexibility creates a need for design governance. A modest CRM connected to a mature service desk can be economical if the two systems agree on customer identity and ownership.
A demonstration should use representative user roles and realistic data. Have a salesperson record an opportunity, a service agent handle a related issue, and a manager view both. Then try a harder case where the agent should see the existence of a customer relationship but lack access to some commercial terms. Ask how the same permissions apply in reporting and exports. A setup that works from the home screen may behave differently when the user opens a dashboard or downloads a record list.
Treat migration as an identity project
Start by sampling existing records. Look for duplicate company names, contacts with shared email addresses, inactive records still appearing in reports, and opportunities without reliable owners. Create a mapping from old identifiers to new identifiers. When you merge records, retain the information needed to trace the decision and test whether related activities and deals still point to the right customer.
Choose migration scope by use. Active accounts, open opportunities, and recent activity may be enough to start for some teams. Historical service records or renewal commitments may need to move with the customer, or remain accessible in an archive. Ask which items can migrate through a standard import and which require API work or a separate tool. Attachments, messages, notes, and custom relationships often need dedicated checks.
Reconcile more than totals. A matching record count can coexist with broken relationships. Compare selected accounts from the old and new systems, including a complex parent and subsidiary, a transferred customer, and a contact whose preferences changed. Validate pipeline reports by owner and stage. Document any differences caused by new definitions so managers can compare future reports fairly.
Design the handoff people will use every week
Use a recent closed deal as the acceptance test. Ask sales to hand it off, then give the service or implementation team only the CRM and the permissions they would normally have. They should be able to find the customer’s objectives, relevant contacts, implementation scope, and commitments they are expected to honor. If they still have to call the seller for every handoff, work out which information was missing and why it never reached the record.
Then test a real change. The customer adds a subsidiary, updates its main contact, or requests an additional product. Check which team sees the change, who can act on it, and whether the original purchase history stays connected. Operationally useful customer records remain understandable as relationships evolve.
The right CRM should handle both a normal customer and the awkward account people keep in a private spreadsheet because the existing system cannot represent it. Bring both to procurement. Their histories reveal the permissions, relationships, and handoffs that should carry the most weight in the decision.