Salesforce dual identity: CRM out-of-the-box AND platform for custom apps.
As CRM application:
- Sales Cloud, Service Cloud out-of-the-box.
- Standard objects, processes, UI.
- Customisation modest.
- Common use case for most customers.
As platform (force.com / Lightning Platform):
- Custom objects, custom flows, custom UI.
- Salesforce Platform license (cheaper than full).
- Build line-of-business apps.
- Use platform features (security, mobile, reporting) for non-CRM apps.
Platform examples:
- HR app (employee self-service, time off, performance reviews).
- Project management app.
- Loan origination app.
- Patient onboarding (healthcare).
- Vendor management.
Trade-offs of platform use:
Pros:
- Shared infrastructure (security, identity, mobile).
- Cheaper licensing than full Salesforce.
- Mature platform with declarative tooling.
- Existing skill pool (admins, devs).
Cons:
- Generic, not specialised for your use case.
- Some Salesforce-specific patterns may not fit non-CRM use.
- Vendor lock-in similar to other PaaS.
- Cost can grow.
Decision criteria:
- CRM-shaped problem -> Salesforce CRM.
- Generic LOB app, leveraging existing Salesforce -> Salesforce Platform.
- Bespoke app with custom requirements -> Heroku / AWS / Azure.
Hybrid:
- Salesforce CRM + Salesforce Platform apps + Heroku apps + AWS services. All connected via integration.
Architectural insight:
- License cost matters. Platform license much cheaper.
- Performance identical to full Salesforce.
- Salesforce Platform users often access Account/Contact in shared way.
- Integration patterns same as full Salesforce.
Common pitfalls:
- Force-fitting non-CRM into Salesforce CRM — wrong tool.
- Building custom from scratch when Salesforce Platform fits — wasted effort.
- Not budgeting for license growth — cost surprises.
Senior architect insight: Salesforce is a platform first, CRM second for the technically-curious. Many highly-successful "Salesforce apps" aren't CRM at all.
The senior framing: the question is fit, not category. If your app shape fits Salesforce's strengths (declarative tooling, identity, sharing, mobile), it's a candidate. If it needs uncommon features (heavy compute, real-time WebSockets), Heroku/AWS may be better.