Sooner or later every growing business hits the same wall. The spreadsheet that tracked your leads has become four spreadsheets, the shared inbox is doing the job of a database, and someone asks, “Should we just build our own CRM?”
My short answer: most businesses should start with an off-the-shelf CRM, and a smaller number genuinely need something custom. The right choice depends less on budget and more on how unusual your workflow is and how central it is to the way you make money. Here is how I help clients decide.
What a CRM is really for
Strip away the feature lists and a CRM does three things. It gives every person and conversation a single home. It shows where each relationship stands. And it reminds the right person to do the next thing at the right time.
If a tool does those three things well for your business, it is a good CRM, whether it cost nothing or took months to build. If it does them badly, no amount of dashboards will save it. Keep that test in mind as you read the rest of this post.
When off-the-shelf is the right call
Commercial CRMs have been refined by thousands of companies with sales pipelines that look a lot like yours. That is their superpower. You get contact management, pipelines, email integration, reporting and a mobile app on day one, and someone else handles security patches and uptime.
Off-the-shelf is usually the better choice when:
- Your sales process is fairly standard: lead, qualify, propose, close, follow up.
- Your team is small and needs to be productive this month, not next quarter.
- You don’t yet have documented processes, so you don’t know exactly what you would build.
- The CRM supports your business rather than being the product your customers experience.
- You would rather pay a subscription than own and maintain software.
That third point matters more than people expect. In why most small businesses don’t need AI yet I wrote that tools amplify systems rather than create them. The same is true of CRMs. A custom build written around a process nobody has defined will faithfully encode the confusion.
When a custom CRM earns its keep
Custom starts to make sense when your workflow is the thing that makes you different, and bending it to fit a generic tool costs you every single day. I see this most in industries with specific operational data that generic CRMs weren’t designed to hold.
Home healthcare is a good example. An agency isn’t just tracking leads. It is tracking referral sources, client intake, caregiver credentials, schedules, authorizations and billing, and those records all need to talk to each other. That is exactly why, at Pinn.Media, we built Pinn.Care as an ERP covering HR, care coordination, caregiver matching, marketing and billing, instead of asking agencies to stitch five tools together. I go deeper on that in my guide to home healthcare agency software.
Signs you may need custom:
- You maintain workarounds: extra spreadsheets, duplicate data entry, or custom fields that nobody understands.
- Your data model is unusual, with relationships a standard contact/deal/company structure can’t represent cleanly.
- You need a client, patient or partner portal that feels like your brand, not a vendor’s.
- Compliance or data-residency rules limit where your data can live and who can see it.
- Per-seat pricing grows faster than the value you get as your team expands.
- You need deep integrations with systems (an EHR, a payment processor, a scheduling engine) that the off-the-shelf options only partly support.
A side-by-side comparison
| Factor | Off-the-shelf CRM | Custom CRM |
|---|---|---|
| Time to first use | Days to weeks | Weeks to months |
| Upfront cost | Low; mostly setup and training | Higher; design, build and testing |
| Ongoing cost | Subscription, often per user | Hosting, maintenance and improvements |
| Fit to your workflow | You adapt to the tool | The tool adapts to you |
| Ownership | You rent it; exports vary | You own the code and the data model |
| Maintenance | Handled by the vendor | Needs a developer or a care plan |
Neither column is “better.” The question is which set of trade-offs you would rather live with for the next three to five years.
The questions I ask before recommending either
When someone asks me to build a CRM, the first conversation is rarely about technology. It is about how work actually moves through the business. These are the questions I work through with them:
- Can you draw your process? From first contact to paid invoice, what are the steps and who owns each one?
- Where does data get re-typed? Every copy-paste between systems is a candidate for integration or a sign of a missing system.
- What does a “good day” look like for each role? The receptionist, the salesperson and the owner all need different screens.
- What reports would change a decision? Not what reports would be nice. Which numbers would make you act differently?
- Who will maintain it? A custom tool without an owner slowly decays.
- What must never go wrong? Lost records, missed follow-ups, a privacy breach. These shape the architecture more than any feature request.
Often, working through this list is enough. Sometimes the answer is “configure the CRM you already pay for properly,” which is a perfectly good outcome.
How I approach a custom build
If custom is the right call, I keep the first version deliberately small. I start with the data model, because a clean structure for people, organizations, activities and whatever is specific to your business will outlast any interface. From there I build the few screens each role uses daily, then integrations, then reporting.
On the technical side I typically work with a modern web stack (Next.js and TypeScript on the front end, Node.js or Python services, and PostgreSQL or Supabase for the database), with role-based permissions and audit trails from day one. Integrations with calendars, payments, phone systems and EHRs are where a custom CRM starts paying for itself, because that is where manual work disappears.
I also plan for automation and AI later, even if they aren’t in version one. A CRM with structured, reliable data is what makes an AI receptionist or an automated follow-up actually useful. Without that foundation, AI has nothing trustworthy to work with.
Common mistakes on both sides
Buying too much
Enterprise CRMs can do almost anything, which means they can also become almost unusable. If your team logs in and immediately feels lost, they will go back to their spreadsheets, and you will pay for software nobody uses.
Building too much
The opposite mistake is trying to build every feature you have ever seen in a commercial product. A custom CRM should do fewer things than a commercial one, and do them exactly the way your business works.
Ignoring adoption
The best CRM is the one your team actually uses. Whatever you choose, plan for training, clear ownership and a short feedback loop in the first few weeks, when small annoyances decide whether people stick with it.
So, which one do you need?
If your process is standard and your priority is speed, buy. If your workflow is your competitive edge and generic tools are forcing daily workarounds, build, or build a thin custom layer on top of what you have. And if you can’t yet describe your process clearly, do that first. It is the cheapest improvement you will ever make.
If you are weighing this decision now, I’m happy to help you think it through. You can see the kind of systems I build in my portfolio, read about my background on my journey page, or tell me about your workflow and I’ll give you an honest recommendation, even if it is “keep what you have.”
