Guides
Most companies that ask us whether to build a custom CRM should not build one. This guide gives you a weighted scoring framework, a five-year cost comparison at a stated seat count, and a plain answer for when buying is correct — read that section first if you suspect it applies to you.
It sits alongside our detailed custom CRM development cost guide and our CRM and portal development service.
Read this first
If your sales or service process is standard, your seat count is modest — say under 100 — and you do not have an unusual integration surface, you should buy Salesforce, HubSpot or Microsoft Dynamics 365 and stop reading this guide. All three are mature, well-supported platforms with large partner ecosystems, predictable pricing and a low risk of the project failing outright. A custom build for a standard sales pipeline and a handful of common integrations is a way to spend £300,000 recreating features these vendors ship for free.
This is not a hedge. Most CRM requirements really are standard: contact and account management, a pipeline with stages, activity logging, email integration, basic reporting. Vendors have spent two decades refining these workflows and testing them against millions of users. A custom team, however good, starts from zero on all of it.
Buy first. If the platform still cannot do what you need after a proper configuration attempt with an experienced implementation partner, you will know precisely which gaps a custom build needs to close, and you will have a far more credible case for the spend.
Framework
Score each criterion from 1 (strongly favours buying) to 5 (strongly favours building), multiply by the weight, and sum. A total below 2.5 favours buying outright; 2.5–3.5 warrants a discovery-level investigation before deciding; above 3.5 is a reasonable case for a custom build, though it should still be tested against the cost comparison below.
| Criterion | Weight | What to assess |
|---|---|---|
| Process differentiation | 25% | Does your sales or service process differ materially from the workflow the platform vendor assumes? |
| Integration surface | 20% | Number and depth of systems the CRM must exchange data with in real time. |
| Regulated data handling | 15% | Whether records contain data subject to sector-specific handling rules beyond general data protection law. |
| Seat count and growth | 15% | Current seats and the seat count expected in five years. |
| Internal engineering capacity | 10% | Whether you have, or are willing to fund, a team that can own a custom system for its lifetime. |
| Time to first value | 10% | How quickly the business needs a working system in front of users. |
| Vendor lock-in tolerance | 5% | How much you mind being unable to leave a platform without a costly re-platforming project. |
Score honestly rather than in the direction you already prefer. The most common distortion is scoring "process differentiation" high because a process feels important internally, when in practice it is a configuration variant a platform already supports. Ask an implementation partner for the platform you are considering to try to configure your process before you score this criterion.
Total cost of ownership
The figures below are indicative bands for a 150-seat deployment with a moderate integration count (four to six systems), not quotes. They exclude the value of features gained or lost on either side — a genuine like-for-like comparison also has to weigh what each option lets the business do, not only what it costs.
| Cost item | Off-the-shelf | Custom build |
|---|---|---|
| Per-seat licences (150 seats, mid-tier plan, 5 years) | £450,000–£675,000 | £0 |
| Implementation partner (setup, configuration, training) | £120,000–£220,000 | £0 |
| Customisation (custom objects, workflow, approval logic) | £80,000–£180,000 plus ongoing dev retainer | Included in build |
| Integration middleware and connectors | £40,000–£90,000 licence plus £30,000–£70,000 build | Built directly into the schema, no middleware licence |
| Admin headcount (platform administrator, 5 years) | £175,000–£250,000 (0.5–0.7 FTE) | £100,000–£150,000 (typically shared with wider engineering support) |
| Renewal price uplift (typical 8–15% per renewal cycle) | £60,000–£150,000 cumulative | Not applicable |
| Initial build | Not applicable | £250,000–£600,000 depending on scope |
| Hosting and infrastructure (5 years) | Included in licence | £40,000–£100,000 |
| Ongoing support and maintenance (5 years) | Included in licence and partner retainer above | £125,000–£300,000 (typically 10–15% of build cost per year) |
| Iteration and new feature work (5 years) | Constrained to what the platform's roadmap allows; extra cost for custom logic | £150,000–£400,000, fully directed by your own priorities |
At this seat count and integration profile, the totals typically land in the £1,050,000–£1,565,000 range for off-the-shelf and £665,000–£1,450,000 for a custom build over five years — close enough that the decision should be made on process fit and integration risk, not on cost alone. The crossover moves sharply in favour of custom as seat count rises, because the off-the-shelf column scales with seats and the custom column largely does not.
The other side
Your process is the differentiator. If the way you qualify, price, underwrite or service customers is a genuine competitive advantage, forcing it into a generic pipeline either destroys the advantage or requires so much platform customisation that you have effectively built a custom system on top of an expensive licence. Energy brokers comparing live tariff data across multiple suppliers are a common example — see our energy broker CRM cost guide and broker CRM development page.
Regulated data flows. Where records must be handled, retained, redacted or audited according to sector-specific rules — clinical records, for example — bending a general-purpose CRM to those rules is often harder than designing the data model around them from the outset. Our patient portal development cost guide covers this in a healthcare context.
Heavy integration surface. If the CRM needs to be the system of record reconciling data from eight or more systems in near real time, with conflict resolution logic specific to your business rules, off-the-shelf integration middleware becomes its own expensive, brittle layer. A custom system can absorb that logic directly.
Per-seat economics break at scale. Beyond roughly 300–500 seats, the licence line in the cost table above frequently exceeds the entire cost of a custom build, and it keeps rising every year through renewal uplifts you do not control.
Questions
No. Buying an off-the-shelf CRM is the right call for most companies. This guide exists to help the minority whose process, data flows or integration surface make a custom build genuinely cheaper and safer, tell the difference from the majority who should simply buy Salesforce, HubSpot or Dynamics.
Yes, and it is often sensible. Buy first, run the platform for a year, and use the resulting list of workarounds and shadow spreadsheets as the requirements document for a custom scoping exercise. That list is far more reliable than requirements gathered before anyone has used a system in anger.
Through a fixed-scope discovery sprint that produces an architecture document, integration map and a phased estimate with a stated confidence level for each phase. We do not quote a build price from a sales call.
It should be contractually yours, in a documented schema, on infrastructure you control or can migrate off without cooperation from the original builder. Ask any prospective supplier to state this in writing before you sign anything.
Send us your seat count, current systems and the process you think is unusual. We will tell you honestly whether it warrants a custom build.
Talk to us