Custom CRM is the option most buying committees rule out in the first meeting, usually with a sentence about not reinventing the wheel. Then the Salesforce proposal lands, somebody works out what Dynamics 365 actually costs across every seat that touches a customer, and renting the wheel stops looking obvious. The honest comparison is not philosophical. It is arithmetic, run over five years, with the integration and administration work included on every side.
This guide puts all three options on the same page under the same assumptions. It covers what each one really is, what a custom CRM competes against once you strip the marketing away, how the five-year numbers compare for a mid-sized UK team, where the money goes in year one, what lock-in costs when you decide to leave, and which situations genuinely favour building over buying. Every figure here is a planning model rather than a quote, and the assumptions sit next to the numbers so you can change them.
The short version: Salesforce and Dynamics 365 are excellent when your process resembles the process they were designed around. A custom CRM earns its keep when your commercial advantage lives in a workflow no vendor has modelled. Most companies are one of those two things and know which, but talk themselves into the other. This sits alongside our build vs buy vs low-code decision framework, which covers the general choice in more depth.
Table of contents
- What a Custom CRM Actually Competes Against
- Salesforce: The Platform You Rent, Not Own
- Dynamics 365: The Microsoft Stack Argument
- Custom CRM: When Building Genuinely Wins
- Custom CRM Cost vs Salesforce and Dynamics 365 Over Five Years
- Where the First-Year Budget Actually Goes
- Data Ownership, Lock-In and the Cost of Leaving
- Integration: The Line Item Every Quote Understates
- How to Choose a Custom CRM or a Platform
- Frequently Asked Questions
- References
What a Custom CRM Actually Competes Against
Before comparing prices it is worth being precise about what is on the table, because the three options are not three versions of the same product.
A platform is a rented process, not just software
Salesforce and Dynamics 365 sell you a data model, a workflow engine, a reporting layer and an ecosystem, all shaped by decades of watching how most companies sell. That shape is the value. When your process fits it, you get years of accumulated product thinking for a monthly fee. When it does not, you pay to bend the platform, and you keep paying every time it updates.
A custom CRM is a process encoded in your own schema
A custom CRM inverts the relationship. You define the objects, the states, the rules and the reports, and the software follows. There is no per-seat ceiling on who can use it, no licence tier gating a field, and no upgrade that quietly changes a behaviour you depend on. In exchange, you own everything: the roadmap, the bugs, the security patches and the hosting.
Low-code sits between them and confuses the comparison
Power Apps on Dataverse, or a Salesforce org built almost entirely from declarative tools, is neither fully bought nor fully built. It deserves a place in the comparison because it is often the right answer, but it should be costed honestly: low-code reduces build time and raises licence cost, and the resulting system is still locked to the vendor that runs it.
The real competitor is usually the spreadsheet
Most teams evaluating a custom CRM are not migrating from a rival CRM. They are migrating from a spreadsheet, three inboxes and someone’s memory. That matters, because it means the comparison is not only between platforms — it is between any system and no system, where almost any option beats the status quo.
| Factor | Custom CRM | Salesforce | Dynamics 365 |
|---|---|---|---|
| Upfront cost | High | Low to moderate | Low to moderate |
| Cost per added seat | Near zero | High | Moderate to high |
| Time to first value | 4-6 months | 6-12 weeks | 8-14 weeks |
| Process fit | Exact | Good, if conventional | Good, if conventional |
| Ecosystem and add-ons | Build it yourself | Very large | Large |
| Vendor lock-in | None | Strong | Strong |
| Who fixes a defect | You or your partner | Vendor | Vendor |
| Hiring pool for admins | Narrow | Very wide | Wide |
Salesforce: The Platform You Rent, Not Own
Salesforce is the default because it is genuinely good, and because nobody was ever criticised for choosing it. Both of those are worth understanding separately.
What you are actually buying
You are buying a mature object model, a permission system that survives audit, a reporting engine, and an app marketplace deep enough that most integration problems have a paid answer already. For a conventional sales motion — leads, opportunities, stages, forecasts — the platform will do in a fortnight what a custom CRM team would spend two months building.
Where the cost really sits
The licence is the visible number and rarely the largest one. Enterprise editions land in the region of £120 to £270 per user per month depending on tier and negotiation, but implementation partners, sandbox environments, premium support, data storage overages and the extra products that turn out to be mandatory all sit on top. Published rates are a starting point; the Salesforce pricing pages list editions, and the real figure emerges only after scoping.
The customisation trap
Every Salesforce org drifts toward complexity. Fields get added, validation rules multiply, flows accumulate, and eventually a release note changes a behaviour something depends on. Organisations that heavily customise Salesforce often end up with the maintenance burden of a custom CRM plus the licence bill of a platform, which is the worst of both positions and the single most common failure mode we see.
Where it clearly wins
Salesforce wins when you need to be live this quarter, when your process is recognisably standard, when you want to hire administrators from a large pool, and when the integrations you need already exist as supported apps. It also wins on credibility with investors and enterprise buyers, which is not a technical argument but is a real one.
Dynamics 365: The Microsoft Stack Argument
Dynamics 365 is usually shortlisted for a reason that has nothing to do with CRM features: the company already runs on Microsoft.
The integration story is the product
If your business lives in Microsoft 365, Teams, Excel and Power BI, Dynamics 365 arrives already connected to all of it. Records surface in Outlook, reports render in Power BI without a connector project, and identity flows through Entra. That coherence is worth real money, and it is the argument a custom CRM has to beat on its own merits rather than dismiss.
Dataverse changes the calculation
Underneath Dynamics 365 sits Dataverse, which means the same data can be extended with Power Apps rather than bespoke code. For a company that wants more flexibility than Salesforce configuration allows but less commitment than a full build, this is the genuine middle path — and the reason low-code belongs in the comparison at all.
Licensing is more complex, not cheaper
Dynamics 365 often quotes lower per-seat than Salesforce, and the total frequently converges once you count the Power Platform capacity, the additional app licences and the storage. The model rewards companies with existing enterprise agreements and punishes those buying à la carte. Our comparison of Microsoft Dynamics 365 vs SAP goes further into how these agreements are structured.
The update cadence is not optional
Microsoft ships two major waves a year and does not let tenants sit still indefinitely. That is good for security and bad for anyone who built fragile customisations. Plan for regression testing every cycle, and treat that as a running cost rather than a surprise.
Custom CRM: When Building Genuinely Wins
A custom CRM is not the romantic choice or the reckless one. It is the correct choice in a specific and identifiable set of circumstances.
Your process is the product
If how you qualify, price, schedule or service is the thing customers actually pay for, encoding it in someone else’s data model is a slow way to lose the advantage. Specialist lenders, logistics operators, clinical services and regulated brokers routinely find that no standard pipeline describes their work. A custom CRM lets the software match reality rather than the other way round.
Seat count makes the maths flip
Platform pricing is linear in users; a custom CRM is not. At twenty seats, licences are noise. At two hundred seats — especially when many are occasional users who only need to read a record or update a status — the annual licence bill starts to exceed what a competent team costs to run the whole system. This is the single most reliable trigger for building.
You need data you are not allowed to hand over
Some organisations cannot put customer data in a shared tenancy without a compliance conversation they would rather avoid entirely. Where residency, retention or auditability requirements are unusually strict, controlling the whole stack is simpler than negotiating exceptions. The ICO’s guidance on data minimisation is a useful lens here, because a custom CRM lets you store only what you need by design.
The honest counter-argument
Building fails when the driver is dissatisfaction rather than differentiation. “We hate our CRM” is not a specification. If the underlying problem is bad data, unclear process or no adoption, a custom CRM will reproduce all three faithfully and more expensively. Fix the process first, then decide.
Custom CRM Cost vs Salesforce and Dynamics 365 Over Five Years
One year of comparison flatters the platforms and five years flatters the build. Five is the fair window, because it covers one full replatform cycle and at least one licence renegotiation.
The assumptions behind the model
The figures below model a UK company with 60 CRM seats: 35 full users, 25 light users, a moderate integration footprint of four systems, and standard reporting. A custom CRM is assumed to be delivered by an external team, hosted on mainstream cloud computing services at UK list prices, and then maintained at roughly 18% of build cost annually. Change any assumption and the ranking can change with it — which is exactly the point.
What the five-year picture looks like
Indicative five-year total cost of ownership, 60 seats, UK mid-market, in thousands of pounds.
| Cost line (5 years, 60 seats) | Custom CRM | Salesforce | Dynamics 365 |
|---|---|---|---|
| Initial build or implementation | £220,000 | £70,000 | £65,000 |
| Licences over five years | £0 | £396,000 | £312,000 |
| Hosting and infrastructure | £54,000 | £0 | £18,000 |
| Maintenance and enhancement | £198,000 | £90,000 | £85,000 |
| Integration work | £23,000 | £54,000 | £40,000 |
| Five-year total | £495,000 | £610,000 | £520,000 |
Why the gap is smaller than either camp claims
Platform advocates quote the build cost and ignore five years of licences. Build advocates quote the licences and ignore maintenance. Run both properly and the totals land within about 20% of each other for a mid-sized team, which means cost alone rarely decides it. Fit, risk and speed decide it, and cost only breaks a tie. For a fuller treatment of the build side, see our guide to custom software development cost.
Where the crossover point sits
Below roughly 25 seats, platforms win on cost almost every time. Between 25 and 100 seats the answer depends on process fit. Above about 150 seats, or wherever a large population of light users exists, a custom CRM becomes financially compelling and the conversation shifts to whether you can deliver it well.
Where the First-Year Budget Actually Goes
The most common budgeting mistake is treating a custom CRM as a screen-building exercise. Screens are the cheap part.
Integration dominates, every time
Across the builds we have delivered, integration consistently takes the largest share of first-year effort — connecting accounting, telephony, marketing automation and whatever bespoke system runs operations. Anyone who has read our analysis of API integration cost will recognise the pattern: the second and third integrations cost more than the first, not less.
Migration is where projects actually stall
Moving twelve years of inconsistent records into a clean schema is unglamorous, politically fraught and routinely underestimated. It is also the phase where a custom CRM has a real advantage: you can shape the schema around the data you actually have rather than forcing it into a fixed model and losing nuance in the process.
Adoption is a budget line, not a hope
Training, documentation, a migration period where both systems run, and someone whose job is to chase usage — none of it is optional, and all of it is missing from most business cases. Budget five to eight per cent of the build for adoption or expect the system to be technically live and commercially ignored.
Data Ownership, Lock-In and the Cost of Leaving
Nobody signs a CRM contract thinking about the exit, which is precisely why the exit is expensive.
What lock-in actually means
Lock-in is not whether you can export a CSV — you always can. It is whether your process survives the move. Automations, validation rules, permission models, report definitions and the integrations built against a vendor’s API are all unportable, and rebuilding them is most of the original implementation cost paid again.
Estimating your own exit cost
A reasonable rule of thumb is that leaving a heavily customised platform costs 60 to 80 per cent of what implementing it did, spread over six to twelve months. With a custom CRM the equivalent risk is different rather than absent: you are locked to a codebase and to whoever can maintain it, which makes source code ownership and documentation quality the things to negotiate hard.
Ownership clauses matter more than architecture
If you build, the contract decides whether you own anything. Assignment of copyright, escrow arrangements, documentation standards and a clean handover of infrastructure credentials are what convert a custom CRM from an asset into a dependency or back again. Our guide to source code ownership covers how these clauses are usually written and where they fail.
Regulatory exit obligations
Data subject access, retention limits and the right to erasure all have to work in whatever system you move to. Confirm that both the outgoing and incoming platform can satisfy the ICO’s right of access requirements before you commit, not during a migration weekend.
Integration: The Line Item Every Quote Understates
Whichever option you choose, the CRM will only be as useful as the systems it talks to, and this is where estimates go wrong on all three paths.
Platforms make the easy integrations easy
Connecting Salesforce or Dynamics 365 to mainstream accounting, marketing or telephony products is genuinely close to configuration. If every system you use is mainstream, this advantage is large and a custom CRM cannot match it without real engineering.
Platforms make the hard integrations harder
Connect either platform to a twenty-year-old bespoke system and the advantage inverts. You are now working inside the vendor’s API limits, governor thresholds and deployment model, with no ability to change the constraint. A custom CRM has no such ceiling, which is why companies with unusual internal systems so often end up building.
| Your situation | Best fit | Why |
|---|---|---|
| Under 25 seats, conventional sales process | Salesforce or Dynamics 365 | Licences are cheap at this scale and speed matters more |
| Already deep in Microsoft 365 and Power BI | Dynamics 365 | Identity, reporting and Teams integration arrive free |
| Large population of occasional light users | Custom CRM | Per-seat pricing punishes read-only and status-update users |
| Your workflow is your competitive advantage | Custom CRM | No standard pipeline describes what you actually do |
| Need to be live within one quarter | Salesforce | Fastest route from decision to working system |
| Strict residency or auditability constraints | Custom CRM | Full control of hosting, retention and audit design |
| Want flexibility without a full build | Low-code on Dataverse | Extends a supported platform with far less bespoke code |
| No clear process, poor data discipline | Neither yet | Fix the process before encoding it in any system |
Budget integrations separately and first
Whatever route you pick, scope and price the integrations before you commit to the platform decision, because integration difficulty is the variable most likely to reverse the ranking. Every connection is also a cybersecurity boundary, and following NCSC cloud security guidance on how you authenticate and segregate those connections is worth doing at design time rather than after the first audit.
How to Choose a Custom CRM or a Platform
The decision is more tractable than it looks if you take it in order and refuse to skip the first step.
Write the process down before you shop
Map how a customer actually moves through your business, in the words your team uses, before you see a single demo. If that map looks like the standard pipeline, buy. If three of the stages have no equivalent in any product you have seen, a custom CRM has just become a serious candidate.
Count your light users honestly
List everyone who will ever need to see or touch a record — finance, operations, support, field staff, senior management. That number, not the sales headcount, drives platform licensing and is the figure most often understated by half when the business case is written.
Run the five-year model with your own figures
Take the table above, replace every number with a quote you have actually received, and extend it to five years. Do it in a spreadsheet somebody else can challenge. A custom CRM business case that only works on optimistic maintenance assumptions is not a business case.
Pilot before you commit
Both platforms offer trials, and a competent partner will build a working slice of a custom CRM — one real workflow, one real integration, real data — for a fraction of the full build. Nothing else resolves the argument as quickly as putting a narrow version in front of the people who will use it.
Decide who will own it on day 400
Platforms need administrators; a custom CRM needs a development relationship. Whichever you choose, name the person or partner responsible for it eighteen months from now during the decision, because unowned systems decay regardless of how they were built.
Frequently Asked Questions
How much does a custom CRM cost to build in the UK?
A focused first release for a single team typically lands between £60,000 and £120,000. A full replacement for a mid-sized company with several integrations and migrated history is more commonly £180,000 to £350,000. Anything quoted below £40,000 is either a very narrow tool or an estimate that has not accounted for integration and migration.
Is a custom CRM cheaper than Salesforce?
Over five years, usually yes above about 100 seats, and usually no below 25. The crossover depends far more on how many light users you have and how much platform customisation you would need than on the headline licence rate.
How long does a custom CRM take to deliver?
Expect four to six months to a genuinely usable first release, and a further six to twelve months of iteration before it does everything the old system did. Platforms reach basic operation faster, though a heavily customised implementation often finishes no sooner than a build would have.
Can we start on a platform and move to a custom CRM later?
Yes, and it is a reasonable strategy — use a platform to learn what you actually need, then build once the process is proven. Budget for the exit cost from the start and keep your customisation shallow so the eventual migration stays affordable.
What about open-source CRM options?
They occupy a middle ground: no licence fee, but the hosting, security patching and customisation burden of a custom CRM without the benefit of a schema designed around your process. They suit teams with in-house engineering capacity and conventional requirements.
Does AI change this comparison?
It shifts it modestly. Both platforms ship capable assistants that a custom CRM must be built to match, which raises the bar for building. Working against your own schema also makes it easier to apply models to your specific data, so the effect cuts both ways rather than settling the argument.
References
Salesforce: Sales Cloud Pricing and Editions
Microsoft: Dynamics 365 Pricing Overview
Microsoft Learn: Dynamics 365 Sales Documentation
Microsoft Learn: What Is Microsoft Dataverse
Microsoft Learn: Dynamics 365 One Version Service Updates
Salesforce Help: Understanding the Salesforce APIs
