What You Are Actually Buying
A CRM built around the way your business tracks relationships, rather than one you adapt your process to fit.
Off-the-shelf CRMs impose a structure. You accept their pipeline stages, their categories, their reporting dimensions, and you shape the way you work to match. A custom CRM inverts that: we model your relationships, your lifecycle stages and your data points, then build the interface and the logic around them.
Our own platform manages client relationships across projects, billing, contracts, communications and deliverables in one system, because no CRM on the market modelled the agency-client relationship the way we needed. That same reasoning applies to any business whose relationships are more complicated than a generic tool assumes.
When Configuring Something Off The Shelf Is The Better Call
Worth saying first, because it is the answer more often than it is not.
If your process fits a standard sales funnel, your data model is straightforward, and a well-configured HubSpot or Pipedrive would genuinely serve you, configure rather than build. A custom CRM you did not need is an expensive way to arrive somewhere a subscription would have taken you.
The point where that stops being true is recognisable:
- The team works around the CRM constantly. Custom fields everywhere, ignored features, spreadsheets filling the gaps
- Relationships carry several dimensions at once (projects, contracts, billing, support) that no single CRM models natively
- The CRM needs to integrate deeply with internal systems rather than through surface-level connectors
- Reporting requirements are specific enough that generic reports do not answer the actual questions
- You are paying enterprise licence fees while using a fraction of the features and missing the ones that matter
How These Projects Run
They start with relationship mapping: every entity in the business that relates to another. Clients, contacts, projects, deals, invoices, communications, documents. How they connect, and how those connections change over time.
That produces the data model and the workflows. What stages does a relationship pass through, what happens at each, what does somebody need to see when they open a client record. Those answers decide the schema, the interface and the automation.
Integration is consistently the hardest part, and it is where the estimate moves. A CRM that does not connect to invoicing, email, project management and communication is one more silo and one more tab to check.
What Moves The Price
Four things, in roughly this order of impact.
How many systems it has to talk to, and how good their APIs are. This is usually the largest variable by some distance. A well-documented API is a known quantity. One that is undocumented, read-only, or rate-limited turns a fortnight into a month.
How much of the data model is genuinely unusual. Standard contacts and companies are quick. Relationships that branch, overlap or change type over time are where the design time goes.
Migration. Years of records from an existing CRM, with duplicates, inconsistent formatting and notes and attachments that matter. This is regularly underestimated and is worth pricing separately.
How many roles need different views. Two is simple. Seven, each with a different scope of what they may see, is a permissions model rather than a preference.
Where CRM Projects Go Wrong
Nobody owns adoption. The system is delivered, the team keeps using spreadsheets, and within three months the real pipeline lives somewhere else. This is the single most common failure and it is not a technical one.
Too many required fields. Every mandatory field is a reason to skip entering something, and skipped entries are what makes a CRM stop being trusted.
Building the reporting last. Reports depend on data captured from day one. Deciding what to measure after six months of use means six months of records missing the field that mattered.
Migrating everything. Bringing over a decade of dead records makes the new system feel like the old one. Decide what does not come.
Questions To Ask Whoever Builds It
Fair questions for us as much as anyone.
- What happens to the data if we stop working with you? You should be able to export everything, including notes and attachments, without help.
- Which integrations are you confident about, and which are you guessing at? An honest answer separates the two.
- How will we know whether people are actually using it? There should be a way to see adoption rather than assume it.
- What is the smallest version that would be useful? If the answer is the whole thing, the scoping has not been done.
- What are you deliberately not building? Every good scope has a list.
- Who trains the team, and when? Adoption is the risk; this is where it is addressed or not.
A Brief You Can Send Anyone
Copy this, fill it in, send it to whoever you are considering. It will get you comparable quotes rather than three proposals answering different questions.
What we do: [what the business sells, and to whom]
How we track it now: [current tools, including the spreadsheets]
What breaks: [the specific thing that fails today]
Who uses it: [roles, and what each needs to see]
What it must connect to: [systems, and whether they have APIs]
What we have to migrate: [how many records, from where, how far back]
What we do not need: [be specific; this controls the price]
How we will know it worked: [one measurable thing]
Built On
Laravel for the backend, business logic, access control and the integration layer. React for the interface, where data-dense views need to stay usable. PostgreSQL for relational storage and reporting queries. Redis for caching frequently opened records. Pusher for live updates so two people never see different versions of a record.
Talk To Us About Your Client Relationships
If the team works around the CRM more than in it, get in touch and we will map the relationship model and scope something that fits. If configuring what you already have would serve you better, we will say so.
This sits alongside the systems we build and the CRM system page covers the delivered product rather than the engagement.