We build custom CRM platforms, so treat the source accordingly. The useful version of this comparison is the one that says where off-the-shelf wins, because that is the part a vendor has an incentive to skip — and if you cannot trust the parts that cost us business, there is no reason to trust the rest.
Off-the-shelf CRM is the right answer more often than people in our position tend to admit.
What off-the-shelf CRM is genuinely good at
Speed to something working. You can be operational in days. A custom platform is weeks at minimum, and the gap is real cost while you wait.
Being finished. Someone else has already thought about email sync, mobile apps, permissions, deduplication, imports, and the hundred small features that are boring to build and painful to lack. That accumulated completeness is the actual product, and it is worth a great deal.
Ecosystem. Integrations exist because thousands of customers wanted them. Your accountant has probably used the same tool. Hiring someone who already knows it is straightforward.
Cost at low seat counts. Five people on a mainstream CRM is inexpensive. There is no scenario where building beats that on price alone.
Not being your problem. Someone else patches it, backs it up, and keeps it running. That has genuine value, particularly if nobody internally wants to own software.
If your process resembles the process the product was designed around, buy it. Configure it well, adopt it properly, and spend your energy on the business. A well-fitting off-the-shelf CRM your team actually uses beats a custom platform nobody adopts, every time.
Where off-the-shelf starts costing more than it saves
The failure mode is not dramatic. It is gradual, and it shows up in three ways.
A spreadsheet appears next to the CRM. Someone builds it to hold the part the CRM could not. That spreadsheet is the most useful diagnostic in your business: its columns are exactly the shape of the mismatch. When decisions get made from the spreadsheet rather than the CRM, the CRM has become a contact database with a subscription.
Custom fields stand in for your real unit of work. Mainstream CRM is built around a sales deal moving through stages to a close. If your business runs on work packages, field tickets, lots, matters, route stops, or loads, you will be modelling those as deals with custom fields. It works, in the sense that the data goes in. What breaks is everything downstream — reporting, automation, and permissions all assume a deal, so you fight the grain of the product on every workflow.
The per-seat bill outgrows the value. At five seats, nobody cares. At forty — including field staff who need read access and nothing else — the arithmetic changes. What matters is not the absolute cost but the ratio: if the bill grows with headcount while value per seat does not, that gap widens forever.
One of these is a configuration problem. Two or three together is a fit problem, and configuration will not fix it.
What custom actually buys you
Not features. Anything a custom platform does could in principle be assembled from products plus integrations. What custom buys is a data model that matches your business, and the consequences of that.
Your unit of work is a first-class object. A work package is a work package, with the fields, states, and permissions a work package needs. Reporting is a query rather than an export.
The operations live in the same system as the customer record. This is the part that is genuinely hard to buy. Takeoff, dispatch, project control, client portals, document control — these usually live in separate tools with their own identifiers, and reconciling them is permanent overhead. In one system they are one record. Our operating system pages go into what that means in practice.
AI has something to work with. An AI layer is only as good as its access. Bolted onto a CRM that holds a third of your operational reality, it can summarise what it can see. Sitting on a platform that holds all of it, it can answer questions that would otherwise take three exports and a pivot table.
No per-seat ceiling on who gets access. Give the whole field crew read access without a budget conversation. That changes what the software is for.
You own it. Covered below, because it is a separate argument that people conflate with this one.
The honest costs of custom
It takes longer. Weeks to months depending on scope. During that time you are running on whatever you have now.
It costs more up front. The comparison is a capital cost against a subscription, and over a short enough horizon the subscription always wins. Over a long enough one it does not. Where the lines cross depends on seat count and how much the mismatch is costing you — which is why the spreadsheet question above is worth answering with a real number.
Someone has to maintain it. Usually the studio that built it, on a retainer. That is not free and should not be presented as free. The honest comparison is retainer versus subscription, not retainer versus zero.
You can specify it badly. A custom platform built around a process you were about to change is an expensive mistake that an off-the-shelf product would have absorbed. This is a real risk and the reason we spend time on discovery before writing anything.
It is not finished on day one. You will miss features the mature product had. Most do not matter. Some do, and you will notice.
A decision framework that does not favour us
Score these honestly.
- Is your core unit of work a sales deal moving through stages? → Off-the-shelf.
- Does a spreadsheet hold the part the CRM cannot? → Point toward custom.
- Do your operations live in tools that do not talk to the CRM? → Point toward custom.
- Is per-seat cost outgrowing per-seat value? → Point toward custom.
- Do you need something working in under a month? → Off-the-shelf, at least for now.
- Will anyone internally own the platform? → If not, off-the-shelf.
- Is your process about to change substantially? → Wait, whichever way you lean.
Mostly ones, fives, and sixes: buy the product and configure it properly. Mostly twos, threes, and fours: the mismatch is already costing you, and it compounds.
The middle path people forget
It is not binary. A common and sensible pattern is to keep an off-the-shelf CRM for what it is good at and build custom only for the operational layer it cannot model — with the two integrated. You get the mature contact and email tooling and a data model that fits your actual work.
This is often the right first step, and it is cheaper than either extreme. Anyone who tells you it is all-or-nothing is describing their business model rather than your options.
If you want the industry-specific version of this question, the construction, oil and gas, and home building pages each go into what the mismatch looks like in that vertical — because "your unit of work is not a deal" means something different when the unit is a work package, a well, or a lot.