TL;DR: in 2026 the decision rule fits in one sentence: if an off-the-shelf CRM such as Pipedrive (from €14 per user per month) covers at least 80% of your sales process without workarounds, buy it; a custom CRM is only justified when workarounds keep piling up, when per-seat licences exceed the three-year cost of building, or when your data requires hosting you genuinely control.

Key takeaways

The short answer: buy first, build when it earns its keep

Always start by looking at the market. Pipedrive, HubSpot, Zoho or Brevo handle a classic sales process very well: contacts, deals moving through a pipeline, follow-ups, reports. If the way you sell fits that mould at 80% or more, a well-configured off-the-shelf tool will be cheaper, live sooner and easier to run than a custom build. A Pipedrive rollout takes one to two weeks, where a full Salesforce implementation runs 8 to 16 weeks according to Cloud Koala.

A custom CRM enters the picture when the remaining 20% is precisely what sets you apart. Multichannel intake that standard tools do not model, an unusual business process, sensitive data that dictates where hosting must sit, or a team growing to the point where the licence bill exceeds the cost of software you would own outright. In those cases, custom is not a luxury; it is the rational calculation.

Why 70% of CRM projects fail, and what it changes about the choice

The figure deserves a hard look: according to DemandSage, around 70% of CRM projects miss their goals, and the dominant cause is adoption, not technology. A CRM the sales team never opens returns exactly zero on investment, whether the seat costs €14 or €165.

That data point reframes the choice. The right tool is not the most complete or the best known; it is the one that matches the team's real work closely enough to be opened every morning. A poorly fitting off-the-shelf tool breeds workarounds: parallel Excel files, double entry, fields hijacked for other purposes. Every workaround is a signal. One or two can be fixed with configuration; a pile-up means the tool does not model your business, and that is where the custom question becomes legitimate.

When an off-the-shelf CRM is the right call

Four situations where the honest answer is "buy a market tool and have it configured properly":

When building custom becomes rational

The tipping point shows up in three columns: workarounds, licences, data.

Workarounds. When the team maintains parallel spreadsheets for what the CRM "cannot do", the real cost of the market tool is no longer its subscription; it is the double-entry time and the errors that come with it. DemandSage notes that a CRM in actual use returns on average $3 to $5 per dollar invested; a CRM worked around flips that formula.

Licences. Run the maths over three years. At €100 per user per month (the price of HubSpot Sales Hub Pro according to Cloud Koala), a 20-person team pays €24,000 a year, or €72,000 over three years, before any setup work; onboarding an off-the-shelf CRM for a 50-person SME is estimated from €15,000 (Pipedrive) to €150,000 (Salesforce) by the same source. A custom CRM should be compared against that total, not against the build quote alone; I detailed the ranges in How much does a custom CRM cost in 2026. With no per-seat licence, each new hire stops inflating the bill.

Data. Some activities handle information that cannot sit on just any cloud: healthcare, child protection, legal, finance. An off-the-shelf tool imposes its vendor's hosting; a custom build deploys on servers you choose, in France if required, with credentials you hold.

Five criteria to settle it

  1. Process coverage: does the market tool model the way you sell without major workarounds? Below 80%, take a serious look at custom.
  2. Team trajectory: multiply the per-seat price by the headcount you expect in three years, not today's.
  3. Channels: phone, chat, forms, messaging apps. The more multichannel your customer relationship, the less standard tools keep up without costly add-ons.
  4. Data requirements: hosting imposed by the vendor or hosting chosen by you; the answer sometimes eliminates a whole family of options.
  5. Reversibility: in both cases, check what you keep when you leave. Full data export on the market side; code ownership and documentation on the custom side.

What I see in the field

Both answers coexist in my own work, and that is the point. For Horus Condition Report, a company operating in French and English, the right answer was Pipedrive: a careful bilingual configuration, not a single line of custom code. The process fitted the mould; building would have been an indefensible expense.

By contrast, on the 3018 helpline, France's national number against cyberbullying, no market tool modelled the actual work: 12 counsellors and managers, 5 channels handled in a single interface, built-in telephony, and minors' data that required hosting in France. The Emma CRM, which I co-built, exists because custom was the only serious answer to that brief. Two situations, two opposite verdicts, one method: start from the process and the constraints, never from the tool.

The final test fits in one question: does your current CRM describe your business, or is your business contorting itself to fit its CRM? As long as the answer is the former, keep it and tune it. The day it becomes the latter, custom stops being an expense and becomes a catch-up.

A simple check to run this week: ask your team how many Excel files live alongside the CRM, and what they contain. If the inventory surprises you, a 30-minute scoping call is enough to set the diagnosis; a scoped intervention like the Sprint starts at €400, the rest is quoted per project. The full approach is on the Custom CRM page.

Frequently asked questions

What is the main criterion for choosing between an off-the-shelf CRM and a custom CRM?

How much of your actual process the tool covers without workarounds. If an off-the-shelf CRM such as Pipedrive, HubSpot or Zoho covers at least 80% of the way you sell with standard configuration, it almost always wins on cost and speed. A custom CRM makes sense when the remaining 20% is precisely what sets you apart: multichannel intake, an unusual process, or sensitive data.

At how many users does a custom CRM become cheaper?

There is no magic threshold, but the maths is simple: at €100 per user per month, a 20-person team pays €24,000 a year in licences, or €72,000 over three years before any setup work. A custom CRM should be compared against that total over time, maintenance included. The bigger the team gets, the wider the gap in favour of custom, since there is no per-seat licence.

Can you start with an off-the-shelf CRM and move to custom later?

Yes, and it is the most common path. A well-configured off-the-shelf CRM stabilises the sales process and builds a clean data history. When workarounds start piling up, that history can be exported and migrated to the custom tool. Starting with the market is never wasted money if the data is kept clean.

Is a custom CRM a risky bet for an SME?

The main risk is scope, not technology. The prudent method is to ship a first version covering the most critical journey within a few weeks, then iterate with the team. Requiring ownership of the code and documentation also protects the SME: it can bring maintenance in-house or switch providers at any time.