Off-the-Shelf CRM or Custom Software?

"Should we buy an off-the-shelf CRM, or have one built for us?" I hear this question far more often than I did ten years ago. The reason is simple: ready-made products got more expensive, per-user licensing became the norm, while the cost of custom development went down. But price is not the whole story.

I am writing this to explain how the decision should be made, not to sell custom software. The right answer is not "always custom". The wrong answer is "always off-the-shelf".

The short answer

If your process is standard — if you work the way everyone else in your industry works — a ready-made product is faster and cheaper. There is nothing to argue about there. A ready-made product carries decades of accumulated practice from your industry; you are not starting from zero.

But if any of these four exist, you reach the limits of an off-the-shelf product very quickly: a multi-branch structure, commission or bonus rules specific to you, your own listing/portal site, and a real integration with your existing ERP. These are the places ready-made products wave away as "custom fields" or "plugins", and they are usually the most expensive and most fragile part of the product.

Off-the-shelf is genuinely good — when

• Your process is standard. You quote, invoice and report the way your industry does.
• You need to start now. If you need to register customers this week, a system that takes six months will not save you.
• Your team is small. For a sales team of three, custom software is often an oversized investment.
• Your process is unlikely to change. You know your workflows will not be turned upside down in the next two years.

Six signs you have hit the limit

1. "We cannot customise that field." Your need is not on the vendor's roadmap. You open a ticket and get "under consideration". Your schedule is no longer yours.
2. Back to Excel. You export data and calculate outside the system. That means your real process lives next to the software, not inside it.
3. Per-user licensing grows. Every field user, every branch, sometimes even a manager who only views reports needs a seat. As you grow, cost multiplies.
4. Integration is "a consultant job". Connecting to your ERP is done by a separate consultant, who comes back with every update.
5. The report never arrives. What management needs is not in the standard reports, and custom reporting is a separate fee.
6. Your data cannot leave. If you change systems, your portfolio, history and notes cannot be exported, or arrive as a meaningless spreadsheet. Locked data means locked customers.

The hidden cost: adaptation and dependency

Ready-made pricing is usually discussed as licence fees. Real cost has three parts, and two of them are rarely mentioned: licence, adaptation (custom development or consulting for what the product cannot do), and dependency (what happens if the price doubles, the product is sold, or the platform shuts down).

Compare three-year total cost of ownership, not monthly licence vs development price: direct payments, adaptation, integration, migration and training, and the opportunity cost of the work your system cannot do — the Excel hours, the double entry, the customers you lose.

Open source: the code is yours, so is the maintenance

Open source looks like an attractive middle ground: no licence fee, full control. In theory, yes. In practice it starts with one question: who maintains the code you changed? Security patches, custom plugins breaking on upgrade, a roadmap that diverges from your needs. Open source sets the licence to zero; it does not set responsibility to zero. If you can carry that with a team and a budget, it is an excellent option. If not, a free product can take you to the most expensive place of all: an unmaintained system nobody fully understands.

Why we no longer build sites on WordPress

This site used to be WordPress. So this is not a rival-technology pitch. WordPress is still excellent for blogs and publishing. The picture changes when you put a company site, a product front door or a customer portal on top of it.

1. Plugin pile-up. Speed, forms, SEO, backups, security, caching — each need adds a plugin. A site with twenty plugins runs code from twenty teams on the same page.
2. Security surface. Plugins and themes need endless updates. Skip them and you stay exposed; apply them and the site may break. The dilemma repeats every month.
3. Performance. A site "sped up" by caching plugins is a slow core with paint on top. Mobile load time is a ranking signal for search engines and AI answers alike.
4. Design limits. The theme is not yours. Change a section and you risk losing it on the next theme update.
5. Structured data and GEO. To be visible in AI search, pages should be produced server-side with clean HTML and correct schema. Plugin-injected schema is often incomplete.

None of this means "WordPress does not work". It means WordPress becomes a load you do not control once your business grows. Acceptable for a blog. Not for a product's main site. That is why we moved to our own stack: Laravel, database-driven content, no plugins.

What AI changed, and what it did not

The real reason for this article: over the last two years, the cost of custom software has fallen much faster than most businesses realise. Not because of a magic tool, but because the repetitive part of the work shrank.

What changed: first drafts of code, screens and data models arrive much faster; testing, documentation and small fixes take less time; connecting to third-party APIs went from days to hours; and as a result, a small system built exactly for you is no longer a large-company budget item.

What did not change: scope (deciding what to build is still human work), architecture and data modelling (the decisions that keep a system alive in three years), security (authorisation, data retention, backups, audit trails), and responsibility (when something breaks, you cannot say "the AI did it").

In short, AI took custom software out of the "luxury" category. It did not make it automatic.

Real estate: where an off-the-shelf CRM hits the wall

Real estate is one of the clearest places to see this decision. For a single office working on standard commission, a ready-made real-estate CRM works well: listings, clients, ads, call notes.

The limit appears when you need:

• A multi-branch structure. Is the agent's pool branch-based or central? How is commission split when a client moves between branches? Many products answer with a "branch" field, not a rules engine.
• Commission and bonus rules. Tiered commission, team bonus, the share of the agent who brought the listing, target bonuses. These differ by company; the product ships with its own assumptions.
• Your own listing site. Publishing your portfolio on your own domain with your own SEO structure, pushing ads to portals automatically, turning enquiries into CRM records, closing listings automatically.
• Integration with your ERP. Accounting, e-invoicing, collection and contracts working on the same data. Double entry means both errors and lost time.
• A field team. Adding a listing from mobile, uploading photos, taking notes on site. A desktop interface used on a phone does not get work done.

When all five are present, the money and time you would spend customising a ready-made product approaches the cost of building your own. The difference: in the first case you end up owning custom fields of a product that is not yours; in the second, the system is yours.

Decision checklist

1. How much of our process is standard, and how much is specific to us?
2. How many hours a week do we lose on work the ready-made product cannot do?
3. How many users and branches will we have in three years?
4. Which systems must we integrate with, and how often do they change?
5. What happens if we need to take our data out?
6. What do we do if the product's price doubles?
7. Who will maintain this system in three years?

How the process works with us

1. Scope workshop. We discuss your current process and where the ready-made product blocks you. Output: written scope, screen list, out-of-scope list.
2. Clickable prototype. You see the screens before code is written. Fixing a misunderstanding here is far cheaper than during development.
3. Staged delivery. Every stage ends with something that works: login and listings, then the ad flow, then integrations. Payment follows delivered work, not the calendar.
4. Data and handover. Migration from your current system is planned. When it is done, the code, database and documentation are yours.
5. Maintenance. Bug fixes and support after launch, then an annual maintenance agreement.

We use AI at every step — to speed up first drafts, generate test scenarios, write integration code, produce documentation. But scope, architecture and responsibility are owned by a person. Your contact is not a tool.

In conclusion

"Off-the-shelf or custom" is a calculation, not an ideology. For a standard process, ready-made is sensible. But with a multi-branch structure, custom commission rules, your own listing site or ERP integration, the limit arrives fast. At that point the right question is no longer "which product", but "with what scope do we build our own system".

You can see how this looks for real estate on the industry page, find similar approaches under other industries, and see concrete work in projects. Tell me about your process and we will work out together whether you should stay with a ready-made product or build your own — and I will tell you whichever it is.

Get in touch and let us talk.