You know the feeling. The SaaS renewal notice lands in your inbox and you stare at it for a second too long. Not because it's expensive (though it probably is) but because you're suddenly doing the maths in your head. You're paying for five tools. Your team uses three of them reluctantly, one of them constantly works around, and the fifth only exists because tool number two can't talk to tool number three without it. You've got Zapier duct-taping the whole thing together. Half your team's time goes into re-keying data that should flow automatically.
And somewhere in the back of your mind, the thought surfaces: should we just build something that actually fits?
This post is the honest answer to that question. Spoiler: for most businesses, SaaS is still the right call. But there's a specific set of conditions under which a custom build wins, and understanding them is what makes the decision.
The honest truth: SaaS usually wins
It's worth saying this clearly, because you're reading a post on a custom software consultancy's website and you should be suspicious of the framing.
Most businesses should be using off-the-shelf SaaS. The economics are usually right: low upfront cost, predictable monthly fee, the vendor handles security patches and infrastructure, and the software matures over time as other customers give feedback. For standard business workflows (invoicing, CRM, project management, HR) the established SaaS options are genuinely good. If your workflow mostly fits Xero, or HubSpot, or ClickUp, use it. Don't build something custom to save a few hundred pounds a month and inherit the maintenance burden.
Custom software for UK small businesses is the right answer in specific situations. The rest of this post is about identifying those situations. If none of them apply to you, the honest recommendation is: stick with SaaS, possibly do a workflow audit to consolidate what you're using, and save the custom build budget for later.
When SaaS stops being enough
Here are the signals that your situation has moved out of "SaaS is fine" territory:
- You're paying per seat for users who barely use the tool. If you've got a 20-seat licence for a tool that 12 people use properly and eight were added because the workflow requires it but don't actually engage with it, you're subsidising the vendor's growth model.
- Your workflow genuinely doesn't match any template. Not "the template is inconvenient", but "the way your business operates is structurally different from the assumptions the software was built on". A specialist trade with an unusual quoting and invoicing flow, for example. A professional services firm that runs projects in a way that no PM tool has ever quite captured.
- You're on tool number eight and they still don't talk to each other. Every new tool is a new integration problem. Every integration is a new failure point. At some number of tools (the number varies by business but it's usually somewhere between five and eight) the overhead of managing the integrations starts to outweigh the value of any individual tool.
- Exports are manual spreadsheets. If the path from data in your system to a usable report for a business decision goes through a manual download-clean-reformat-upload cycle, you've already outgrown the SaaS stack you're on.
- Compliance or data requirements the vendor won't support. Regulated industries have requirements that generic SaaS vendors often won't accommodate: specific data residency, audit trail formats, custom access controls, integration with regulated third parties. When the vendor's answer to your compliance question is "that's on our enterprise roadmap", you're not getting it.
- Data sovereignty concerns. Some businesses, for reasons ranging from contractual to competitive, don't want their operational data sitting on a US-headquartered SaaS platform's servers. A custom build gives you full control over where data lives.
When custom software is actually worth it
Four specific scenarios where a custom build wins in the UK small business context:
Replacing three or more SaaS subscriptions that together cost £5,000+ per year. The break-even maths here are straightforward (more on that below). If you're paying £400/month across four tools that together cover one operational function, a one-time custom build that replaces them entirely often pays for itself within two years, and continues to save money every year after that. A specialist UK recruitment operation that Orchestrix audited was running eight disconnected systems for their compliance and onboarding workflow. After consolidation into a single custom HRIS, the saving was £24,520/year and 14 hours/week per worker recovered. The subscription bill disappeared. The admin hours didn't come back.
A workflow so specific that no SaaS exists for it. Some businesses have genuinely unusual operational models. A bespoke manufacturer with a complex quoting workflow tied to materials costs and lead times. A specialist training provider with accreditation tracking requirements that no off-the-shelf LMS handles correctly. When you've evaluated five SaaS options, read the case studies, and every one of them is wrong in the same fundamental way, the custom build conversation is worth having.
A customer-facing product or portal where the software is the service. If your clients need to log in somewhere to see their account, track a job, access documents, or communicate with your team, and you're doing this via email threads and manually-shared Dropbox folders, a custom portal is often the right answer. Not because it's cheaper than SaaS (it probably isn't, in year one), but because it's the product. The portal is the experience your clients have of you. Off-the-shelf portals built for generic use cases will always look and feel like someone else's product with your logo on it.
Data or compliance requirements that force control over the stack. If your business operates in a regulated sector (financial services, legal, medical, but not CQC-regulated care, which isn't Orchestrix's domain) and you need specific audit trails, access controls, or data handling guarantees, a custom build is sometimes the only way to meet those requirements without paying enterprise SaaS prices.
The 18-month rule
Here's the maths, stated as plainly as possible.
A custom software build is a one-time capital cost. A SaaS subscription is an ongoing revenue cost. The question is whether the capital cost is lower than the cumulative subscription cost over a reasonable time horizon.
Some examples using real ballpark numbers:
- A £15,000 custom build replacing a SaaS stack that costs £500/month breaks even at month 30. After that, the custom build is free every month.
- A £15,000 custom build replacing a SaaS stack that costs £1,000/month breaks even at month 15.
- A £25,000 custom build replacing a SaaS stack that costs £1,500/month breaks even at month 17.
The rule of thumb: if the combined monthly SaaS cost for the function being replaced is above £700/month, a properly-scoped custom build often breaks even within two years. Below £300/month, the maths rarely work out.
There are a few things this maths doesn't capture. Custom software has ongoing costs: it'll need maintenance, dependency updates, and the occasional bug fix. A managed retainer handles this for a fixed monthly fee, but it's a real cost and the comparison isn't always as clean as the numbers above suggest. On the other hand, SaaS costs escalate. Most vendors increase prices annually, add per-seat costs as you grow, or change their pricing tiers. The £500/month tool today is often £750/month in three years.
The honest version of the 18-month rule: if the break-even is under 24 months and you have good reason to believe the underlying workflow isn't going to change significantly, a custom build is worth serious consideration.
What a custom build actually includes
Worth being specific, because "custom software" can mean anything from a 50-line script to a multi-year platform project. What Orchestrix builds for the UK small business market is web applications: browser-based tools, no installer required.
A standard custom software build includes:
- A full web application with authentication and role-based permissions, designed around how your team is actually structured, not a generic permissions model
- A PostgreSQL database schema you own, with migrations documented
- A deployment pipeline so updates aren't a manual process
- User documentation written for the people who use it, not for a developer
- Source code in a repository you control: GitHub or GitLab, full history, no lock-in
- A handover session with a screen recording covering the architecture and how to extend it
- A 30-day bug-fix window at no additional cost
The build starts at £1,000 for a small focused tool, and most SMB projects land between £3,000 and £15,000 depending on scope. Larger applications with more complex data models and permission structures sit above that. Every project is fixed-fee and scoped after a workflow audit, so you have a firm number before a line of code is written.
What you lose by going custom
This section exists because most custom software pitches skip it. Here's what you're signing up for when you choose a custom build:
No vendor patching. The SaaS vendor patches their software automatically. With a custom build, someone has to stay on top of dependency updates, security patches, and infrastructure maintenance. That's either your own developer, an Orchestrix managed retainer, or a future technical hire. It's real work and it doesn't go away.
No vendor support SLA. If a SaaS tool goes down, there's a support team and a status page. If your custom build has an outage at 11pm on a Friday, the resolution depends on whoever is responsible for maintaining it.
Slower feature additions. SaaS vendors ship features continuously, funded by their whole user base. Custom software evolves at the pace you're willing to fund. That's not necessarily a problem (if the software does exactly what you need today, you don't need it to evolve) but it's a different mental model from "new features arrive in the product update email".
You carry the maintenance cost. It won't be expensive if the build was done properly: a well-built application on a stable stack doesn't require constant attention. But it's not zero, and it's worth being honest about that in the build vs. subscription comparison.
How to decide in 15 minutes
Run through these questions honestly. If you answer yes to four or more, the custom software conversation is probably worth having:
- Are you paying more than £600/month combined across SaaS tools for one operational function?
- Have you switched tools in this category in the last two years and still aren't happy?
- Does your team describe the current tool with a swear word at least weekly? (This one's diagnostic, not decisive, but it correlates.)
- Does your workflow require a manual data export or re-entry step at least once a week?
- Have you looked at five or more SaaS options in this category and found all of them wrong in the same fundamental way?
- Does your business have data handling or compliance requirements the vendor explicitly won't support?
- Is the software itself part of what your clients pay for?
If you got to yes on four or more of those: a workflow audit is the right next step. Not a build: an audit first. The audit maps the workflow properly, identifies what actually needs to be built, and produces a fixed-fee proposal for the build. You can act on the audit output regardless of whether you proceed with Orchestrix.
If you answered yes to one or two: there's probably a SaaS consolidation or business automation answer that's cheaper and lower-risk than a full custom build.
The decision isn't permanent
One last thing worth saying. Custom software vs SaaS isn't a one-time lifetime choice. A business that's correctly on SaaS today might be the right candidate for a custom build in three years as its workflows mature and its subscription bill grows. And a business that builds something custom today might, in a different part of its operations, be perfectly well served by an off-the-shelf tool.
The question to answer is always: for this specific operational problem, at this specific moment in the business's growth, what's the highest-value solution? Sometimes that's SaaS. Sometimes it's automation. Sometimes it's a custom build. Occasionally it's "leave it alone, it's fine".
If you're genuinely at the decision point and want an independent view on which answer applies to your situation, start with a 15-minute triage. Bring the subscription bill, describe the workflow, and you'll get an honest assessment of where you are, and whether the custom software vs SaaS question even needs answering right now.