Picture a forty-person business. It has an accounting package, a CRM, a shared inbox, a stock system, a job-scheduling tool, three WhatsApp groups and a spreadsheet that connects all of them.
Every one of those purchases was reasonable. Each solved a real problem on the day it was bought. Together they form a workflow that nobody designed — one where the same customer's details live in four places and "can you check…?" is the most common message in the building.
When the next problem comes along, the default answer is another subscription. This article is about when that is the right answer, and when it is time for software that fits the business instead.
The problem is not SaaS
Be clear about this first: most of your software should be bought.
Accounting, payroll, email, file storage, HR — these processes are close to identical from one business to the next, heavily regulated or both. Mature products handle them better than anything you would build, and someone else keeps them compliant. Building your own accounting system would be a waste of money.
The problem is not SaaS itself. It is using SaaS for the one or two workflows that are not standard in your business — the way you take orders, schedule jobs or handle requests — and then filling the gaps between tools with people.
Count handoffs, not tools
The number of subscriptions is a distraction. The number that matters is handoffs: each point where work moves between people or systems by re-typing, copying and pasting, forwarding an email or asking someone to check something.
Try this with one real order, job or request from last week:
- Write down every step from the moment it arrived until it was finished and invoiced.
- Next to each step, note the tool and the person.
- Circle every point where information was entered again, or where someone had to ask someone else.
The circles are the cost. They are where time goes, where errors get in, and where work waits. Our homepage shows what this typically looks like for a single order — six steps and five handoffs before anyone invoices.
With that picture in front of you, there is one question to ask of any new tool: does it remove a handoff, or add one? A new product that needs its own data entry and its own login, and still needs the spreadsheet beside it, has added one — however good its feature list is.
A four-question test for each workflow
Do not make one decision for the whole business. Make it workflow by workflow, with four questions.
- Is this process standard in your industry, or is it how you compete? If a competitor could run it identically without anyone noticing, it is standard. If your pricing, turnaround or service depends on doing it your way, it is part of how you compete.
- Would a product fit without a spreadsheet beside it? Trial the tool on your messiest real example, not the vendor's demo data.
- How many other systems does this data need to reach? An order that must reach stock, purchasing, the customer and accounting is an integration problem as much as a software problem.
- What happens at your next growth step? Double the volume, add a second site, take on a larger customer. Does the tool's pricing, and your workaround, still hold?
| What you find | What to do |
|---|---|
| Standard process, a product fits cleanly | Buy it. Configure it. Move on. |
| Standard process, but the data must reach several other systems | Buy it, and invest in integrating it properly rather than re-keying. |
| How you compete, and every product needs a workaround | This is the workflow to build around. |
| Not sure yet, and the process is still changing | Wait. Settle the process before choosing software for it. |
What "software that fits" actually means
Software that fits is not a rebuild of everything you use. It is a Custom Operations System at the centre of the one workflow that is genuinely yours — orders, customers, purchasing, inventory, a dashboard, some automation — connected to the standard tools you keep, such as your accounting package.
Three things make it practical for a business of 10–100 people rather than an enterprise project:
- It starts with the workflow. A discovery such as our AI Workflow Blueprint maps how work really moves before anything is designed. What that involves, and when it should conclude "don't build", is covered in What's in an AI Workflow Blueprint.
- It is built in phases. The part that removes the most handoffs comes first and is usable on its own. Initial builds are typically £3,000–£10,000, not a single large leap.
- You own it. Code, infrastructure and documentation are yours, so the system cannot be repriced at renewal or switched off by someone else.
The honest costs on both sides
Neither route is free, and it is worth being clear-eyed about both.
Another subscription costs more than its pricing page suggests. Per-seat pricing grows with your headcount. Workarounds cost staff time every day. And leaving is harder than joining — our article on what you actually own when you try to leave a supplier covers what to check before you sign.
Software that fits has an up-front build cost, needs someone inside the business to own it, and will need changes as you grow. Hosting and ongoing changes can be handled by an optional Care plan, or by your own team, since you have the code. Our piece on why nobody can price your software from a blog post explains where custom software costs actually come from.
The useful comparison is not subscription price against build price. It is the total cost of the workflow as it runs today — including the handoffs — against the cost of running it in a system designed for it.
The pattern we recommend
For most operational businesses in the 10–100 person range, the sensible shape is:
- Keep buying the commodity tools: accounting, payroll, email, documents.
- Build around the one core workflow that is how you compete — often the path of an order, a job or a request through the business.
- Integrate the two, so information is entered once.
For a worked example in one sector, see Your Wholesale Order Book Is Still Email and Excel, or our page on custom operations software for wholesale and distribution.
A sensible next step
Trace one workflow and count the handoffs. If the circles are few, you probably need a better tool, or no change at all. If they are many — and the same information keeps being typed into different systems — the answer is unlikely to be another login.
Either way, you are welcome to show us your workflow. We will tell you honestly whether building around it makes sense, or whether you would be better off buying. Or read more about custom operations software built around how your team already works.
