When work slows down, the first question is usually "do we need new software?" A more useful question is "which part of our system is failing, and why?" The answer decides whether you should keep what you have, connect it to something else, improve it, or build something new.
Start with the work, not the software
Pick the three processes that matter most to the business. For most organisations they are collecting money, serving customers and reporting to management. For each one, follow a single real example from start to finish: one fee payment, one hotel booking, one delivery.
Write down every place the information is written, typed, copied or checked. Ask the people who do the work, not only their managers. The places where staff copy information by hand, wait for someone, or keep a private spreadsheet are where the problems are.
The four options
Keep
The tool does the job; people are not using it as intended. Ask: does the feature we need already exist? Has anyone been trained on it? Is the problem a habit rather than the software?
The fix is usually training, a short written guide or a small settings change. This is the cheapest option, and it is more common than vendors like to admit.
Connect
Two tools each work, but staff copy information between them. Ask: can each tool export and import data, or connect directly to other software through an API? Which of the two should be the "source of truth" for each piece of information?
Typical examples are payments and accounts, a booking system and the restaurant till, or a stock system and the online shop. Connecting them removes the re-typing and the mismatched numbers that come with it.
Improve
The system nearly fits. Ask: can it be configured to match how we work? Is the data clean? Could one repeated step be automated? Are the reports built on the right information?
Improvement often means cleaning up the data, adding a missing workflow step, automating receipts or reminders, or rebuilding a report that management actually reads.
Build
Build only when you have confirmed that configuring or connecting what you have cannot meet the requirement at a reasonable cost. Ask: is the process stable enough to build around? Who will maintain the system in three years? What will it cost to run, not only to build?
Building is the right answer when there is no system at all, when spreadsheets have clearly hit their limits, or when the software cannot be adapted. Even then, build the most important part first.
A quick test
| What you notice | Likely option |
|---|---|
| The feature exists but nobody uses it | Keep, with training |
| Staff re-type the same information into two tools | Connect |
| Numbers from two departments never agree | Connect, or improve the data |
| A spreadsheet sits beside the "real" system | Improve |
| Month-end reports are put together by hand | Improve the reports, or connect the sources |
| The core process is not supported at all | Build, starting small |
Costs people forget
- Cleaning the data. Moving messy records into a new system moves the mess too.
- Training. Every role needs to learn its own screens, and new staff need it again later.
- Running costs. Hosting, licences, internet, backups and support continue after launch.
- The changeover. For a period, staff may need to run old and new side by side.
Questions to ask any supplier, including us
- Who owns our data, and how do we export all of it?
- What happens if we stop paying or you stop trading?
- What does it cost to run each year, not only to build?
- Who trains our staff, and what documentation do we get?
- What are your support hours and response times, in writing?
- Will it keep working when the internet or an AI service is down?
Want a second opinion?
On a free 30-minute call we can go through one of your processes together and suggest which option fits.