If a tool exists that does your work well enough, use it. What remains is finding out whether such a tool exists and what “well enough” means for you.
By an existing tool I mean software that is already finished and that you usually use on a subscription, like an accounting package or a calendar where customers book appointments. Custom software is built for your business, around how you work.
What an existing tool gives you
You can start with it today and you can try it before you decide. Its makers take care of updates, security and the servers it runs on. If other businesses already use it, most of the teething problems are gone too.
Say you are a hairdresser and you want customers to book appointments online. Every hairdresser has roughly the same need, and tools exist for it. Something custom would add little.
On the other side, you are renting the product. Its makers decide on the price, on new features and on what disappears. You adapt your work to the tool, and that goes smoothly until you want something it cannot do. Check as well whether you can get your data out if you ever want to leave.
When custom software is worth it
Custom software fits when your way of working differs from what the tool assumes, and you want to keep that difference. Say an installer calculates quotes with his own rules, grown over the years, and those rules are exactly what he is good at. A standard package forces him to simplify them or to keep a spreadsheet next to it.
It can also be work spread across several tools, where someone retypes data from one into the other. If that comes back every week and takes a quarter of an hour or an hour each time, you are looking at the signs from When has your spreadsheet grown too big? Sometimes a connection between those tools is then enough.
If the software is itself your product, there is little to rent. If you have an idea for an app you want to offer to customers, it still has to be made, and the first question is what goes in a first version.
Custom software brings costs of its own. Besides the building itself there is the maintenance, the hosting (the servers the software runs on) and someone who knows the code when something has to change. In return you get software that works the way your business works, and that you can have changed when your work changes.
Combining them or moving later
You can also combine the two: you keep the tool and have a piece built next to it. Felix is set up that way: a boutique keeps working with its own till, and Felix connects to it, looks at who bought what and when, and suggests which customers the owner should contact.
That works if the tool has an API, an agreed way for other software to talk to it. How that goes is explained in Connect your software to your till and your accounting. For simple cases there are connector tools like Zapier or Make, which link two programs without anyone having to program.
You can also start with a tool and move later. That way you learn what you really need, and you already have data and experience when you have something built. Do check beforehand whether you can export that data in a usable form.
How to make the choice
- Describe the work in steps, from the point of view of the person doing it: I open this, I press this, then this should appear.
- Find a few tools that promise to do that work and try them with an ordinary working week. If a tool has a trial period, really use it.
- Note every place where it doesn’t fit, and write down how often it bothers you. An annoyance once a year weighs differently from something that costs you time every week.
- Compare over several years. A subscription keeps running and can rise with the number of users. Custom software costs most at the start, and maintenance remains after that.
If you come to me with this question, an existing tool can be the right answer.
Keep your list from step 3. It shows which detours you are accepting if you choose the tool, and it is a good start for a brief if you later have custom software built.
Read next
What's your idea?
Tell me what you're trying to build. I try to reply within one working day.