Off-the-shelf software fails when your users are not average

We are not snobs about buying software. Plenty of organisations should start with something that already exists. Payroll, email, a well-known CRM, a bookkeeping tool: do not commission a custom cathedral when a house will do. The Always 49 opinion starts later, at the moment the bought system begins to shape the work instead of serving it. That is when people invent the second system, the unofficial one, and the unofficial one is usually a spreadsheet, a shared inbox, or a person who has become human middleware.

You can see the failure in the user experience long before anyone calls it a software problem. Extra fields that do not match the job. Required steps that exist because the vendor’s model of a “company” is not your company. Training sessions that are really translation sessions: here is what the button is called, here is what we actually mean, here is the order you have to do it in or it breaks. That is not digital transformation. That is your staff compensating for a product that was designed for a market average.

Custom software is not automatically better. Custom software that is built as a pile of requested features is just an expensive way to recreate the same confusion. The reason we exist as a UX-first studio is that the alternative to a bad platform should be a clear model of the work, then a product that fits. Wishy’s job is charitable giving, not “a website with donations somewhere”. Staff Checks is about screening candidates with less friction. Comtec is a portal for translation work that staff and clients can actually inhabit. Those products are specific because the users are specific.

There is a cost conversation that needs more honesty than it usually gets. Licences look cheaper because the invoice is familiar. The hidden cost is operational drag: slower onboarding, more support, more errors, more people keeping shadow records “just in case”. Founders tell us they wanted time back. Mark Burton’s line about scaling a business so he could get time is the version of success we take seriously. Software that consumes time in workarounds is not an investment. It is a subscription to friction.

How do you know you have reached the point of a custom build? When the workarounds have names. When a new starter’s first week is learning the exceptions. When a small change in the real world — a new partner, a new check, a new depot — takes a ticket with the vendor and a prayer. When the people doing the job can describe a simpler path than the one the system forces. That description is a specification. It is also a cry for help. Our Discover workshops are often just a structured way of hearing it.

Building the replacement still has to be humble. You do not throw away everything the off-the-shelf tool got right. You keep the mental models that already work, and you remove the humiliation. Humiliation is a strong word and we mean it. Software that makes a competent person feel incompetent is rude. Rude software gets abandoned, or it gets used under protest, which is worse because the data then lies. UX-first development is partly an etiquette. Treat people as if their attention is scarce, because it is.

Integration is usually part of the answer rather than a holy war between buy and build. Real organisations already have a landscape. The custom piece should be the part that holds the unique work, then talk politely to the systems around it. That is architecture in service of users: fewer double entries, fewer “export to CSV and pray” moments, fewer opportunities to get the same fact wrong in two places. We would rather a smaller product that fits than a grand platform that demands the business become generic.

If you are living with a tool that almost works, do not start with a feature list for a replacement. Start with the people who have been making it work anyway. They already designed the better product in their heads. Always 49’s job is to listen, model it, prototype it, and only then write the software. Off-the-shelf failed them because they were never the average customer. They should not have to keep paying for that fact.