We have been building Comus for over a year. The first version had no graphical interface at all. It was managed entirely through PHP.
That was deliberate. We wanted to test the concept and prove the market before sinking serious time into a full product. No point polishing something nobody wants.
So we built a bare-bones alpha and trialled it with one client. We gave them a very favourable no-cost deal in exchange for being the guinea pig. It went well. The client was happy, and three weeks later we ran Comus for a second event with some tweaks. That one went even better.
Both of those clients are now booked in to use the system for their annual events next year. Between them that is up to 8,000 people. For a system with no interface, that is a decent result.
More businesses were showing interest at this point, and the tempting move was to keep signing them up. Instead we chose to stop and rebuild properly. The original system was largely hardcoded to prove the concept, and you cannot scale on that.
Alex has since spent hundreds of hours redeveloping it from scratch into a true MVP: a professional product that is polished and cohesive rather than held together with string.
The rework has not been smooth. Making the system dynamic and adding the features people actually need created bugs we never anticipated, which created fixes, which created more bugs. The whole thing has taken about three and a half months, well beyond what we planned.
But it is nearly there, we are launching in the new year, and it was worth the wait. Prove it works first, then build it properly. In that order.