I Shipped Kashi. Getting People to Use It Was Harder.

I spent months building Kashi, a point-of-sale and business management platform for small businesses. It had sales, inventory, customer records, quotes, credit tracking, reports, online catalogs, and WhatsApp ordering. It ran on web and mobile.

It worked. That was the part I knew how to do.

Getting a business to use it was a different problem.

The launch

I had treated shipping as the finish line. Once Kashi was online, I expected people to recognize how much work it could save them.

I ran Meta ads for about a week at roughly MXN 200 a day. The campaign brought around 100 visits per day, about 20 registrations, and five trial users.

All five canceled.

The numbers were small, but they were enough to challenge my assumptions. People were willing to click and some were willing to sign up. That wasn't the same as being willing to change how they ran their business.

At first I wondered whether price was the problem. I added a free plan. Usage was still limited to one or two active users.

Removing the price had made Kashi easier to try, not necessarily more useful to adopt.

I had built too much before learning enough

Kashi covered several jobs that a small business might need to do. But each new feature also meant more decisions for someone opening the app for the first time.

I spoke with or observed businesses such as flower shops, hardware stores, cafés, pet groomers, and grocery stores. They didn't have the same workflows. Some used notebooks, spreadsheets, WhatsApp, or older software. Those tools had limitations, but everyone already understood them.

A new product had to be more useful than their existing process after accounting for the effort required to switch.

I could demonstrate an inventory dashboard. The owner still had to enter their inventory. I could demonstrate customer management. They still had to change a habit. My feature list described what Kashi could do; it didn't prove why a specific business needed it this week.

What I'd do differently

I'd choose one type of business first. I'd watch owners do the work and find one repeated, costly problem. Then I'd build the smallest workflow that solves it.

I'd also decide what counts as activation before launch. For a point-of-sale tool, recording the first real sale might matter more than registering an account. For quoting, it might be sending a quote to a customer.

Then I'd measure how many people reach that point, talk to the ones who don't, and change the product accordingly.

Kashi taught me to build software. Launching it taught me that the real test starts when someone has the option not to use it.