How to test ideas quickly?

“This idea is brilliant! I’m sure there’s a market for it and I’ll build it better than anyone else!”

I couldn’t be more wrong.

When I started my first company, I developed a mobile app from scratch.

We wanted real estate agents to be able to manage all of their business on their phones!

One of the key tasks was comparing a house to other similar ones, and assessing the market price

There were already a lot of players on the market competing for this, but arrogantly I believed “we can do something better”

And so I set out on a journey to build a tool to compare houses. The process was manual and excruciating! Agents hated it

But for some reason, my inflated self though this would help their business

After several months and dozens of complaints, eventually I realized I was wrong. But this wouldn’t be the end of it!

I set out to talk to my competitors, trying to figure out a way to integrate them on my App

One more time, arrogantly I said “their interface is not that good, we can build something better!”

And so I spent 2 months negotiating contracts, creating API integrations and developing the necessary screens on our App

It somehow worked. Agents could do their home appraisals and get a price

But it was far from ideal. A lot of agents eventually started to ask for “the real deal”, that is, true access to the platform

Looking back, I spent over 6 months (and a sh*t load of focus on other important areas) that could have been avoided by simply asking users whether they were missing this feature

In the end, all of them already had what they wanted. I was just trying to reinvent the wheel

Now, I’m trying to fight this urge to always build from scratch.

First ask users, then do a quick MVP and only after a proven concept truly start developing

I still fail way more than I would like. Coding is extremely easy and makes me feel good!

Marketing, sales, branding, positioning… That’s messy. That’s harsh. That’s UNPREDICTABLE

But in the end, it’s what really brings money through the door

Are you over-engineering something that could simply be tested by asking users if they truly want it?