He Built It in Two Hours. He Knew When to Stop.
Taylor Matzeller built CustomGradient in two hours. Two weeks later, he decided to stop. Not because it failed. Because he was honest enough to know when something isn't worth more of his time.
Most builders don't know when to quit. Taylor Matzeller does.
CustomGradient started as a question he had been carrying for a while. What if, instead of letting users upload profile pictures, you assigned them one? An image generated by an algorithm, unique to their name, deterministic and impossible to fake.
Two hours later, he had his answer. It worked.
The Experiment
The MVP was simple: input any name, get a gradient back. Same name, same gradient, every time. Taylor spent the next week polishing the algorithm and building a UI clean enough to ship. You could download your gradient and share it.
Then he asked the obvious question: is this viral enough?
The answer was no. So he made it weirder. He added zodiac signs and a two-name compatibility check. Enter two names, see how well your gradients align, let the algorithm tell you something about your relationship. It took another week to build.
The plan was: launch, add Google ads, market on TikTok. Straightforward.
Then Google rejected the website for ads.
The Moment Most Builders Spiral
Taylor didn't spiral.
"I felt a bit frustrated, mostly because I didn't know anything about the requirements for ads."
He read Google's policies. He understood the reasoning. A website with minimal content doesn't qualify. The rule exists to prevent exactly what he was trying to do: monetize something thin.
That realization didn't feel like failure. It felt like information.
"I came to the conclusion that it probably wouldn't be worth continuing the effort to build and grow this for now, since I have other projects to prioritize."
No pivot. No desperate marketing push. Just a clear read of the situation and a decision to move on. CustomGradient is live, technically working, and listed for sale.
Why He'd Do It Again
The product might be done. The idea behind it isn't.
"What if, instead of allowing users to upload profile pictures, we assign an image to them that is deterministically generated by an algorithm."
CustomGradient was a two-week test of a concept Taylor has been thinking about for much longer. It answered the technical question. It taught him something about distribution and monetization. And it gave him a clean break from the bigger, heavier projects he was carrying.
That's worth two weeks to most builders. Taylor just had the clarity to say so out loud.
CustomGradient is available at customgradient.com.
Follow Taylor's work at @taylormatzeller.
