I finally ran 5K under 20 minutes. But the harder challenge was waiting behind my computer.

The plan was simple: start at 4:05/km and increase the pace by 5 seconds every kilometer.

4:05 → 4:00 → 3:55 → 3:50 → 3:45.

The first three kilometers went exactly as planned.

Then reality kicked in.

After 3 km, I simply couldn't increase the pace anymore. From that point on, I was just trying to stay under 4:00/km.

And to make things even more interesting, I decided to change my route during the final kilometer.

The gravel path looked too muddy, and I didn't want to get my nice running shoes dirty. 😂

A couple of sharp U-turns later, I somehow managed to find either a headwind or a slight uphill in every direction.

But I finished.

5K in 19:56. Average pace: 3:59/km.

But this run also reminded me of something I've been working on while building Race Day Coach.

Finding the right running route shouldn't be so complicated.

Before the test, I spent time thinking about where to run. I wanted a relatively flat 5K course without too many sharp turns or unnecessary interruptions.

And that's exactly why I've been working on a route planner for Race Day Coach.

The idea is simple: open a map, add a few points, create a route, check the distance and elevation profile, and eventually use that route to build a race strategy.

Sounds easy, right?

Well... building it is another story.

Connecting points along actual roads and paths requires routing technology, map data, and a lot of calculations.

There are services that can handle this for you, but they come with a cost.

And when I think about hundreds or thousands of runners creating and editing routes, those costs become an important part of the equation.

So now I'm exploring another challenge:

Could building my own server-based routing system make more sense in the long run?

It would require more development work and an upfront investment, and running your own infrastructure certainly isn't free.

But it could give Race Day Coach more control over costs, performance, and future development.

That's one of the things I enjoy most about building this platform.

I encounter a problem as a runner, then sit down at my computer and try to figure out how to solve it.

Some days I make progress. Other days I discover five new problems while fixing one.

But little by little, it's coming together.

Sub-20 5K: done.

Now I just need to teach those map points how to communicate with each other.