Runna Reviewed

I learned to dislike running in school. Then, in basic training, I learned to loathe it. It took many years before I started to actually work at running. Slowly, but surely, I found that while I didn’t exactly like running, I wanted to like it. In early 2025, a friend emailed me and another friend, challenging us to run the Berlin half-marathon in March 2026. We signed up for the ballot, and got drawn to participate. At that point, the furthest I’d ever run was eleven or so kilometers, and I knew I needed to do some serious work. That’s when I signed up for a Runna plan. After using it to train for a 10k in September 2025, I used it to train for the Berlin half, and finally the Runway 10k in June.

Runna is generally speaking good at what they do, which is generating a structured training plan and provide a simple way of running each individual exercise with guidance from your training watch or smartphone. This is core functionality, and it makes sense that they are good at doing this. If they weren’t, I don’t think they would have seen anywhere near the success they have. That, however, is only one side of the equation, and Runna is certainly not without its flaws.

Runna are touting their use of AI combined with humans who perform quality control and improvement of plans. The AI bit, and particularly the feedback part of it, is (in my opinion) being sold as far more capable than it actually is. The feedback was never all that useful. It was always generic (despite having access to detailed data), and interpreted even very minor elevation changes (five meters over 1500 meters of running) as major hills. This means I can’t rely on the feedback as part of my post-run analysis.

After each run, the app would ask me to indicate (thumbs-up or down) how the run went. When indicating it went badly, I got a short list of relatively generic options to choose form, and that was that. Just as frustratingly, there was no way to provide any feedback at all if I indicated the run went well. If I feel it was too easy, I should be able to provide that feedback. I ended my long runs in weeks 21 and 22 of the Berlin half plan early (7,5/18 and 13/20 kilometers, respectively). Now, while I don’t necessarily expect Runna to suggest changes after a single truncated run, when truncating two in a row, they really should have recommended some changes – or at the very least helped me diagnose what was happening.

(What was happening was a bad case of accumulated fatigue, and I needed to cut WAY back on my training for a little bit in order for my legs to recover. Because I didn’t know or understand the signals, I wasn’t able to diagnose it.)

If you, at this point, are expecting me to say that Runna is bad, I won’t, quite simply because it isn’t bad. What Runna has done for me is give me solid plans that have helped me achieve a lot of my goals, including running two half marathons this year, and break through the 60-minute barrier in the 10k. It has done so in a format which I have been able to push straight to my watch.

While Runna has worked for me, I am not going to continue using it. The lack of options to analyse my workouts is a show-stopper for me – especially at the price point it is being offered. Part of the issue here (for Runna) is that a year ago I didn’t have the experience to see if a proposed plan was realistic or not, or to question a plan. Having run regularly for a year, and grown into a better and more knowledgeable runner, I now do. In other words, the value proposition Runna presented me with that made sense a year ago, no longer does.

As for my relationship with running, running regularly and to a specific plan has healed it. After Berlin, it took two days before I wanted to go out running again, and I am signed up for another half marathon in September. That is a stark contrast to after I ran the 10k in Oslo in 2022 and 2023 – I was burned out and wanted nothing to do with running at all. This feels like a huge improvement – and much more sustainable.