Garmin has the muscle power to turn Trainingpeaks into the optimal AI-enabled Training platform, that offers the current C2A communication, certain customer data analytics and then a lot of AI automation. You will still have the required human factor in coaching, supported by AI, that most are predicting to be the way forward in this market.
~3× ARR (or 10-12x EBITDA), which is relatively conservative for a profitable, niche vertical SaaS business with strong recurring revenue and low customer churn.
I just want to address the AI topic. There’s a German platform called MATS that already has an AI coach built in. You can set your target competitions and need to enter your available time slots each day (you can automatically copy this for upcoming weeks). Then the AI seems to generate a reasonable training plan. You also have the option to fill out a questionnaire every morning with questions about sleep, how well you feel, injuries/illnesses, etc., and the AI adjusts your planned workouts accordingly.
If you don’t manage your slots properly, the AI will go a little wild(for example, schedule two 2.5-hour threshold bike sessions in one day)
my 2 cents TP sold at the best time, given that they started to advertise a lot (podcasts etc ) about 1 year ago they kind of knew the best days are over. and they would have had to invest heavily or take the money, while they still could.
Did not read much of this. facts are that as a company you must be profitable and grow to keep existing. Some of that profit has to be in R & D so you can keep selling whatever you are selling. Sometimes that R & D is purchasing a patent or company that has already done the leg work into the R & D that can be incorporated or integrated into their product.
I mean, Wahoo continues to have their App to control stuff, even though they’re deeply embedded with Zwift now.
So I think it will likely still chug along, but it won’t be the core focus.
My sneaking suspicion is that you will see some TP analytics come over to Garmin Connect, and maybe the UI/UX experience within TP gets an upgrade (which I think everybody would welcome). And Virtual, IMO, has really taken a sizable leap forward over the last six months.
If they don’t do a one-year TP Premium subscription included with a Tacx trainer purchase this fall to boost TP Virtual ridership over the indoor season, I would be shocked.
My biggest concerns are that 1) Garmin seems to be increasingly more difficult to play with, and I’m not seeing a better quality product. Just yesterday I had to go thru the $99 refurb replacement unit/7 months out of warranty on my daughter’s Forerunner. I’ve done this with a 935 2x and and Edge 530 1x. My COROS is the best $200 I’ve spent for a training wearable…ever. 2) with Connect, although the claim seems to be full customization, I HATE anything “steps” being forced down my throat…and I haven’t found a way to eliminate/discard steps visibility. I don’t give a ----- about steps. I’m a little concerned Garmin will find a way to force feed crap via TP and screw up a relatively easy platform to use…
but also cross bull shitting, like adding the vo2 guestimator of garmin into tp lol .
I cant wait for the day garmin stops the vo2 guestimator as iam sick and tired to tell people its useless.
I do not have a lot of visibility into the broader TrainingPeaks plan, but the ongoing work to support WKO5 is continuing, and they are fully committed to ensuring it remains fully operational across all supported operating systems.
More broadly, the future of performance analytics is clearly moving (moved?) toward AI-supported tools, and I believe one of the goals of this acquisition is to accelerate that development. The important distinction is that these tools appear intended to strengthen human coaching, not replace it, by helping coaches analyze more information, identify patterns faster, and make better-informed decisions.
There’s no need to be hyperbolic. There are plenty of AI tools for training out there that will not KILL anyone. “100% serious” get a grip… There are plenty of people putting together substandard training plans with AI but to think that they are following these plans to death is ridiculous and ignores the successes.
Seems pretty similar to the other AI platforms. I use Humango to good effect. Mostly for the basic workouts, but it does take some fanagling to work out how to schedule the weeks.
Public AI models, something that all of you guys would be using (not defense or private AI models) have an error rate of over 7%. There are no AI tools available for a normy to build things that are at a level where you can completely assume safety. Now, you could say the same thing for an inexperienced coach, but an inexperienced coach isn’t hallucinating and making shit up. I have a grip on this. We’re several years away from a public AI model being a programming tool. Analysis tool? It’s great at that.
ditto…they are buying the strategic asset i.e. the data moat… TP has 25 years of real coaching decisions which are ripe for real AI / ML training… The SaaS business will continue to accumulate real data…
I’m sorry are you living in 2023? Publicly available models are perfectly fine for using as programming tools. Why do you think every tech company is pushing their use (before tokens got so expensive). Again, you said these models would KILL people (100% serious). Don’t lose your guiding argument here. Just because a training plan has an ‘error’ what does that even mean? Certainly the error doesn’t include a workout ‘shoot yourself 1x100% intensity’.
Agree with this - Garmin UI design in general across devices and Connect is pretty poor, and TP in my opinion is less so, so hopefully they leave it alone!
