There’s been a handful of interesting announcements or things this week that didn’t (or don’t) quite need their own post, but I wanted to briefly touch on since I’ve seen some questions on them. These cover everything from the crazy circular sale of Withings back to itself, my time speaking at the HRV Summit last week in London, Favero’s super cool white paper on power meter accuracy, and today’s announcement of the AIRhub core (for a bunch less than before).
Actually, spoiler about this post: Nothing about this is brief. Sorry, not sorry. But, I think it’s interesting…and thus hopefully you do as well. With that, let’s dig into it.
Nokia Sells Withings…Back to Withings:
Oh Withings, someday someone will undoubtedly write a book about you. The question is: What will the title be?
Will it be the story of a successful fitness company cratered into the ground by their ‘loving’ parent company? Or will it be the story of a company that managed to flourish after escaping their parent company? That’s the question still to be decided.
For those following along in the drama that is Withings – back about 2 long years ago they sold themselves to Nokia for a cool $191 million. At the time, myself and just about everyone else said that sounded like a horrible idea. The reasons were many – not the least of which that it’d been a long time since Nokia actually built consumer products that people wanted. Outside of Finland, their brand was no longer associated with much good in the consumer tech realm – let alone the fitness tech realm. Instead, they became like Kodak and simply licensed their name to be applied to various products. Sure, they still did enterprise stuff, but that’s irrelevant here.
Nokia wasted no time in proudly erasing the Withings name from products, re-branding them Nokia Health.
But that would unfortunately be the only tangible thing to happen within the new Withings. After that, they more or less stopped releasing new products. No, simply releasing a new variant of the same watch doesn’t count. Nor does removing the bedside clock from the sleep sensor and calling it a new product. Sure, technically it was/is, but that’s not why Withings became so successful.
They became successful because they were the first WiFi weight scale…ever (my initial review of it from 8 years ago), and a really damn good one at that. They became more successful because they completely re-thought what a classical timepiece should be and created the high-end Withings Activité watch, which didn’t look digital yet had all the activity tracking built-in. They then brought that to the masses with the Pop variant.
All of which were (and are) still great products.
But while they were busy building no new products, Nokia was busy ruining the apps for their existing users. It would be months of frustration as Withings forced users over to new and re-branded apps that were so full of bugs they could have been mistaken as an ant farm. And then some products Withings abandoned entirely. Got a Withings Baby Scale (as I do), good luck with that. How about their Home cameras (also have one)…uhh…yeah, about that.
All of this swirled into a storm of speculation on whether Withings would be sold off, or put to death. Nokia themselves decided to pour gasoline on the fire by writing a company-wide e-mail saying they were studying the options.
Ultimately all was answered this past week when one of Withings past co-founders actually re-bought the very company he sold off. That’s as beautiful a blend of French brazenness as you’ll ever get.
The question being – can Withings recover?
Assuming they take back the Withings name (since obviously, the Nokia name didn’t work out), I think there’s a slim possibility. The challenge the company faces is that the fitness/health market is very different now than it was when Withings went to the dark side. I’m sure they noticed, but there’s a boatload of new entrants which are now taking their market share. And products that used to be unique, are hardly so. Countless brands have classical timepieces with activity tracker smarts in them. And there are more WiFi scales than ever before. Sleep trackers? A dime a dozen.
Sure, Withings still has one of the best WiFi scales out there (if not, the best), but that alone won’t float the company. It’s gonna take a long time to undo the app damage that ‘Nokia’ did, as well as repair the numerous lost retail partnerships too – many retailers stopped carrying the Nokia products as well.
Hopefully, we’ll see something rise from the ashes. My guess is that either the holiday 2018 or CES 2019 (January) would be logical points to do so. But that assumes the sale closes quickly, and that Withings still has the funding (and ideas, besides a hairbrush) to power through development of a new product. Here’s to hoping!
Favero Assioma IAV Update:
Last week Favero announced an update to their Assioma pedal-based power meter. Assioma came out last summer following on the Gen1 variant – BePRo. It’s an awesome unit, especially because it’s the least expensive pedal based option that hasn’t had any real teething issues. Favero doubled down on that last week though, with their big announcement that essentially does two things:
A) Increases their claimed accuracy to +/- 1%
B) Adds support for elliptical chainrings (aka oval rings, osymetric rings, Q-rings, wonky-ass rings, etc..)
We’ll talk about the first one in a moment, but the second one is them doing a solid job of edging out Garmin Vector in terms of chainring compatibility. See, very few power meters actually support elliptical chainrings. Sure you could always install them with any power meter of any type, but it wouldn’t mean you had accurate power numbers. That’s because the power accuracy varied based primarily on cadence, and as that fluctuated so did accuracy. The variance would generally be up to about 4%, again, dependent on cadence.
The trick to ‘fixing’ this was dependent on accounting for changes in angular velocity. When it comes to power meters on the market today that support this, there’s actually only a couple: PowerTap Hubs, PowerTap P1, ROTOR power meters. That’s it!
(Nitpickers corner: Both Power2Max and WatTeam have somewhat claimed to support elliptical chainrings, but I’ve yet to see them or anyone else prove it properly)
None of the biggies you might expect like Quarq, SRM, Stages, or Garmin support it.
In any case, it’s not so much a difficult problem to deal with, it’s just a case of getting it implemented. And Favero now did that. Woot to them!
Next, was their increase to +/- 1%. This is a whole interesting bucket, as they’re making a claim that other power meters are less accurate because they aren’t measuring some of the angular elements quick enough. So, let’s chat about this in two pieces:
A) Their study that they put together
B) Whether their claim about other power meters is valid
First off, their study is without question the most impressive self-published power meter study I’ve seen in years (if not ever). Seriously, nobody has done anything like this in terms of detail or data. I’m really impressed with their work here. You can read the entire 27-page study here (PDF).
The study is basically divided into detailing two portions. Part of it focused on round-rings, and part of it focused on elliptical rings. The elliptical ring portion all makes total sense, as it shows how they account for changes in angular velocity as you pedal. All good there.
And the other portion talks to round-rings, specifically in cases where angular velocity varies. Why would it vary? Their examples include cases of a hard climb where your cadence is very low and you’re iterating through part of the pedal stroke at different speeds (the point where you want to die basically). They also call out some trainers where the inertia may be slightly off and as a result you’re compensating that with slight changes in angular velocity that you might not realize.
This claim is actually pretty interesting, and they show a bunch of data that appears to support their claim of slight accuracy changes in these edge cases of about 0.4% (in some situations). On the surface I’d agree with their claim that those specific situations could/would probably alter accuracy slightly. But, as to whether or not that’s meaningful? Hmm, that’s a much tougher nut to crack. It’s been a long time since I was fighting out cadence at 50RPM on a climb. Or, that I had such a bad trainer where inertia was wonky.
Then there’s the aspect of whether others are wrong, or that others aren’t looking at data points fast enough. On the wrong piece, that’s also plausible – but it would take a bunch of testing to look at. As for evaluating speed of data collection, that’s definitely not the case. Most of these units are actually making calculations far more times per second than they transmit.
Which, brings us to a slightly funny point about all this. Despite these increased claims in accuracy, they’re less accurate in brief stop/start pedaling areas (like on a bike path, in city traffic, etc…) where they show zero-value numbers for 1-3 seconds upon starting pedaling again. I discovered this in my Specialized Power Meter review two months ago, and Favero has come back with some reasons as to why this is happening:
“We confirm that the different behavior of Assioma and Specialized is a direct result of some technical design choices, mainly the type of measurements filters (stability Vs. reactivity-oriented), the ANT+ configuration (event Vs. time-based), the ANT+ Power message type (crank torque CT Vs. power-only PO and others) and the measuring point (Dual Vs. Single-sided).
Every power meter we had the chance to test (bePRO, Assioma, and the main competitors) has a different instantaneous behavior, but nothing should change when it comes down to average power. From our experience, there is no best or worst choice in absolute terms, just different trade-offs and advantages depending on the specific situation.
Of course we can only speak for us, so we can tell you that we chose these software solutions and protocol (event-based, stability-oriented, dual-sided, etc.) because in our tests this software design proved to be able to guarantee an excellent measurement stability in the overall road cycling experience, and this seems to be noticed and appreciated by both bePRO and Assioma’s end users.
Although we can understand it may appear “aesthetically” unpleasant in this specific case, we definitively do not consider Assioma’s behavior in short pauses as a sign of inaccuracy, but mostly a sign of our power meter’s ANT+ protocol implementation (common to some competitors as well).
Besides all that has been said above, if you make similar tests with future firmware release of Assioma, please consider:
• tests on short pauses (up to 3s) are deeply affected by the position of the legs/pedals at the moment of stop and restart;
• each bike computer model has its “coasting/stopping” detection that has to zero the power data after a certain amount of time;
• bike computers usually save data every 1s (while the ANT+ messages are 0,25s – based) so they give back already “averaged” data, which are not the actual “raw” ANT+ messages we can observe during a lab test;
• any power meter comparison also has some intrinsic time alignment issues.
I hope this info was of any help and will be useful to clarify the cause of the Assioma-Specialized discrepancy in short pauses. In any case, thanks again for your article, every feedback from you is always very welcome and we are very glad to be able to give you our official position in these cases.”
(Note: I was going to try and shorten this quote, but I think just providing the entire thing makes the most sense)
In any case, technically what they say above is largely correct. Except this one tidbit:
“…we definitively do not consider Assioma’s behavior in short pauses as a sign of inaccuracy…”
That part is definitely not true. As shown in my tests, it doesn’t capture the first 1-3 seconds of power data, and no amount of lacking that makes it more accurate. If it’s missing data – it’s missing data. This has nothing to do with ANT+ messages, since I’ve seen it on no other power meter on the market (and I’ve seen all of them).
The biggest question is whether or not this inaccuracy matters to most. And that, I don’t know. Certainly, it’s not going to have any appreciable impact on your normalized power or other longer metrics. Nor is it likely to have any impact on sprints, unless you do sprints from a standstill every time. That’s because it’s the 0 to non-0 value that’s tricky. Not something like 250w to 750w.
But ultimately, that’s kinda like evaluating whether or not low-cadence but high variability pedaling is meaningfully impacting to accuracy, compared to other units that don’t account for it. We’re talking solid minutia here.
Either way – I wouldn’t have any qualms on recommending Assioma to folks. It’s an awesome power meter for the price, and even more so now if you have elliptical chainrings. And more than all that – I’m just impressed with their whitepaper, seriously, other power meter companies should follow that lead.
You may remember AIRhub, which is kinda like an outdoor trainer that can provide resistance on the road. Higher performance athletes use it to get more resistance training in places that otherwise wouldn’t provide a good training day. For example, on local bike paths. Same goes for riding with slower riders, by adding wattage to ‘equal’ the playing field. Of course, some folks laugh at these suggestions and say just add brakes or such, but most know that’s not exactly the same effect.
In any case, I tested their 2nd gen unit – the AIRhub Pro ($856USD) this past February in Australia and found it does exactly what it says it does. This after testing the far more expensive AIRhub Original ($1335USD) the previous year. While I actually liked the less expensive AIRhub Pro over the original, my issue was the same with both: They were simply too expensive to become more accessible to the masses.
There are boatloads of people (including myself) who may want to train with friends/family that simply aren’t going to be able to keep pace. For example, when I’m out for a ride with The Girl, it’s not viable for her to match most of my efforts on the bike – even drafting.
So AIRhub today announced a new product – the AIRhub core, which gets it down to $595USD. But there’s a catch: It’s only the wheel hub. It simply takes the AIRhub Pro and removes the wheel. You still need to build a wheel around it. I had said before that the price point needs to be between $499 and $699 to get critical mass. And, to that end, they technically hit that range dead-center. But my comment assumed it was still a wheelset.
At $595 you’re still looking at another $75-$125+ for the labor at a bike shop to build the wheel and true it, plus the parts you need, which at an absolute minimum for spokes/rim/etc will set you back another $100. So let’s say $200 for fun. So now you’re back at $795…not all that far from their costs.
On the flip side, if you wanted to put this in a more high-performance wheelset, then this definitely makes more sense as you can skip the higher cost of the other units.
So it’s definitely good to see them getting the price lower, but I just gotta believe that magic price point is still sub-$500 for a wheelset. That makes it affordable to the point where coaches can buy one or two as well to loan to athletes to try out for certain training days, and perhaps even for non-UCI WorldTour Pro teams to have them on-hand too. Like your local triathlon club to ‘slow down’ one or two of the front of pack folks.
(Note: Photo via Terrain Dynamics, the company behind AIRhub)
FirstBeat HRV Summit:
Finally, I wanted to very briefly mention that I spoke last Thursday in London at the FirstBeat HRV Summit. This was the first time I attended this event, and presented on ‘The Future of Wearables’. My presentation dove into some of the trends in wearables, and of course as always, where things are going south. The presentation had some similarities to the one that I presented a few weeks prior at the Connect IQ Summit, though it was about 92% new goodness. And yes, I complied with your popular request.
I don’t know if the entire session was recorded – but if so, I’ll be happy to publish it. I’ll be publishing my thoughts on the Connect IQ summit a bit later this week (I was waiting for some other minor news to drop to tie into that post, but figure I’ll just go ahead and get my tidbits out sooner instead).
In any case, as for the HRV Summit, it was pretty interesting. I was only able to attend for the first day – but it was still pretty fascinating. The morning was filled with details from deep medical studies on the benefits of exercise, as well as how they’re leveraging HRV data, and of course details on how FirstBeat does their stuff (though, that was just a small portion of the overall day’s sessions).
The title of one super-interesting talk from Jonathan Myers, Ph.D., was “Cardiorespiratory Fitness, Health Outcomes, and Health Care Costs: The Case for Fitness as a Vital Sign”, but in reality, it probably could have been titled “85 studies that show that if you don’t exercise you’ll die”. That said, it was fascinating and I’m hoping they publish it somewhere.
This then segued into a talk from one of the sports scientists with Arsenal FC, and how the team leverages a boatload of data sources to determine how much load a player receives in training each day. Again, another talk I’d love to see published somewhere. Hopefully, FirstBeat follows what other conferences do in publishing the talks on their YouTube channel.
As for the event, it was held at the Lord’s Cricket Grounds. I don’t know anything about cricket, but apparently this locale is famous. I did note that they don’t let anyone step on the grass. Unless your name is ‘Steam Roller’, because the steamroller wandered all over the grass all morning long. Still, at least it was pretty grass (that apparently, I didn’t take a picture of).
I’d definitely go back again, it was interesting in that it was a slightly different audience than I’d usually present to (which is more device-platform industry specific). In the case of FirstBeat, it had both a sports and medical slant to it. Of course, there were still numerous industry folks there from all the brands you probably know: Garmin, Suunto, TrainingPeaks, Strava, and many more. And, even better – I actually got to meet a ton of DCR Readers that were there. Woot!
So with that, this post is already a gazillion times longer than I envisioned. Thus I will end it. Thanks for reading!