Zwift Details Challenges With Building Their Own Hardware

image

CyclingTips today published a great ‘From the Top’ podcast episode about the history of Zwift, talking to both Jon Mayfield and Eric Min. Those two, being two of the co-founders of the platform, with Jon bringing the initial development/technical side – and Eric bringing the financial/business side. The roughly 70-minute long episode takes you from Jon buying his first road bike, onwards to realizing how much basic “dumb trainers” sucked a decade ago, and then deciding to fix it forever. Wade Wallace of Cycling Tips did a great job weaving together the two sides of the interview (done separately) to form one cohesive picture. You can listen here on Spotify, or Apple (or anywhere else for that matter).

However, the historical story of Zwift is not what caught my attention. Not because it isn’t interesting, but because I’m personally pretty familiar with the story. I met Eric and Jon back in 2014 at Eurobike for a first look and ride on the platform. They had rented what was essentially a glorified coat closet inside the Eurobike convention halls. Actually, I think it was legitimately a coat closet – not kidding. Both Eric and I looked like the teenagers we apparently were (seriously, go look at the post).

What caught my attention was a short back and forth discussion on Zwift’s hardware ambitions, to ultimately build smart trainers and smart bikes. I’ve written frequently about this over the last couple of years. From Zwift’s first job posting advertisements, to Zwift detailing it on CNBC, layoffs to make room for hardware teams, and various other things along the way. Thus, when Zwift says hardware – I’m all ears.

Within the episode, there’s a few chunks dedicated to talking about their hardware ambitions with Zwift’s CEO, Eric Min. It’s no secret that Zwift has struggled with getting their hardware off the ground, and Eric confirms that, saying:

“The microchips is the biggest issue with supply chains. We have the auto industry chasing the same components as we are. It’s just a problem for everyone. It’s causing delays for everyone, including us. But it doesn’t change our strategy.

 

We believe that we have to be manufacturing hardware in order to grow our business. It’s not to take business away from our partners, I just don’t think they’ll make enough. That’s the problem. We’re just trying to grow the pie.

 

It’s going to take as long as it takes. Supply chains, logistics costs – all of this is a real thorn in our business plans. But it’s the world we live in at the moment, and there’s not much we can do about it other than just persevere and work as hard as we can to get the products we have in mind to market. It’s definitely taking longer than we expected.”

The discussion then dives deeper for a bit on that, before interviewer Wade Wallace circles back to asking Eric about how Zwift “walks the tightrope” of trying to build hardware that will compete with the very partners they are 100% dependent on. Eric responds that a good example they see is Google and Android, where Google builds Android (the platform), and yet still competes with Samsung and countless others with phones.

And indeed, this is a reasonably good analogy. Of course, all analogies have flaws when poked too hard – but at a high level, this is a reasonably good example. From there, he expands on why this is required for their hardware partners:

“Think of that as a model for us. We want to help augment the supply that we think we need to grow the market. And we also want to set a reference for our partners, to force them to, [pauses] not force…encourage…them to innovate”

The challenge here of course, is that at present, Zwift simply isn’t reference leader for these hardware innovations. In fact, in every scenario, it’s their hardware partners that are waiting on Zwift, not the other way around. For example:

A) Elite Sterzo Smart (in-game steering): This device sat waiting on Zwift nearly a year from initial Eurobike announcement and when Zwift had units, before introduction.

B) Wahoo Direct Connect Wired Connectivity: We’re now 10 months since this was released, with still no sign of integration from Zwift – despite almost all competitor platforms supporting it.

C) Steering in smart bikes: This was nearly 6 months after Zwift rolled out steering, until the industry’s top-end bike was able to go live with it, despite having it ready for months. This readiness was mirrored by virtually every other smart bike company that had to wait on Zwift.

D) Lack of any trainer certification program: While Zwift ‘certifies’ hardware, in reality, there’s actually no testing of the hardware against any known standards. It’s merely a sticker, based on whether the hardware meets a spec sheet listing.

There are of course countless more examples, some older vintage, and some still waiting on people’s desks in Zwift HQ. Albeit, there are some more recent cases where Zwift has been relatively speedy. For example, the recent Elite Rizer (steering + climbing) was implemented the same day it went live to consumers, this past August.

Now it did seem like slightly later in the conversation Eric admits Zwift might be the roadblock. While talking about a slightly different aspect of Zwift, he noted that:

“I would like to see more people innovating within our platform. And I think part of that is just…we’re not making that particularly easy for people to do that.”

And again, there’s numerous exchanges and details in this interview that are worth listening to, even for people that think they know most of the Zwift story. Eric discusses which companies he sees as competitors (and which ones not). He talks about the major chain name-brand hotel that wanted to be put into Zwift, as well as how the investment rounds worked, and what COVID did to their product plans (and how they handled it).

Unrelated to the hardware, as you listen to the episode, it’s clear that as much as Jon loves what he’s created, it’s hard for him to reconcile a billion-dollar valuation with the old days of an agile start-up company, noting that he “never had the aspirations for it to be so big”.

And you can hear how he still desires to scratch some of those startup itches, beyond what Zwift aims to do, saying “I kinda wish that next person would build that next experience…it doesn’t have to be a threat to Peloton or other big companies, Zwift, I suppose.”

There are countless great sections of the interview from a detail standpoint, but listening to Jon talk in the last 10 minutes or so about the impact Zwift made on him (the person who created it), and inversely, what interests him now, is interesting. I suspect it’s something that many founders struggle with as their babies grow far larger and more successful than they could dream of.

With that, thanks for reading – and go give the episode a listen, either on Spotify, or Apple. Plus wherever else podcasts are found.

Post a Comment

Your email address will not be published. Required fields are marked.
If you would like a profile picture, simply register at Gravatar, which works here on DCR and across the web.

You can click here to Subscribe without commenting

Add a picture

*

30 Comments

  1. Charles

    Zwift is standing on one leg waiting for an actual gaming company to come along and knock it over. Everything about Zwift is literally terrible. Graphics, UI, game design, MMO interaction, racing, management, communication, control, everything! If you think Zwift is good you lack knowledge of games period. I would bet money Zwift isn’t even around 10yrs from now. It is obvious Zwift management is absolutely clueless.

    • Fred Lee

      I’m not particularly into games, but Zwift does create a pretty nice virtual world for cycling.

      I mean, it’s lacking loot-boxes and stuff. Is that what you’re hoping for?

    • Reid

      As an example, look at the bike mod for Grand Theft Auto. That takes a high production value AAA game in an unintended direction, but the graphics and atmosphere beat the pants off of zwift.

      I am not saying it is close to a zwift replacement, but it is an example showing how zwift can be completely outclassed in some areas by even an aging game. Imagine if something like that was designed from the ground up to be a direct competitor…

    • Those are some pretty strong opinions charles… not only will zwift be around in 10 years, but I anticipate them buying peloton.

    • While I agree that Zwift as a platform needs a technology overhaul, I don’t think it’s in some horrific shape. As Jon talked a bit about in the podcast, they wanted to delineate between ultra-realism and having bit of that cartoonish look.

      Still, one can achieve that look but offer better graphics realism – those two can coexist. And I think that’s what most want. Same goes for a better user interface for menus and such. Or countless other features folks have asked about that are more gameplay specific.

      While I agree it’s highly likely Zwift will be around in 10 years, I don’t think it’s likely Zwift will buy Peloton. Frankly, the opposite is far more likely. In order for Zwift to grow bigger than Peloton, they’d have to shift to not being a 3D game (as that limits mass-market appeal in this segment) – or at least offering other portions of the app that aren’t 3D based (but likely more instructor based). However, as Eric pointed out in the podcast, they have no desire to change that course. They don’t see non-3D platforms as direct competitors (which is horrifically wrong, but hey…).

      Looking at growth rates, it’s clear Zwift is no longer at the same momentum as Peloton (nor even the same ballpark. Peloton’s most recent earnings call easily demonstrate that, especially once you roughly guesstimate Zwift’s user base and expansion over the past year. Which doesn’t take away what Zwift has done – they’ve grown massively, and have (rightfully so), a very strong and passionate user base (including myself). But I don’t think most Zwifters really grasp how much bigger, more sticky, and faster growing Peloton is.

    • inSyt

      True, Zwift graphics/physics/fps reminds me of NFS Underground. Would love to see Codemasters create a UCI branded cycling simulator, like they do with F1, with a similar pricing scheme.

    • Mark

      Personally I think the biggest hurdle at the moment is how poor the pack dynamics are… Far too much weight put on top line power Vs wkg — 80kg+ riders flying up inclines like natural born climbers and reach end of races with lower wkg avg. Could also do with introducing *some* sort of skill to descending, e.g you can’t pedal 300w round hairpins

      If they sort that out and then allow private race events so organisers could create competitive leagues they’ll open themselves up to a new revenue stream as well as get people getting a whole lot more interesting when cash prizes could be on offer.

    • Fausto

      :”which is horrifically wrong…”. Since you are no dummy and have worked in corporate America, I’m sure you know that what a company says publicly is not always what they perceive internally.

  2. David W

    It’s not just trainers that are waiting for Zwift. It’s ALL hardware. For example, after years of trying I still can’t get Zwift to work on my Windows 10 laptop with Bluetooth. Despite a lot of useless help from their tech support. The exact same hardware and software work perfectly with TrainerRoad, Sufferfest (now SYSTM), Rouvy, and FullGaz. If everyone else can make it so that I just turn my laptop on and it works why can’t Zwift? After years? My suspicion is that they just don’t care. I would think that poor connectivity would be a huge block to adoption but they apparently don’t agree.

  3. Nik

    It feels like Zwift employs no more than 1 software developer. So not a lot of development work can get done. It’s a mystery why the rest of the company thinks this is OK.

    The Bluetooth thing is one of many egregious shortcomings. Come on people, Bluetooth is a mature technology, not some recent fad.

    • AC

      1 software developer, 5 people in the support forum explaining why promised enhancements haven’t arrived yet, and 10 marketing people surveying users to see just how much they’d be willing to pay.

    • Xavier Neys

      And probably a high rotation level. I spent many years in IT to recognize different types of bug and lack of QA. My guess is coders don’t stay long because the code base is a huge mess.
      Lack of QA is also a huge problem. A simple example: yesterday’s yellow jersey social ride in Paris, big turn up, 1000+ riders, and we started in the wrong kit…
      One thing I know is if Zwift design hardware the same way they design software, I won’y buy it.

      PS: I do not use Bluetooth and I haven´t had any disconnect issues in months. Stages smart bike and HR strap connected to Tacx ANT+ receiver on Win10.

  4. The fact your original post has no comments is really funny.

  5. jww

    Haven’t listened yet, but the CEO explains Zwift’s initial plans to produce hardware (years ago) on the CURRENT day semi shortages?

    Seems a convenient story to tell the partners.

    Even with today’s myriad supply chain issues, trainers seem to be in ample supply. Wahoo and Tacx both shipping immediately from their dot coms.

    Thanks for flagging this!

    • Robert

      It’s a *lot* easier to obtain components you pipelined 12-24 months ago since you had mature in-production products, than to get components to ramp-up production of a new product without established demand. And it’s a lot easier to run a supply chain in these difficult conditions if you’ve been running one for a decade or two than if you’ve been delivering software all that time.

    • jww

      Yeah. I’m certainly not denying it’s easier to get components that were ordered earlier vs. ordered later.

      Point was it’s questionable logic for the CEO to argue Zwift decided to produce trainers in 2019 because there’s not enough trainer manufacturing supply in the world to satiate new Zwift users. Logically, they would’ve needed to predict a COVID-style event for that to ever occur.

      CEO used Google-Android-Samsung example as the conversation stopper, though Google isn’t brazen enough to claim they built Pixel because they’re worried Samsung etc can’t produce enough Android phones.

      Seems the honest answer was Zwift wants a more turnkey solution, plus capture some of the trainer mfg profit pools. That’s not so evil to me.

      This was an awesome podcast!

  6. verolanuova

    ZWIFT is somehow stuck in between… they raised a lot of VC capital and therefore they need to growth a lot to justify those valuations.
    Probably the financial fundamentals with the potential pool of subscribers (15$/mth x 12 months or less x potential pool of hardcore cyclists) do not bring the projected revenues high enough, so a move to HW was a must…
    At this point they are struggling to balance the mission to either become a viable HW manufacturer (lame excuses on microchip provided…) or to really develop more the SW to cater to more “recreational/fitness” audience (like Peloton..). Good luck to them

    • Robert

      “lame excuses on microchip provided” – meanwhile, here’s another startup using the same excuse:

      “Apple may slash the number of iPhone 13s it will make this year by up to 10m because of a shortage of computer chips amid a worldwide supply chain crunch”

    • Greg

      I think that you’ve struck on a key step that they are missing. Peloton is not growing faster than Zwift because they have hardware, it’s because they appeal to everyone. All the people I know using peloton are NOT cyclists. They join for the workouts; an area where Zwift has massively failed. I see no reason why there couldn’t be structured workouts lead by a group leader. They need to make it more personal. Not just cyclists pedaling around a virtual word communicating with their thumbs.

    • Agree. The Zwift pool is ultimately rather limited in its current form, at least or a company of Zwift’s financial aspirations. Instructor-led workouts would seem obvious. And it’s easy to transition from the ‘celebrity’ type group rides they do now, to instructor-led group rides. And then keep on going from there – establishing the true stickiness that Peloton has in terms of membership churn rate.

      Then from there you slowly get to the point where there’s no 3D engine at all for certain rides – it’s just an instructor. Somewhat inverse of what Peloton has done with scenic rides this past spring, in making those instructor-led for folks that want that – but don’t want to be staring at the inside of a workout room.

      Peloton continues to expand its training plan bits. It’s not super coherent yet in terms of something like a 12-week build. Rather, it’s just a 12 week collection of workouts. But the main puzzle pieces are there to do it, and they’ve been slowly putting them together technologically. Not to mention ERG mode set-point accuracy on the Bike+ that frankly beats most of the smart bikes these days.

      As for Zwift’s HW troubles, undoubtedly chips are a piece of that pie – but everything I hear says that’s a very small piece in the grand scheme of the issues the project is facing.

      I would disagree that all Peloton users aren’t cyclists. I’ve been really surprised with how many ‘bike people’ in the industry I run into that use a Peloton bike for a portion of their workouts each week. Everything from CAT 2/3 cyclists to CEO’s of some of the biggest names in the sports tech industry. Everyone says the same thing: It just works…every.single.time.

      And for people with limited time, that’s a key component of it. I mix and match. Over the last 7 days I’ve done TrainerRoad workouts, Peloton workouts, and Zwift workouts/rides. They all serve a purpose, and no one was any harder than the others. All can hurt equally and improve fitness equally. In one platform you’re staring at blue bars, another platform you’ve got music synced to the workout, and another platform you’re looking at a 3D recreation. To each their own.

    • Greg

      Yes, agreed, I could have phrased that better. All the Peloton users that I personally know, in a sleepy suburban Canadian city, were not IRL cyclists initially.

    • A Tallent

      I will not speak for everyone but the Peloton/Ifit/Instructor led classes or format have zero interest for me. If I wish to get screamed at while I am riding I will just go outdoors. Riding is my escape from people not my draw. I used to manage group club rides in Southern California so did the groups scene in some of the best cycling areas on the planet. I never thought I would enjoy Zwift but it is a tool for me when I cannot or choose not to ride outdoors. I would prefer to lose the Dinosaurs and other animals and weather taking up resources but the graphics on my W10 laptop are fine as long as I am not in enormous group rides. Zwift coming up with a 3500 dollar bike or treadmill has less bearing than if it just works on the equipment I choose to use. If it doesn’t I just move on to something that does.
      As always appreciate the insight and the discussion – even the doomsday, “I hate my life”, nothing is right in the world crowd that just reminds me how great I have it on a regular basis.

  7. Klaus

    There was Steering for everyone available with the Kommander from Keith or other ESP32 based solutions. But Zwift has nothing better todo then breaking her own protocol to disable this and other solutions for people who never would buy the Elite Sterzo Smart because they possible use the Wahoo Climb.

    I think there is still no license and documentation available for Steering.

  8. Alex

    The only logical explanation why Zwift’s leadership team is still in place is that their investors are behind the hardware push.

    Not only is the hardware situation a total dumpster fire, the shift of focus also meant they didn’t really improve the game in any meaningful way in over two years. For the first time I notice a significant drop in activity from “power users” on my activity feed. Even though that’s obviously not representative for the overall state of the user base.

    Personally I went from *very loyal customer* to *customer who is staying on at least until there’s a better option* to *I might just cancel anyway*.

  9. ChillFMM

    I do not buy the pioneering view they try to sell. Zwift is a Bkool copy with 1 bilion dollar venture capital backing.
    Bkool started in 2011 and by 2014, when Zwift started, it had all the features that Zwift is still is promising its users (It even started with its own hardware and by the time Zwitft started you could even use ant ANT+ FE-C trainer in the platform). But they built a bigger community and at the end that is what counts. But pioneers, nope.

  10. mpulsiv

    I’m not sure why folks label Zwift as a game. It’s a training/racing platform. I don’t even bother with graphics. When I train, Netflix is running on my large monitor and Zwift running on my small USB monitor. When I race, my eyes are on the wattage, not pretty graphics in the background.