Blister Labs: Product Manufacturers vs. Researchers & Reviewers (Ep.213)

Question: can engineering actually help us discover real-world characteristics about products that will be meaningful to end users? Or, when we try to start getting sophisticated with fancy-sounding systems that attempt to capture quantitative data, do we primarily run the risk of having data lead people astray?

This gets at the heart of a tension that sometimes (and perhaps often) exists between product manufacturers on the one hand, and engineering researchers on the other.

Thomas Laakso, the vice president of product and operations at DPS Skis, was recently in Crested Butte to discuss these issues with some of our Blister Labs faculty members and over a hundred engineering and computer science students at Western Colorado University. And in this conversation that you are about to hear, Thomas and I sat down in Blister HQ with Dr. Sean Humbert, one of the top roboticists in the country, a passionate skier and biker, and one of our key faculty members at Blister Labs, to think through the risks, the concerns, and the potential rewards here.


  • Will Data Set Us Free? (6:12)
  • Variability & Defining Uncertainty (14:05)
  • Data & Real-World Applicability (16:58)
  • DPS Pagoda Update (29:03)
  • Quantifying the User Experience? (34:34)
  • Interrogating Engineering (37:42)




7 comments on “Blister Labs: Product Manufacturers vs. Researchers & Reviewers (Ep.213)”

  1. Jonathan, I would love to participate in a back-and-forth discussion with anyone about the topics in this podcast. I suspect an active discussion won’t really happen via posts on this webpage, especially during the off-season. Like, if I post here, would Sean or Thomas actually reply anyway?

    Please let me know if there exists a more appropriate webpage or website out there where I can join in on some active discussions about Blister Labs topics. Like, is there a thread somewhere at a Western Colorado University website? Or anywhere?

    Or is this page here my best shot at getting a discussion going?

  2. Got to say I was slightly disappointed in this one after the awesome build up you gave us Jonathan ;).

    I happen to be an engineer in a complete different field and I think to me this tension you guys are talking about exists within my daily job… If I had to frame it I would talk about Craft vs Theory. After years of experience crafting a solution to a problem (which is essentially what any manufacturer is doing out there) they basically develop their own model based on their own gut-feel experience based decision making.
    That’s why you see shapers in ski or surfboards doing some very advanced tweaking without really being able to articulate what they are doing but essentially doing “the right thing”.

    I see that when I go skiing with my old French alpine guide, does he know snow and the snowpack the way the youngster do. Nop. Does it get it right, yes. Every single time.

    Another example (French again), is wine making. Does the bio-dynamic small craft wine maker knows what makes her wine the best ever. Nop. And when the industry start trying to replicate or map it, they fail miserably because they try to reverse engineer overnight something just way too complex and subtle.

    So I’ll have to agree with your DPS guy, in the same way we are able today to provide science backed tools to my French alpine guide to help him taking the right decision in the snowpack. Or in the same way we can today help wine makers really understanding what’s in their wines etc etc.
    We should help DPS shapers with new tools but we still need their creative craftsmanship otherwise we will end up with horrible undrinkable wine, I mean Salomon skis, I mean you know ;)

    My biggest issues with engineers who do not appreciate the importance of craft is that they think food, ski, cars, etc can be all standardised and designed eventually by a model. Keep the human in it, this is where the magic happens


  3. Yes, Jonathan, this is criticism directed at you!

    You scared the “cr**p” out of me with your opening statement saying you are no longer making the digital version of the Buyers Guide free without putting it in the context of non-Blister Members. By the end of a fairly long rant, you clarified it, please do that up front in the future. I was heading towards the cliff (or about to jump ship) before you pulled the rope tight and saved me.

    • Ha, you get no apology from me, Scott. We’re creating a culture here where people actually have to read – or listen – to the end of the statement – don’t just draw conclusions off of a headline or first sentence. :)

      That said, we very much appreciate your enthusiasm for the Guide, and we’re all extremely happy that you didn’t go jump off a cliff, ha.

  4. I think that both can be right.
    With enough measurements, and a well calibrated model, I think we will certainly be able to make some predictions about the behavior of skis for a certain skier.

    At the same time, I am sure that we are a long way off from being able to 100% percent rely on lab measurements and pick or recommend skis from that.

    So, I think it can be a bit of both. Some useful inferences to be made from measurements combined with actual on snow reviewers.

    Aftel all, when I ask for a recommendation for a ski, the Blister reviewer who helps me is using a mental model of skier Input and ski behavior , even though they are not me, and might be recommending a ski length they have not skied.

  5. I agree with the Blister Labs folks; a regularized model should be able to give you some insight into an ill-posed problem. That does not mean you can predict everything, but if you can calibrate it to be consistent with what reviewers are seeing, then inference about new obversions should be forthcoming.

    I have worked in a couple fields where it was stated that “you can’t model what we are doing, that is just not possible it is too complex” and that was proven wrong again and again. If it is about taste, about design, about experience that is very hard and very challenging. However, if the problem is “how planted do you think this ski will be that you have never been on before?” that could be tractable.

    I am curious to see how this all pans out. I think either extreme view; not possible or impossible to model are unlikely to be correct in the long run.

Leave a Comment