14 thoughts on “The Modularity Revolution. How markets are created”

  2. Are the slides you presented available publicly? Some of them are very hard to read in the video.

  3. Excellent as always. Some suggestions:

    1) I suspect a uniform weighting of your 10 factors is not likely to reflect reality. Do we really expect the need to remodel my house to tolerate your technology to have the same impact as having to modify my behavior? You should be able to quickly come up with an improved set of weights by looking at the difference in means between populations that have a trait and those that do not. The standard deviation of those means (not flat standard deviation) will inform you whether each factor is even statistically useful in making predictions or should perhaps be discarded. You may only have 7 factors, not 10.

    2) A refined model should then mean that your scatter plot of innovation adoption time vs. weighted modularity score will be a lot less noisy.

    3) That in hand, I am sure we will all be greatly entertained to learn whether the mass adoption of alternative energy will occur before or after the global heat death moment of no return. I understand that there are crocodilian fossils to be found north of the arctic circle from a previous warmer epoch in our history and would like to know whether I should be buying my retirement beach real estate in Nome AK or Bali. I can only assume you’ve already privately worked that out, but was curious if you’d share. 🙂

    1. A refinement of the model is absolutely essential. As is replication in other regions. In particular the notion of leverage of prior art which I weigh as a single vote can be improved by having a list of enablers act as multiple vote. I suspect that is far more influential than other suspects (like capital, policy, law, geopolitical, energy, etc.).
      Regarding warming, Nome AK is 64 deg. 30’ N. My vacation home is at 61 deg. 21’ N. Make of that what you will.

  4. Have you looked into going even further back than your current technology diffusion curves go? Say looking at how long it took for the printing press[1], the cotton gin, the moldboard plow, etc to get to 90% adoption? Of course these are harder to analyze. My guess is that if you look on a large enough time scale, you’ll see that innovation does happen faster today than in the past.

    [1] See second figure of this link: https://en.wikipedia.org/wiki/Global_spread_of_the_printing_press

    1. Adoption does happen faster but my research indicates that adoption speed is not caused by the fact it occurs recently. Correlation is not causation.
      It’s important not to get lulled into a sense of complacency and wait for Moore’s law to take care of time sensitive problems. The anomalies will kill you (literally).

  5. Thank you again Horace for an excellent insight. ( Many people know stuff…very view understand it)

    As a side issue. Horace I see you do not mention “personal” robots as a possible “new” product/technology. I think it will be massive and it will fit Apple’s strenghts much better than cars.

    1. I have to have some data to begin thinking about the trajectory. Personal robots have not, as far as I know, reached market zero meaning they are somewhere after invention and perhaps before market formation.

    2. Pieter, I’ve long believed personal robotics is a market that, when the time is right, will be massive. It will have a far greater impact than the touch screen revolution of the last ten years has had.

      At some stage Apple will move beyond selling devices with screens and branch into new markets, I agree robotics would fit them well.

  6. I am not sure if I understood rightly your research.
    You are trying to derive a model from historic data of successful market saturation to be applied to guess the adoption speed of future technologies, right?
    Ignore everything before m10, when you are there your product has been understood and everybody in its market can be persuaded to buy it and has the means to buy it (that is you assume penetration does not depend much on knowledge [advertising], functionality and price after m10).
    After m10 penetration velocity can be derived by your push/pull model and, if you are right, velocity can be managed by entrepreneurs in ways that are different from what’s going on today (where everybody push on price or advertising or both).

    That is very very interesting.
    What I miss is the link from your data about technology adoption and the choices an entrepreneur must make for a sing product using that technology.
    There are indeed products (I think about Apple) that have higher velocity with less advertising and higher prices than competitors, so a link could be present.

    Have you tested your model against single product adoption, like iPhone against Galaxy? How is the score of the products in your yes/no chart?

  8. Really enjoy your blog and podcast. Thank you for sharing your analysis and insights. In terms of your current thinking, could you outline the up-to-date factors that drive speed of adoption in the diffusion curve?

  9. Horace,
    I love this work and the direction you are going. I had two thoughts that might be relevant.
    1. I would want to try a regression analysis on your modularity factors to determine the relative weights.
    2. In the IT business, I have heard some Latin American resellers suggest that their country follows US technology trends (for IT) by 5-10 years. I wonder if you might just need to look at penetration in one or two other countries. If you find that the lagging country is missing some of the factors that are present in the US, you will have a nice piece of confirming evidence.

    I would be happy to help with the regression analysis; feel free to reach out.

    Keep up the excellent work!

    Peter Laudenslager

Comments are closed.