Making A Better Bike Using Data Analysis

Lynne Kiesling

This Baseline Magazine article from late June describes how Trek, a bicycle manufacturer in Wisconsin, uses sophisticated technology and data analysis to improve the quality of its bikes, particularly its high-end bikes that elite athletes use.

It’s 1997. Paul Andrews is taking a 20-minute spin on a Trek Y-Foil road bike. The bike is black, revealing its unpainted carbon fiber tubing. Attached at three strategic spots are sensors. These sensors, affixed to the bottom bracket, the head tube and the chain stay, are wired to a small “data acquisition unit,” a black box attached to the frame.

In effect, Andrews that day was taking an electrocardiogram of the bounces and stresses his route took up and down the hills of this farming country.

Their objective is to make the bike as light as possible while still absorbing as much road shock as possible, and at the same time making the bike more durable by reducing torque and pressure from the road and the rider. Sounds like a complicated dynamic programming and design program, and it sure is.

Trek has years worth of such data in its database, which it can mine for regularities. This use of technology and attention to empirical data has paid off: Trek’s bikes have consistently improved in quality over the past decade (my husband loves his Trek mountain bike!), and they are the bikes of choice for several professional riders, including one very famous one who recently won his sixth Tour de France. For more casual athletes the attention to design and time is not a big deal, but it is for Lance Armstrong:

But the vehicle he rides does matter. Saving the 32-year-old Armstrong as little as 10 watts of energy over the course of a 120-mile stage of the Tour will speed his trip by one minute. Not much? Last year, in his record-tying fifth Tour win, he edged German rival Jan Ullrich by 61 seconds after 2,125 miles of racing.

Trek uses their data gathering and analysis to put in lighter carbon fiber sheets in places on the bike where the additional weight was not necessary. The article describes the design software they use, how it incorporates the data from their trial rides, and how these tools enabled them to build better bikes to power Lance Armstrong to victory.

I’m not just mentioning this because I’m a cyclist and a technology weenie; there is interesting economics here. Trek has been using its combination of data analysis and design for almost a decade, and as their designs and materials progress, the older ones filter into their retail products. In fact, if you are so inclined you can buy an exact duplicate of the Trek Madone bike that Armstrong rode to victory in the 2003 Tour, for between $5000 and $7000. Compare that to the $2000 that Trek used to be able to charge for the top end of its range. But even the $2000 bikes in 2004 are significantly better than they were a decade ago, as a result of this technology and design filtration from the requirements of elite athletes into the standard retail range.

I consider this yet another manifestation of how improvements in technology, design, and materials that are driven by “super-users” improve our quality of life and how much bang for the buck we can get through technological change. It’s like so many other manifestations of this — sure, your typical bicycle consumer isn’t buying a fancy schmancy Lance Armstrong bike, and probably can’t afford one. But s/he still benefits from the market meeting the technological and design demands that such intense consumers possess.

2 thoughts on “Making A Better Bike Using Data Analysis

  1. Even down at the $400- or $500-level bike technology has been improving, often filtering down from improvements first seen at the high end.

    Now please stop talking about expensive bikes.

    My ol’ Trek is just fine.
    My ol’ Trek is just fine.
    My ol’ Trek is just fine.

    (I have to keep saying this to myself, so that when I next take it into the bike shop for $100 of fixing up I don’t end up thinking I need to spend 10- or 20-times as much for something new. Even if something new would be really cool. And much lighter. I could be faster. It would look nice. I would look nice. I would be really cool.

    Such are the delusions of bike lust.)

  2. I am looking for some archival data on common damage to entry level road bikes which may need improving. Your help would be much appreciated.


Comments are closed.