Skip to content


Market standards – reflections on bicycle drivetrains

The chain drive is a venerable and apparently immutable part of bicycle design. It is highly efficient and long-established, with on-going incremental innovation. This is largely evident in chain width, which is driven by (sorry) other incremental innovation in the drivetrain on bikes, largely reflecting the number of sprockets on the rear wheel. It should be noted that the drivetrain consists of a number of discrete parts – hub, cassette (sprockets), derailleurs, chain, shifters and cranks and chainrings.

The two major component manufacturers in cycling – Shimano and Campagnolo – have each carved out their own particular specifications in this area, to ensure a lack of inter-operability between their components. Shimano is currently offering a 10 sprocket system and Campagnolo’s now goes up to 11 (I’m not sure they recognise the Spinal Tap reference). One cannot, with ease, mix and match drivetrain components from the two. The relatively recent entry of a third party – SRAM – into the full component market has complicated matters a little further. SRAM had previously been known for base-level products like chains, cassettes and shifters and then naturally progressed to other parts. In moving into full component systems, SRAM have adopted Shimano’s cassette and chainring spacing, so these parts (cassettes, chains and chainrings) are interchangeable but offer their own, incompatible shifters and derailleurs, brakes etc. Shimano has even made compatibility between subsequent generations of its own components problematic, with recent iterations of its combined brake levers and shifters (so-called STI units) requiring new generation derailleurs and even brakes too (due to variations in ‘cable pull’).

There is also evident segmentation in component ranges, reflected in part by the number of ‘speeds’ on offer – 8, 9, 10 and 11 sprocket systems co-exist in the marketplace for road bikes (and 8, 9 and 10 in the mountain bike market). Even within a particular specification – eg 10 ‘speed’ – numerous variations will be offered by each manufacturer, reflecting different price points and quality levels. For road components, Shimano has Dura Ace at the top of the range, followed by Ultegra, then 105 (all 10 speed) and then Tiagra and some more I don’t really know about, as the lowest spec Shimano component any of my bikes has is 105 (on my hack/commute bike, and even that has some Ultegra bits). Campagnolo has Super Record, Record, Chorus, Centaur and Athena. SRAM’s range is narrower – Red, Force, Apex and Rival. In each case, differentiation of the ranges is through materials, specification, build quality and, to some degree, feature.

However, all of these are based on chain drives. In the last couple of years, into the bike market has come a different, albeit similarly venerable, approach – the belt drive. What has now made this viable is technological progress in materials – rather than rubber-based belts, a new synthetic, polyurethane-based material. The Gates Corporation‘s “carbon belt” is an unbreakable, lightweight material requiring no maintenance, no lubrication and offering incredibly long life. Quite a number of bike manufacturers (from bespoke, like Seven cycles, to mainstream, like Trek) offer belt drive models. However, the belt is best-suited to single speed or hub gear-based bikes, so this tends to rule it out for road race models. Again, there is market segmentation by product going on here.

So why aren’t ALL (or even most) new bikes belt-drive now? Cost may be one aspect – a chain can be very cheap (a few quid at manufacturer scale), whereas the belts – at this point – are anything but. More significantly is the incompatibility with existing drivetrains (and, in fact, the design of existing bike frames). To run a belt drive, one needs a belt drive compatible ‘chainring’, a hub gear (as it won’t work with derailleurs) and a frame that can accommodate the belt. How hard can it be to ‘accommodate’ a belt, you might wonder. The answer is “very”. By its very nature, the belt in a belt drive is a single, unbroken band. A chain, on the other hand, is made of lots of separate links and can be (more or less) easily split and rejoined. To fit a chain to an orthodox bike, you break and then re-make the chain, fitting it through the rear triangle of the frame. A belt drive can’t be split, so the frame has to split instead! This has to be designed in at the concept stage – it’s not something you can really retro-fit to an existing bike.

So, there is a de facto market standard of chain drives, and this is reinforced by the incumbent players – much or all of their product ranges reflect chain-based transmission (though Shimano and I think SRAM have some expertise in hub gears, which is a belt requirement). Note too that the belt drive option comes from a new entrant to the market, not one of the existing players. For certain applications, the belt drive’s advantages are very compelling (I’m attracted by the thought of a tandem timing belt drive, for instance – and this is much easier to fit to existing kit). These advantages may well outweigh the cost aspect, so it’s potentially a top-down market disruption, as Nick Carr would describe it. For the product segments where belts are deemed possible, it may well become the dominant form of drivetrain in coming years.

Another aspect of the bike component market where market standards carry less sway is non-drivetrain components – here you find many more rival manufacturers as the dominant firms have not been able to successfully impose exclusive product specification – brakes, for instance (though Shimano’s recent updates to their STIs have been an attempt at doing just that, one might argue). Maybe I’ll save that for another post…

Posted in design, disruption, market standards, technology.