We will drown in sensors before we ever build a true internet of things

Just the way the smart home has single-purpose devices as opposed to overall intelligent systems, the development of intelligent roads, athletes and any other system made up of multiple components will feature single-purpose sensors for years before we ever get to unified systems– if we ever get to unified systems.

This is unfortunate for consumers who will have to wrangle many apps and also because having multiple platforms can slow the pace of innovation, but thankfully sensors are getting cheaper and we can at least fulfill some of the promise of the internet of things while we wait for an eventual standard or service to unify things to arrive.

I offer you today’s example of a fun single-purpose sensor for a niche audience: The Sony Smart tennis shot tracker. This $200 device affixed to an array of tennis rackets comes with an Android or iOS app and tracks up to 12,000 shots in memory. It can’t share information back to the app you use to track your fitness. So in order to, for example, get a sense of how many calories your strokes burned, you’ll have to pull data from this sensor and others and link them via spreadsheets, or perhaps use one of the up-and-coming Android Fit or HealthKit frameworks from [company]Google[/company] or [company]Apple[/company].

Sony's tennis tracking app.

Sony’s tennis tracking app.

There are similar devices for runners, such as the runScribe that my colleague Kevin Tofel profiled. The $99 device uses sensor data to measure and track metrics including stride rate, contact time of your foot with the ground, impact, and the angle your foot pronates, or rolls. There are a number of possibilities for more sensors or other single-purpose monitoring devices in clothing, especially for athletes, and sensor-infused shirts, socks and bras are on the market today.

Niche markets are everywhere

Athletes aren’t the only segment of users getting an array of single-use sensors. I think we’ll see them play a role in cars, too. Aficionados can hook computers into their cars via engine modifications or with OBD port devices like [company]Zubie[/company] or [company]Automatic[/company]. That lets people gather metrics from their cars and share them with their home or services like If This Then That, or just load them on an app.

Cars are a great example of why we’ll want single-use sensors. In many cases, cars already report a lot of data back to the dealer, mechanic or insurance firm. And any consumer with a smartphone can turn on an app like Waze or Google Now and get a sense of certain types of data, such as when she was driving and where. The challenge is that data isn’t openly shared yet, so while an automaker might not want to share all information with a mechanic or even a consumer, the consumer might want her own way to determine how her brakes are wearing.

My whiteboard scribble of car data. Conclusion was that we'll have a ton of relatively cheap sensors providing similar info to different parties.

My whiteboard scribble of car data. Conclusion was that we’ll have a ton of relatively cheap sensors providing similar info to different parties.

Or think about insurance companies. Companies such as [company]Progressive[/company] are offering customers lower rates on their auto insurance if they plug a box into their car to track various data. Progressive CEO Glenn Renwick said in March that he can’t wait to start getting automotive data from carmakers, so he can get out of the device business — sending customers boxes to plug into their cars. But given that auto manufacturers may want to charge Progressive for that data, it might make more sense for Progressive to get it from another provider or keep sending out boxes to customers.

As sensors get cheaper, and as we’re better able to gather and aggregate data through middlemen such as IFTTT or platforms from Apple or Google, the number of sensors should increase exponentially. That’s not because we really need five different sensors tracking different aspects of our movement, or five different automotive sensors telling us our miles per gallon. Rather, it’s because there are no standards, no real data marketplaces and no real reason yet for all the companies in these different elements of each ecosystem to work together.

The economics of sensors

Before you get excited and launch a sensor product, here are a few things to think about to see if it’s likely to become a successful business.

  • Cost: Adding a sensor to a device makes it more expensive, so you need to find an audience that will spend big to get this data. Athletes tend to spend significant amounts of money on their hobbies, for instance, so a $200 shot tracker or $100 device to measure your stride becomes a good gift or a worthwhile training tool for the dedicated.
  • Utility: This ties to cost, but it’s not enough to just track the runner’s steps, as the runScribe shows. You’ve got to come up with extra data that’s uncommon but relevant to your niche audience.
  • Partners: The goal with any of these sensors should be to link in with the other tools your customer is using. So runScribe should eventually tie to RunKeeper or the Jawbone fitness tracker. The combination of your utility and audience will determine the economics of those partnerships.
  • Start thinking of devices as shared or personal: This is an interesting one for most device-makers, because most of them think of their devices in relation to smartphone apps. A phone is a one-on-one device, which means that a lot of my sensors feel very personal, even if they could be shared. A perfect example is the runScribe. For $100, I’d love to share this device with others — my stride probably doesn’t change all that much so I can’t see a lot of value in wearing it everyday. It seems like a cool thing to offer at high-end running stores or gyms. Customers could clip it on in the store, take a few laps and get the data back on what shoes they should buy.

This last element is probably worth a post of its own at some point, or at least a conversation among many of the device makers we’ll have onstage at our Structure Connect event in October, so I’ll leave us with my vision of each person having dozens of sensors, ballooning to a hundred or more over time. Eventually I hope we figure out how to connect them via standards, middleware or services like IFTTT or other platforms, but for now, we’re just going to have to live in a fragmented world with a ton of apps.