2012-02-25

The curse of success.

Expectations.

A few weeks ago I bought a Polar FT60 heart rate monitor along with a FlowLink USB interface that is supposed to make it possible to transfer data from the device to my computer.

Let's say a bit about my expectations when buying the product.  I already own a Polar heart rate monitor, but after I moved to a new house I have mislaid it.  After looking for it for a couple of weeks I decided to just get a new one.  In any case, the heart rate monitor from Polar I had earlier allowed me to transfer the measurements as files and then plot my heart rate as a function of time.  So this is what I minimally expected to be able to do when I got the new monitor.

What I got.

It turns out I can't.  I can not get the files from my watch onto my computer.  What I can do is to download a really flimsy piece of software from Polar which will take the data from my watch and ... upload it to Polar's website.

So then I can get the data from there?  Of course not.

How about getting those plots of heart rate as a function of time?  Nope, you can't have that.

You can have a summary of your workout;  how much time you have spent in different "heart rate zones", but actually getting a plot of your heart rate over time: no.

If you want those plots you will have to get a different watch and pay extra for a different piece of software that, "oh by the way" only runs on Windows.

Okay, so this is a huge step back from the previous products from Polar that I've owned.  And to be honest, I feel cheated.  This is not what I thought I was buying.  Why would I think that Polar made a product that was greatly inferior to something they made 5-6 years ago?

In fact, I don't think the product is worth its price if I can't get the data off it.  A more reasonable price for the product would be somewhere in the $40-50 range.  And it should come with a very clear warning that it doesn't do what previous products did.  I think Polar should take more care in labeling their products "WARNING: WILL NOT PROVIDE ACCESS TO ACTUAL HEART RATE DATA".

What is wrong?

So exactly what went wrong here?  Well, the main problem is a quite widespread one in the consumer electronics industry:  the manufacturer makes some assumptions as to what the customer wants or needs and then makes the mistake of thinking that they are the ones best positioned to add that value to their product.

The software products that Polar provide appear, at best, barely adequate.   In particular their online offering is clumsy, confusing and largely useless.  It is the sort of software that creates more work artificially rather than make things easier.  And it doesn't even fullfil its primary objective:  properly displaying captured data.

But if it is so easy to spot, why do they do this?  In short: because they are successful.  As long as Polar are making heaps of money they are not likely to examine what they could be doing better with any eagerness.  It is very likely that if you ask any manager at Polar about their software products and their online services, they will be proud of these.  Because they don't know any better and because they have no reason to think they are lacking.

There are three problems with this.  The first is obvious:  they make products that are annoying.  The second is that if they have a vulnerable flank when it comes to competition (and before you say that there is no real competition:  that's what Kodak and Nokia thought too).  The third is that they have settled for their current plateau -- I think they could have been more as a company.

Partnerships?

So what if Polar partnered with third party software vendors?  It would appear that they, to some degree, have done so.  But it isn't very well done.  It is really unclear what products work with what software and the website does a poor job of directing the user.  There are a lot of pages on the Polar web sites, but very little useful information when you actually look for something.

I'm not sure who is to blame here.  Polar appears to be a heavily marketing-driven company and not a technology company, which might make it hard for them to partner with more technology savvy companies.

Reading between the lines.

I've been in the software industry for a lifetime and Polar has a very familiar smell.  If you look at their product range and in particular the different technologies for communicating with their watches and handling the sensor data, they are all over the place.  There is no obvious and consistent direction.  Rather than having a clear focus and direction they are a bit like Nokia used to be before they got credible competition:  trying to fill every imaginable niche and completely neglecting the developer.

They probably do the electronics and sensor bits well, but all the infrastructural parts and the software seems to be the result of a very poorly lead engineering department.  It is also unlikely that this will improve -- because they make tons of money.  You can't convince people who make tons of money that they are doing something wrong.

What they ought to do.

I think there are a few things Polar should do.  The first is to really let go of the software parts and try to focus on doing the hardware.  Their insistence on doing software and "services" only damages their product because they do it so badly. In the short term their software engineers should be working on making it easier to integrate with Polar and forget about doing consumer-facing stuff.  In order to do consumer-facing software they need to clean up their act considerably.

If you open up and allow third parties to build on your solutions you get others to invest in your product and to depend on its success.   Forget about clammy handshakes and signed contracts:  open up to all developers.  Everything needed to develop software for Polar products should be easily available on the website.  Without even requiring registration.

The prime example of getting others to invest heavily in your platform is the iPhone.  Apple could have chosen never to open it up to third party developers -- to only allow their software to run on it.  They didn't.  And because of this the iPhone is now perhaps the single most important device in the history of computers.

Polar is probably never going to have this kind of mass appeal, but if they opened up their products to third party developers, Polar would become a much, much more valuable brand than they are today.

(Actually, Polar could have mass appeal.  Sooner or later the field of biometric telemetry for health monitoring is going to explode and someone has to make the hardware.  The problem is that I don't think it is in Polar's DNA to capitalize on this opportunity).

I think Polar should:
  • Publish the specs for protocols and data formats for all their watches.  It doesn't really matter if the specs are badly written or messy;  having some specs is better than having none.
  • Publish source code for drivers, parsers and such on, for instance, Github
  • School their executives in open source and the Maker Movement.  Here's an exercise I think would be useful:  buy some Arduinos, hire a hardware hacker and send the entire management team on an off-site to learn how to program hardware.  Make them understand that this is somethign a lot of people know how to do.
  • Build a proper software engineering department that spends more time writing software that makes it easy to innovate on top of Polar products than crank out third rate software products.
With regard to their product lines I think they need to tighten things up quite a lot.  Polar is living in the early 2000s with regard to over-segmenting their products.  You cannot create buzz for any one product if you have a gazillion semi-equivalent products you are trying to push.  Each product will end up getting a weak identity -- which in practical terms means consumers have no idea what to buy.  But I shouldn't have to tell you marketing-savvy people this?  Surely your expensive mercantile educations covers such trivial insights? :-)

As for me:  if I could just get the measurements off my watch and into a CSV file or something I can work with, I'd be really happy.  If you know of some way to do this, please let me know.  Even just importing the values into a spreadsheet is much more useful than Polar's useless personal trainer website.

Oh, and if you happen to be an executive at Polar:  I would love to learn more about your company.  If you cover my travelling expenses and donate a decent amount of money to Doctors Without Borders or Digitalt Personvern I'd be more than happy to spend a weekend discussing the above.

12 comments:

  1. I think you should take a look at Garmin and see if their products meet your requirements better. At the very least, they've published documentation for external developers here: http://developer.garmin.com/

    I've taken one of my bike rides from last summer, http://connect.garmin.com/activity/97848707 and exported it to TCX format here: http://dl.dropbox.com/u/12580451/activity_97848707.tcx

    I think that one would contain most of the data you want.

    PS: Not in any way connected with Garmin, just a happy user of their HRM, after I got tired of my Polar HRM not speaking to my Mac.

    ReplyDelete
  2. Wow, that looks just like what I wanted. Given that format it is trivial to write a parser that can convert the data to whatever I would need. I checked out the Forerunner 610 and although it is about twice the cost of the Polar, being able to export the data so easily makes it worth the higher price. If I had known about this when I was in the store I would never have even considered buying the defective Polar.

    I wonder if Polar would be willing to refund the FT60 on the grounds that they did not adequately inform me as a consumer that their product does significantly less than I had reason to believe. (I don't feel like being a prick to the sports store that sold me the watch, though I will have a word with them and inform them that they are selling a defective product and that they ought to either warn customers or drop Polar products)

    ReplyDelete
  3. Knowing you, I'm slightly surprised, that you have not already proposed a fix for this issue. After all the data travels trough you MAC and onto their web site. How much effort would it take to capture a copy?

    ReplyDelete
  4. Sometimes I don't want to have to pull the MacGyver stuff. Sometimes I want the stuff I buy to not be broken.

    Put in a bigger perspective: it upsets me when I see companies make products that cannot fulfill their potential.

    ReplyDelete
  5. Sounds like a dreadful user experience!

    I can heartily recommend Garmin Forerunners. Yes, their included training diary isn't that useful, but at least it is running locally, and you can both look at graphs and export data to gpx files (which includes HRM data).

    It is also possible to retrieve data directly from the ANT service, which is the service receiving and storing data transferred from your watch. (I use http://www.matstroeng.se/quickroute/en/ -- aimed for Orienteering.)

    ReplyDelete
  6. I feel your pain. Just now, I wanted to see this movie at the website of my ISPs "digital services" branch, and after I paid and all, there was a spinning beachball-like animation running for about a minute, until the web page helpfully replied "Plug-in Failure". Dunno whether I'll be charged, dunno what to do. It's in beta, I guess.
    --
    chr

    ReplyDelete
  7. Christian Nybø: you are not alone in disliking the state of that. believe you me.

    ReplyDelete
  8. I had much more fun than that.

    I retried starting the movie, on the second try I got a message that my IP address was outside of Norway.

    Third try started the movie.

    Now, if I want to see a movie over at Apple iTunes, they realize that I have little bandwidth and download the movie before it starts. This player did not appear to precache the movie, as there seemed to be little network activity after pressing "Pause" in the player. With my 1.5 mps, direct streaming did not work.

    Instead I rented the BlueRay hardcopy at the parlor down the street, upgraded the software on my PS3 to adapt to the disk's new encryption key or method and finally watched the movie. Take that, Pirate Bay!

    Does not seem like I'll get a refund for the online experience, though. On the website of said digital services branch there is a question mark. Prior to buying something on the site, it seems the customer is expected to follow the link at the question mark to read up on the FAQ, especially "Technical Requirements". At least that's what customer support told me through email today.

    So there, I got to donate NOK 39 and spend my weekend discussing the above with an executive at said digital services branch. ;)

    ReplyDelete
  9. Although I agree with your conclusions I feel I should mention you can export data from the polar online site. My exercise site (http://turan.no/ even has a parser for it (https://github.com/torhve/Turan/blob/master/apps/turan/polaronlineparser.py)

    But indeed, the garmin products are much more developer friendly.

    ReplyDelete
  10. @thveem: it is actually a bit worse than that. it would appear that the FT60 doesn't store HR samples -- and if it does, it won't export them during sync. I should probably take this up with the store where I bought it since I specifically asked if it would store HR samples. on the other hand, Polar didn't label the product in any way to indicate that it was a considerably lesser product that the Polar watches I have owned before. I didn't think I would have to check specifically to figure out if the product they sold me was defective by design.

    I hate products that are marketed in a manner where key features are drowned out by all the useless bullshit features people aren't going to need anyway.

    ReplyDelete
  11. Right, this is worse than I thought. And basically useless for a being a HR monitor.

    ReplyDelete
  12. Polar indeed has taken the direction of becoming more 'closed' or 'proprietary' with each succeeding product generation. The S610, released ~2003, can use any generic IR dongle and when the original watch band breaks, it could be replaced with a generic band at any watch shop. The RS400 released in 2006 came with a custom band which is only available from Polar and requires Polar's own "lipstick" IRDA dongle. But, the S610 and RS400/800 can still directly transfer data to the PC and the data can be subsequently exported as a text file for analysis. I suspect the motivation for the Flowlink interface and web-based upload is plainly to sell additional hardware.

    Perhaps the FT60 not storing HR sample is because Polar wants to skimp on flash memory. It should take less memory space to store the max/min/average of an exercise session than to store samples at 15s intervals.

    ReplyDelete