Wednesday 30 January 2008

Google tests QR Codes

A minor story is doing the rounds on the Interwebs about Google's experimental adoption of the 2D Barcode for use in their print advertising. Google's 2D Bar Code info page is pretty thin on information but they make it clear they are using QR Codes* which have gained some some momentum in Asian markets over the past few years.  Joel Spolsky reckons they are just as stupid an idea as CueCat's barcode palns were when they tried something similar 7 years ago. He was definitely right then and CueCat definitively lost that battle but I wonder if he's as spot on as he usually is with this one.

He makes one important claim that is central to the use cases that QR codes (and Semacode from which it is derived) are supposed to enable - "typing URLs is not hard".

I beg to differ. Typing anything on almost any handheld device is a hassle. Yes the iPhone has a good virtual  keyboard and my Samsung Blackjack has a very good QWERTY Thumb-pad but entering any sort of complicated or long string of data into any mobile device is painful. Yes it's possible but it's not intuitive and the lack of fundamental capabilities like cut\paste on a large percentage of mobile platforms makes it even worse.

Typing is not very hard even on a 12 digit standard phone and it's certainly pretty easy on my Samsung Blackjack but something like certainly presents a nearly insurmountable obstacle in terms of usability. The general use case for these codes is to provide a simple mechanism for users to lift usable data from a physical object (which might be a display screen) nearby that may be time consuming, hard or damn near impossible to enter manually into a handheld device in the time available. In terms of use cases think for example about a user on an escalator in a metro system seeing an interesting advert for a concert they might want to check out wanting to capture the relevant details as they pass, or a user at a trade show wanting to get more technical data on an item on display. You can certainly take the time to accomplish these things by hand but the idea is to make it easy enough so that the potential number of people acquiring and making use of the data is increased significantly.

The above fairly tricky URL for a Manfrotto Tripod on Amazon UK becomes the following QR Code:


(There are quite a few online QR Code Generators - I used this one to generate this)

That captures reliably on my Samsung Blackjack in about 2-3 seconds. I haven't bothered to try and type out the full URL but even on a PC I'd be hard pushed to enter that in under 20 seconds even if I chose to navigate by going to Google and searching for "Amazon Manfrotto MN190ProB" so the convenience and usability arguments have some merit.

I used the QuickMark QR Code reader for Windows Mobile Smartphone to test that code. It's available for free (once you register) and it's small (255kb) and quick to install. They support a pretty wide range of platforms and at a quick glance appear to support a pretty comprehensive selection of the most popular handsets. QuickMark's reader is pretty snappy - I can unlock my phone, navigate to the reader, launch it and read the above code in about 10 seconds. QuickMark also understands some additional meta formatting so it will give you a link to the URL if the encoded data is a URL, or prompt you to add the data to your Contacts if it's a contact card format, send and SMS if the format is SMS and a whole lot more. Apart from QuickMark there are a few other offerings, Google themselves have put some work out on Google Code to encourage folks to build an Android QR code Reader and there is also a sample QR code reader under development for the iPhone.

Microsoft started some limited experiments with 2D Barcodes using the QR Code format via Windows-Live Barcodes which was open to the public to play with briefly in late 2006 but it seems to have gone into lock down at some point since then and I get access denied pages when I try to link to it's current (supposed) home at

There are some limitations - the most obvious one is that mobile phone cameras are pretty pathetic and even though the QR code standard supports codes containing up to 3-4K of text the practical limitations of camera resolution (fairly low quality 2 megapixel) prevents my phone reading anything more dense than 145 character codes. That prevents some interesting use cases at the moment but it is more than sufficient for linking a phone to a fairly complex URL that can then do a lot more - especially if the link contained geo coordinates for example.

Anyway I'm clearly a fan. I've played with QR codes a lot in the past and I've done a lot of work with Barcode scanning equipment and believe that the reasons why barcodes are so useful in a commercial\industrial context (reliability,convenience,speed, context control [scan that particular box, read the code on that particular bag ..]) could apply very easily to consumer contexts if there was a sufficient critical mass of consumers equipped with devices that made it easy to consume them.

That last point is a big problem though, while I now think that a majority of the current generation of mid range consumer hand held devices are capable of supporting a 2D barcode reading application today only a very, very small minority actually have one installed and working. Unless the manufacturers and vendors pile onto this almost universally (as happened in Japan) then it's simply not going to get the momentum it needs to take off.

So unlike Joel I am quite glad to see that Google has joined Microsoft in doing some experimental work with them as that might be a sign that we will start to see a significant enough percentage of devices becoming available that support QR code scanning by default. Once that happens then maybe we will all join the Japanese in having our phones tell us about the nutritional content of our burgers.

* QR code is trademarked by Denso Wave, inc.

No comments: