53 comments

[ 117 ms ] story [ 1172 ms ] thread
"Now, I'll be the first to admit that a website about salt is not the most riveting thing in the world. But that's exactly the point! The costs associated with setting this up are close to zero."

Just because we can doesn't mean we should. Why put a QR code on the salt packet if it serves no purpose. What is the cost associated in the number of people who wasted their time scanning that code?

I totally agree QR codes are great in that they are cheap to deploy, but I've seen very few uses in the real world for consumers that are actually useful. Only one I can really think of using myself and which was useful is the Google Authenticator app and the BlockChain.info equiv.

Based on the salt comment, I went through my cupboard and scanned several products. Here's the results:

• The Red Bull can leads to a full page advertisement for Red Bull Mobile, which is more than useless.

• A generic oak lettuce leads to a recipe page containing, you guessed it, salad suggestions.

• A cleaning spray lead to the companies Facebook page, with a very large image asking me to like their page.

• A bottle of reasonably pricy champagne lead to a page full of gleaming reviews, and a copy of the information that was on the label anyway.

• A well known bleach product (scanned weeks back) also contained a link to their Facebook page.

It's hardly enough of a reason for me to scan these in the future.

You are correct. It's simply another advertising portal. I completely ignore QR codes for the reasons you mentioned. I think of them as hyperlinks in another form that must be "clicked" with an app.
I was just peeling a banana this morning that had a QR code on its sticker. I suspected it would be a page about bananas, but unless it told me this banana I was about to eat had been recalled, I really couldn't care less.
Seems some technology inspires development thereof, but few are interested in using it. Fun to create, boring to use.

I've an itch to put a QR code on my business card. Having no illusions that anyone will in fact care enough to scan it or be wowed by the results if they do, I've refrained.

Momentum strikes me as a big issue. Not just the effort which goes into using it (dig out phone, fire up app, position camera, read results), but the average payoff in doing so (as adjacent post notes, QR codes give a consistent "so what? WTF did I waste my time for?" experience).

The only time I ever used a QR code was at AnDevCon, when a speaker gave a prize to whoever could scan the QR codes on his slides the fastest. It felt gimmicky and silly.

I've seen them on movie posters, bags of coffee, and at work, but I've never felt compelled to use them. Has anyone seen a great implementation of QR codes?

The big problem for me is, if I want to find out about a movie, product, etc. I want an objective opinion, not a sales pitch. I know whatever is on the end of the QR code probably won't help me, but Googling for movie reviews, product reviews, etc. will.

You are not a representative sample. If you look at the statistics for QR codes, you'll see that they are fairly widely used.

I agree that going to a product's QR code for an objective review is silly. But to see, say, a movie trailer or get an instruction manual for a bit of furniture, I think they're very useful.

> You are not a representative sample. If you look at the statistics for QR codes, you'll see that they are fairly widely used.

I would be fascinated to see those statistics, please share them.

Certainly! Here are stats from the London's bus timetable QR http://shkspr.mobi/blog/2012/01/real-qr-statistics-from-tfl/ http://shkspr.mobi/blog/2012/04/tfl-qr-followup-5000-scans-p...

Here are stats from Tesco (large UK retailer) http://shkspr.mobi/blog/2012/10/no-one-scans-qr-codes-apart-...

Finally, some QR stats from the back of a train ticket. http://shkspr.mobi/blog/2012/12/no-one-scans-qr-codes-apart-...

So, in the UK at least, well designed QR codes get a significant response from "normal" people.

Bus time tables, giveaways and product advertisement are normal things that people are interested in. A train ticket with a QR code is going to be sitting in someone's hand on the train while they've got their phone in the other hand. If they are curious about getting a 2012 games ticket, or if they're just bored they might play with this QR reader thing.

A salt package however... Maybe you're in a restaurant, do you pull out your phone? Do you hide away the salt package just so you can read the QR code later? This isn't even described, so why do you want to read it?

We already have a ubiquitous machine-readable way of identifying things, barcodes. Just because you own a barcode scanner doesn't mean you go around scanning the multitude of barcodes on your stuff. In fact, we get kind of blind to the number of barcodes on everything because they are used on everything and given without context.

If QR codes are overused on everything, maybe leading to things as mundane as an infosheet on salt (oh, this is high in sodium, who knew?), or forgotten broken links, they're just going to be ignored except when given appropriate context.

Are "normal" people really interested in getting detailed information documenting the unique ID of the salt package they purchased?

I was rather hoping for some more independent statistics, those links are all to posts on your blogs with your numbers but anyway let's take Tesco for example.

Tesco is huge in the UK but let's be very conservative and estimate that they have 10 million customers a month (in reality I expect they have more customers than that in a day) apparently 6000 of those are engaged enough with QR codes to use them which is 0.06%. That's not a very impressive statistic from where I'm sitting.

I don't know about great, but I'm in the market for a new car. When I went to a Subaru dealership they had QR codes on their vehicles which led to a page with stats, videos and other information. I had already researched the vehicles I was looking at, just wanted to see them in person and in various colors, so it wasn't that useful to me that day. However, in the past I've done some window shopping without the prep, would've been nice.

EDIT: Wehicles? I need more sleep.

I found them useful at bus stops in Japan. Scan it to go straight to the online live timetable listing for the stop you're at showing the current time/date (upcoming 5~10 buses). Sure you could type in a URL and populate a form or search for+install an app, and many stops have expensive digital displays, but this was really quick and easy.
According to the senior editor of The Atlantic back in January 2012, 'QR Codes Are the Roller-Skating Horses of Advertising'

http://www.theatlantic.com/technology/archive/2012/01/qr-cod...

  > QR codes are an intermediate technology at best, a novelty at worst.
Well, I'm not talking about using it for advertising per se. I'm looking at what happens when every item you own has a machine readable tag. Hence "Internet of Things" .

And, most intermediate technologies tend to stick around for a long time. Look at AM radio - supplanted by FM, DAB, satellite - and still going strong.

AM has a couple of things working in its favour. Power requirements and incredible range with those power requirements. It's no coincidence that sports, religion, and talk shows are popular on that band.

QR codes could be useful but between URL shorteners, geofencing, and RFID tags they are yet 1 entry in a crowded arena. I'm not exactly sure why it didn't take off but I assume a certain fruit company didn't think of it so didn't care to include it.

AM was not an intermediate technology - it was supplanted by FM it wasn't created with FM and satellite already available.

Passive RFID is going to be way more useful.

I never scanned a QR code in my life, albeit I gotta admit I've already searched on libraries to create them(talk about totally not scratching my own itch!)

the best use case I've ever thought for QR Codes is vandalism.

If you get to the point that things are scanning each other silently and automatically, QR codes are a fine solution.

It would be much easier for a hypothetical 'smart' refrigerator to read a QR code on an item being placed inside than trying to get a clean barcode read from arbitrary angles or visually recognizing every possible object.

Similarly, a Google Glasses sort of tech would have a far easier time offering context-aware services if it could passively identify objects by a QR code. No-one would actively scan a poster QR code to order tickets to a movie. But if you were looking at a movie poster with a QR code, you could say "remind me to buy tickets to that" and the glasses would have a much easier time than if they had to visually recognize the particular poster and parse out whether it was for a film or tv show or stage production and figure out where to order said tickets, etc.

So... as a shortcut to effective computer vision and understanding that human beings will not ever actively scan them, sure. Why not?

+1 can't overstate removing inertia of work from the equation.

for similar reasons, RFID or similar should also carve nice niche over time as Internet becomes ubiquitous and costs decrease?

While I appreciate your points, personally, I'm wary. I've seen stickers overlaid on top of printed codes, and/or social hacks to dupe people into downloading malware.

I went to the media lab not too long ago and got to see VR codes. Imperceptible patterns ftw. This doesn't necessarily relate to your points about the opportunity cost of placing QR codes on packaging (after all, VR codes are designed for digital displays), but I think QR codes are sort of a stepping stone.

The QR malware is an area I've addressed [0] - there are some hacks out there, to be sure.

I love the idea of imperceptible patterns but then you have the problem of how you let the user know that there's a pattern there to scan! It's the same issue with NFC - you need a symbol that a user will recognise to mean "Scan me"

[0] http://shkspr.mobi/blog/2011/12/how-to-prevent-qr-hijacking/

Fair point on imperceptible patterns (until we have something like google glass :b).

On the matter of addressing QR malware, you must appreciate that in order for the technology to be safe and widely adopted, the _masses_ must be educated on all the points you bring up. Every time I realize how such information needs to be understood by the lay person, I'm led to believe that QR codes won't become 'the internet of things'. My 2c.

The problem isn't getting a user to recognize that they need to scan something... it's that the user needs to scan something! We need to get things automatically scanning other things.
(comment deleted)
In your example, you forgot the cost of having less surface for the logo and general product advertisement, and ugliness.

  > Compatible with every phone with a camera.
  > No need to build or use a dedicated app.
Surely this is incorrect, except for phones with built in QR code readers - which certainly isn't every phone with a camera.
Notably both Android and iPhone require you to install a third-party app in order to scan QR codes. I think if (when) this changes, QR codes will really become frictionless, since you no longer need to instruct the end-user how to first install a scanner app with a help label next to the QR code.

Eventually people will "just know" what to do with QR codes, just like they already know what to do with http:// or www. prefixed strings.

Google Goggles scans QR codes and it came installed on my Android phone.
He might be referring to competing systems like Microsoft's Tag which require a proprietary reader app, or the various image recognition apps for scanning billboards and movie posters. QR Codes are freely and openly licensed.
As others have pointed out, I meant that there is a generic QR scanner for every phone. I didn't need to download the "Condiment Scanner" which was incompatible with "Heinz Scanner" etc.

Interestingly, Windows Phone has a QR scanner built in to its visual search tool which is pre-installed with every device. Some Android phones have it installed. ISTR that BlackBerry comes with a scanner pre-loaded as part of BBM - but I haven't checked.

The only good use of QR codes I have ever seen is in the Nintendo 3DS game Pushmo. It allows you to create levels in the game and share them publishing a image with a QR code in them.
URLs killed QR codes. Get over it.

Easier to read, easier to share, easier to remember.

Harder to enter on a touchscreen. There's no reason they can't coexist...
QR codes are annoying to use even on phones that support them. Get a short URL and forget QR codes.

Typing proving to be too much? How about working on better OCR?

Homographic disambiguation. It's hard for even the best AI to tell the difference between 1lI!| etc.
If the AI has a hard time differentiating, I'm sure your human users will have as much of a challenge.

Pro tip: Don't register A1lI.com and expect people to be able to type it in.

The amount of time it takes to make an OCR-friendly URL on a product is zero.

Tell me how you enter a QR code on a touch screen. At least the url you can enter, QR code fails, as your argument 'entering them on a touch screen is harder'.

That problem is pertinent to both, so QR codes are still worse than urls.

Your argument is invalid.

Now, if you mean reading the QR code or URL on any device, let me tell you that QR codes are dots, if you misread a dot you fuck up the meaning. So OCR engines can EASILY read a URL and the same rule applies, if they miss a dot, they may confuse an I for an L or a 1, but still easier than a QR code, because 26 letters an 10 numbers are easier to configure in an OCR engine than irregular dots on a square.

URLs win.

I rest my case, your honor.

s/touchscreen/touchscreen device/. Even iPods have cameras now.

OCR has been failing us really badly for 40 years now, and not for a lack of trying. In fact, OCR is so bad that the one way we use to distinguish humans from machine is by forcing them to do OCR.

QR codes were designed with targeting/aligning areas as well as tons of redundancy built in, there's a reason why they can be easily scanned in realtime from a video feed at an odd skewed angle in the dark, whereas OCR doesn't even work properly on books scanned on a flat scanner.

OCR has always sucked and will always suck. Sorry.

The salt packet is cute, but I don't think QR codes can go small enough to dominate this area for long.

There are inherent limitations: http://www.qrstuff.com/blog/2011/11/23/qr-code-minimum-size

Besides the regular Qrcode there's also Mirco Qrcode[1]. Also the Reed–Solomon error correction in qrcodes take up a lot of datablocks. If a new standardisation for mirco qrcodes would allow for a version without error correction you could store even more.

[1]http://www.qrcode.com/en/microqr.html

"Tiny sachets of salt an pepper. Created in their millions. Given away for free the world over. Each stamped with a unique ID which can be recognised easily by a computer."

Um, do we really need every salt packet to have a unique ID?

If we really do, then the cost is NOT trivial. What machinery are you going to have in place to assign and track IDs? What server is going to hold all those IDs and respond to queries about when/where/what the salt packet is? What high-speed ink printers do you have in place on the assembly line to print these codes?

Well, it's unique to that brand of salt packets, or SKU.

Still, it's not like we'll run out of IPv6 addresses, is it :-)

Well, to play devil's advocate against myself, food products are an area where you want unique IDs on things.

You want to be able to scan that head of spinach and have the reply be "PRODUCT RECALL: E.COLI CONTAMINATED".

But I don't think QR codes are really the answer here. I want it to be wireless so when I walk up to the fridge my smartphone starts going off with warnings.

Robot barf on everything? Please, stop.
2001:0db8:85a3:0042:1000:8a2e:0370:7334

That's the internet of things.

  |2001:0db8|
  |85a3:0042|
  |1000:8a2e|
  |0370:7334|
Or that, if you like squares.
That's a beautiful way of formatting IPv6 addresses.