I asked for the menu.
The waiter pointed at a small black-and-white square stuck to the table.
Of course.
There was no menu.
There was a portal.
So I took out my phone, opened the camera, positioned it above the table, waited for the little box to appear, tapped the link, waited for Safari, dismissed something asking about cookies and finally arrived at a PDF containing twelve dishes.
I had come for lunch.
I had accidentally entered a digital transformation programme.
The embarrassing thing is that I cannot even pretend to hate QR codes.
I love them.
I am obsessed with QR codes.
Especially dynamic QR codes.
A static QR code is loyal. It points somewhere and stays there.
A dynamic QR code has options.
Print it today. Change the destination next month. Send people to a campaign on Monday, a form on Tuesday and an emergency apology page by Friday afternoon.
The square remains exactly where you printed it.
Behind the scenes, it has completely reinvented itself.
I find this disproportionately exciting.
A printed object that can change its mind.
Technology has rarely understood me so well.
Then one of mine stopped working
A while ago, I had a dynamic QR code that worked perfectly.
Mostly.
The QR scanned.
The destination existed.
The website was online.
On Wi-Fi, everything worked.
On mobile data, nothing.
My beautiful little square had apparently become a matter of national telecommunications security.
Naturally, I became obsessed with finding out why.
My first diagnosis was:
The telco has blacklisted my QR code.
This is not technically the most precise explanation.
But it is not completely ridiculous either.
Dynamic QR codes often contain a short URL that redirects the user to the final destination. Bitly, for example, describes its dynamic QR codes this way: the printed code can stay the same while the destination behind the short link changes.
That redirect creates another link in the technical chain.
And links can be judged.
Google Safe Browsing maintains systems for checking URLs against lists of potentially unsafe resources, including URLs reached through redirects.
Telecommunications companies can also operate security layers capable of blocking suspicious domains, websites or URLs. Legitimate links can occasionally be caught in those systems too.
So yes: a QR journey can work perfectly on Wi-Fi and fail on a mobile connection because something in the network or security stack dislikes a domain or redirect.
Without access to the specific operator's logs, I cannot say that the telco definitely blacklisted the link.
But that did not stop me from explaining the situation to everyone as though Vodafone had convened an emergency board meeting about my QR code.
I had apparently become too dangerous for 4G.
This was not the QR code's original career plan
The QR code was developed by DENSO WAVE in 1994 for industrial use, originating in Japan's automotive manufacturing environment.
The problem was practical: traditional barcodes could not carry enough information efficiently, and factories needed something faster and more capable.
Thirty-two years later, it is stuck to a café table asking me whether I want avocado.
Quite a career.
The QR code has now colonised almost every available flat surface.
Menus. Posters. Packaging. Receipts. Museum labels. Hotel rooms. Conference badges. Parking meters. Shop windows. Business cards.
There is almost certainly a QR code somewhere explaining how to use another QR code.
And most of this happened for a very good reason.
QR codes solved a genuinely useful problem:
How do you connect something physical to something digital without making people type a URL?
Point camera.
Tap.
Done.
Brilliant.
But businesses eventually discovered something even more attractive.
A QR code does not only connect a physical experience to a digital one.
It can replace part of the physical experience altogether.
The menu disappears.
The brochure disappears.
The registration form disappears.
The printed instructions disappear.
Sometimes the employee disappears.
And sitting in the place where all those things used to be is a tiny square saying:
You do it.
We did not remove the friction
This is the part I find more interesting than the QR code itself.
Because one of the great promises of digital transformation was the removal of friction.
And sometimes we genuinely removed it.
Other times, we simply moved it across the table.
The restaurant no longer has to print menus.
Excellent.
The customer now needs:
- -a smartphone
- -a charged battery
- -a functioning camera
- -an internet connection
- -a working redirect
- -a browser
- -a mobile-friendly website
- -enough remaining psychological stability to reject the newsletter popup
From the restaurant's perspective, the process has become more efficient.
From the customer's perspective, lunch has acquired system requirements.
That distinction matters.
Because companies often measure the steps they have removed from their own operation and call the result customer convenience.
The customer is measuring something else.
The customer is measuring how much work remains.
We did not always remove friction. Sometimes we just reassigned it.
And QR codes happen to be one of the most elegant pieces of administrative outsourcing ever invented.
No uniform.
No salary.
No onboarding.
Just a small square and the quiet assumption that you brought the necessary equipment.
A QR code is a door, not a room
None of this means QR codes are bad.
They are extraordinary.
A QR code on packaging that opens detailed sourcing information can add something useful.
A QR code beside an artwork that launches an audio guide can improve the experience.
A QR code on an event poster that lets someone register immediately is better than asking them to remember www.something-events-2026-registration.com.
And dynamic QR codes are even better.
You can update a campaign after printing.
Correct a destination.
Change seasonal content.
Measure scans.
Run different destinations without producing thousands of new physical materials.
That is genuinely useful.
The mistake is assuming that putting a QR code on something means the problem has been solved.
It hasn't.
A QR code is a door, not a room.
It transports somebody somewhere else.
What happens next is the actual experience.
And sometimes what happens next is extraordinary.
Scan.
Loading.
Cookie banner.
Accept.
Newsletter popup.
No.
Location permission.
No.
Download our app.
Absolutely not.
Create an account.
Why?
Verify your email.
I wanted soup.
At some point the printed menu begins to look like one of humanity's great technological achievements.
Paper had excellent uptime.
One tiny square. An impressive amount of infrastructure.
This gets even stranger with dynamic QR codes.
What appears to the customer as one tiny square can depend on:
- -the phone camera
- -the operating system
- -the mobile network
- -DNS resolution
- -the short-link provider
- -the redirect domain
- -security reputation systems
- -the destination server
- -the CMS
- -the consent layer
- -the webpage itself
That is quite an infrastructure diagram for discovering whether the restaurant has tiramisu.
And this is the hidden trade-off of modern digital systems.
Their simplicity at the surface is often created by extraordinary complexity underneath.
When everything works, that complexity disappears.
Which is exactly what good technology should do.
But when one dependency fails, the entire illusion collapses.
The customer does not care whether the problem is DNS, SSL, Safe Browsing, a redirect rule or Mercury being in retrograde.
They scanned the square.
Nothing happened.
For them, your QR code is broken.
This is why testing the destination page is not enough.
If a QR code is important to a customer journey, test the journey.
Scan it.
Try another phone.
Try mobile data.
Try Wi-Fi.
Check the redirect.
Check whether the page actually works on mobile.
Check whether the user understands what they are scanning for.
And if the information really matters, give them another way to reach it.
A short URL.
A printed instruction.
A person.
Sometimes even a menu.
Redundancy sounds terribly inefficient until the efficient thing stops working.
Efficiency has a point of view
QR codes belong to a much larger story.
Self-checkout.
Chatbots.
Apps.
Automated customer service.
Online check-in.
Digital ordering.
Passwordless authentication that somehow requires three passwords.
Technology repeatedly allows companies to make their operations more efficient by inviting customers to participate in them.
Sometimes that is a fantastic trade.
I would rather check in for a flight on my phone than queue at a desk.
I would rather scan a QR code and pay instantly than wait fifteen minutes for someone to bring a card machine.
The point is not that self-service is bad.
The point is that self-service is still service design.
The customer must receive something valuable in exchange for doing the work.
Speed.
Control.
Flexibility.
Convenience.
Better information.
If the only meaningful beneficiary is the company's cost structure, then we should at least stop congratulating ourselves on improving the customer experience.
Because that is where the language becomes suspicious.
We call it frictionless.
We call it digital-first.
We call it self-service.
The customer calls it:
Apparently I work here now.
The QR code won
And yet I remain devoted to the QR code.
I will continue generating them.
Tracking them.
Redirecting them.
Testing them.
Changing their destinations without reprinting anything and feeling, for reasons I cannot fully explain, like I have briefly defeated the laws of physical media.
The QR code deserves its victory.
It took a thirty-year journey from automotive factories to the restaurant table and became one of the most successful bridges between atoms and pixels ever created.
But its success also reveals something useful about the way businesses think about technology.
Every time we automate something, digitise something or remove something from the physical experience, there is a question worth asking:
Did we actually eliminate the work?
Or did we simply give it to somebody else?
Because the QR code won.
The menu became software.
The table became an interface.
Lunch acquired a dependency stack.
And somewhere along the way, we quietly changed the meaning of convenience from:
“This requires less effort.”
to:
“We no longer have to print it.”