How to know which poster worked, without analytics

You put the same code on a window, a flyer, a table tent and a van. Two weeks later business is up a bit. Which one did it?

The usual answer is to install analytics, or to pay for dynamic codes that track scans. Both work. Both also mean an account, a monthly fee, a cookie banner, and a service whose outage breaks paper already hanging in the world.

There is a much older trick that costs nothing, needs no account, and works on every phone made this century. You change one word in each code and let the message tell you where it came from.

The trick, in one line

A QR code does not have to carry the same thing in every place you print it. Make each copy slightly different, in a way you will see when somebody uses it, and the difference is your report.

Three of the code types on this site can do it out of the box.

1. SMS codes: a keyword per location

An SMS code carries a number and a message. The person scans, the messaging app opens with your number and the text already written, and they press send.

So write a different message into each one:

• Window code: “WINDOW”

• Table tents: “TABLE 4”, “TABLE 5”, and so on

• Van: “VAN”

• Flyer drop: “FLYER — MARCH”

Now every text that arrives tells you which piece of print produced it, in the first word. Fifty table tents, one phone number, fifty different answers — and no analytics anywhere.

It works past attribution too. A code on a bay in a car park sending “BAY 12” tells you which bay before anybody types anything, and a sticker on a lift sending “LIFT 3” does the same for a fault report.

2. Email codes: a subject line per placement

An email code can carry a subject line and a first sentence as well as your address. The subject is the useful part.

• “Enquiry — shop window”

• “Enquiry — trade show”

• “Enquiry — packaging insert”

Everything lands in one inbox and sorts itself. If you have ever set up a mail rule, you already know the rest: one rule per subject, one folder each, and the count in the folder is the number you wanted.

This is the version to use when the reply needs to be a real conversation rather than a text — support, quotes, applications, complaints.

3. Link codes: one page, several addresses

If your code points at your own website, give each print run its own address that redirects to the same page — one for the flyer, one for the window, one for the van.

Your web host's logs, or whatever stats page it already gives you, then separates them without you adding anything to the site. No third-party script, no consent banner, nothing new to explain to visitors.

Keep the addresses short and stable. A code is printed once and lives on a wall for years; the address inside it has to still work in three years, so pick something you will not want to tidy up later.

What this does not tell you

Being straight about the limits, because they matter.

You learn about the people who acted — texted, emailed, clicked through. You learn nothing about the ones who scanned and changed their mind, which dynamic codes do count. If the difference between a scan and an enquiry is the thing you are trying to study, this is not the tool.

But for the question most people actually have — which of these four things is worth printing again — counting the enquiries is the honest answer anyway. A scan that leads nowhere is not a result you wanted to pay for.

Try it on one thing first

Pick the two placements you are least sure about. Give each an SMS code with its own keyword, or an email code with its own subject line. Print, wait a month, count.

It is the cheapest experiment in marketing: the codes are free, the change is one word, and the reporting is a folder in the inbox you already read.