Measuring the effect

Every measurement begins with a sentence you should write down before buying: what is this screen hanging there for? The answers differ – and they lead to different figures.

One goal per surface is enough: give a screen four jobs and you get four mediocre results. Also write down when you would be satisfied – not «more revenue» but «on days with screen advertising, more lunch specials go over the counter than on comparable days before». That sentence is the difference between an analysis and an opinion.

There are three levels, and they get mixed up constantly. The first is the technology: did the device run, did the content run, how often, how long. That is hard measurement; the player logs it itself. The second is the opportunity: how many people were within range, how many could have looked. That can be estimated, with honest imprecision. The third is the effect: did the screen change anything, in sales, in atmosphere, in the number of queries. That can only be grasped by comparison, never as an absolute figure.

Anyone who claims to tell you exactly how many people saw a slide is selling an extrapolation as a measurement. That is not dishonest as long as it is called that: footfall times a visibility factor times the share of playout time gives an estimate you can plan with, but not a count.

And some things simply cannot be measured. Whether somebody entered the shop because of the window, whether the waiting room feels calmer – there is no sensor for that, only the eyes of the staff who stand there every day.

The most reliable figure in any signage system is also the least spectacular: the playout log, known in the trade as proof of play. Every player records which file it showed, when and for how long, and reports it back. The result is a list that answers three questions: what ran, when, where.

On your own surfaces this is a control instrument. You can see whether the promotion really started on Monday, whether the Christmas slide ran on past New Year and whether the screen in the basement actually played or only claims to be switched on. In our advertising network the same log is the billing basis: anyone booking at third-party sites pays for playouts, and those have to be provable.

On third-party surfaces, watch three points. Is billing per site or as a flat rate across a network? What happens to playouts that did not run because of an outage? And do you get the log as a verifiable file or only as a pretty summary? That last answer says a lot about the provider.

If you want to know how many people come past at all, you need a source on site. The simplest is the counter at the door, which many shops already have; it delivers daily curves and weekly patterns and costs no data protection debate.

Anonymous sensors go further: small devices that count people in the field of view without storing images. They only output numbers – passers-by, people facing the screen, average dwell time. Gaze detection measures whether a face was turned towards the screen; thermal sensors count bodies rather than faces and work in half-light too. What is permitted and how to handle the duty to inform and anonymisation is covered in the chapter on network, power and security.

The shop till is underrated. It is the most honest data source in the building: it knows what was actually bought, in what quantity, at what hour. For most small businesses an analysis by item and period says more than any sensor on the ceiling – and nobody has to hang up a camera for it.

Effect shows up in comparison. A single figure says nothing – two hundred lunch specials sold is a lot or a little, depending on what came before.

The cleanest route is comparing two branches: two similar sites, the same item, the same week, screen advertising in only one. If you only have one site, compare periods – two weeks with, two without, alternating, so weather and holidays average out. And rotate the item being promoted: if the effect appears with three items, it was not the item but the surface.

An example from a furniture store: an armchair had been in row three for months and sold moderately. For two weeks it ran on the screens in the entrance, then two weeks not, then two weeks again. The till figures per period showed a clear pattern, and a hunch turned into a decision about which pieces go on the screens in future.

A screen ends at the pane of glass. Anyone who wants to know whether somebody then acted builds a bridge to the phone – usually a QR code. It is often used wrongly: too small, shown too briefly, too low down, with a destination that is unreadable on a phone.

Equally useful are promotional codes quoted at the till, and signing in to the guest wifi when the screen points to it. Also look at the tools you already have: searches for your name during and after a campaign, views and route clicks in your Google business profile. Those curves are not proof, but they move reliably along with everything else.

You do not need twenty key figures. Five or six that somebody actually looks at are worth more than a report nobody opens.

Note for every figure who looks at it and what happens when it tips; a figure without an owner is decoration. And do not add anything up into an overall score – a score built from six figures says nothing any more about what to do.

Once a month, twenty minutes: availability, playout, content age, effect. Four blocks, one page. Decisions follow every time, and they belong on the same page. If a site failed three times, a replacement unit goes into the store or the connection goes on the test bench. If a slide has been running for five months, it gets replaced or deliberately marked as permanent content. If a promotion produces no response, change the message first, not the hardware.

The most underrated figure is availability. Design gets discussed endlessly, outages rarely – yet a black screen costs more reputation than a mediocre slide. The mediocre slide goes unnoticed. A dark device in the shop window, though, is read: nobody here is looking after anything. And an installation still running last winter's promotion in June says more about your business than you would like.

That is why availability stands at the top of our report, ahead of all the effect figures. Anyone who first makes sure that every surface shows something current during every opening hour can then calmly turn to the two seconds this book is about.