16 September 2026 · 4 min read
A catalogue that measures itself. What 340 sessions said
Three days at a trade show, eleven tablets, fifty hours of active time. A paper catalogue will not tell you what anyone was looking for. This one did.
A salesperson at a trade show gets two minutes. The visitor stands there, looks, asks about something that is not on the table. A paper catalogue weighs what it weighs and ends where the paper ends. It cannot show a kit where products connect. It cannot tell you afterwards what people actually cared about.
I built a tablet app that runs without internet and, once the show is over, hands back a report on visitor behaviour. One file, no build pipeline, font embedded locally. In an exhibition hall it has to open instantly, not wait for venue wi-fi that is not there anyway.
What I measured
Those last two numbers are two different brands running on the same engine. Seven out of ten people who picked up a tablet at all went into a specific product page. That is not a number you can get from paper, because you cannot get any number from paper.
Same engine, two different behaviours
The interesting part was not in the summary but in the breakdown. At the first brand people entered through the product grid: one hundred and seventy-six entries. At the second, through the kit strip: two hundred and seventy-five. Same app, same layout, two different patterns.
Kit screens from the catalogue8
Average session length split too: five and a half minutes at one brand, almost fourteen at the other. Fourteen minutes on a tablet at a trade show is not browsing. It is a conversation in which the tablet is a third person.
The design lesson is simple and boring: there is no single right way into a catalogue. A brand selling individual products needs the grid. A brand selling complete solutions needs the kit strip on top. Had I shipped one entry point for both, one of them would have lost half its interactions and nobody would have known.
Why offline
Running without a network was not a technical whim but a risk calculation. An exhibition hall is the worst possible radio environment: thousands of devices, steel structures, an organiser network shared by every exhibitor. An app that needs the network will fail at the worst moment, which is when somebody important is standing at the stand.
The guide for the sales team, six screens6
So the catalogue asks for nothing. Everything, product photography and colour swatches included, sits on the device. Statistics accumulate locally and only come off after the show. Zero dependencies means zero reasons to fail.
What I would not repeat
I moved product data by hand out of the asset management system, because there was no export that would hand it over whole. That was the biggest time cost in the project and the least design-like part of the work. Doing it again, I would start with a conversation about the export, not with the interface.
The second thing is the lack of lifestyle photography. The catalogue fell back on packshots, products on white. It works, but it is a catalogue, not a story. A photograph of a person in the field sells differently from a jacket on a hanger, and at a trade show that gap showed.
Two languages, one tap
The show is international, so the catalogue is bilingual, switched with one tap in the corner of the screen. That sounds like a detail, but it was the only interface element the sales team used consciously and mentioned unprompted. The reason is mundane: at a stand there is no time to translate a product name, and switching language mid-conversation is faster than hunting for the word.
The design consequence is that both versions have to be complete from day one. A catalogue where half the descriptions are Polish only looks worse in front of a foreign client than an English-only catalogue.
What the report contained
The post-show report is not a list of clicks. It is three things: which routes people took into the catalogue, how long they spent in each section, and which products were opened most often. Nothing more, because nobody would read more anyway.
The stats page (code protected)
The most practical part is the list of products that were opened but were not physically on the stand. That is a ready answer to the question of what to bring next time, and the only thing in the whole project that turns into a purchasing decision with no interpretation at all.
What this means for a client
A catalogue that measures itself costs roughly the same as one that does not, and after three days it hands back a map of what people at the stand were looking for. That map is worth more than the catalogue, because it tells you what to make, what to promote and which way to walk the next visitor.