Online, every intent is measured, predicted and acted on. Step outside and the record stops. TRAKKER learns how people move and decide in real space, and tells you what to do about it.
| # | Visitor | Category | Zones | Dwell | Returns | Score | Notes | Contact |
|---|---|---|---|---|---|---|---|---|
| 1 | J. Keller | Buyer | 6 | 18m 12s | 3 | 92 | Asked for pricing twice | |
| 2 | M. Suarez | Buyer | 5 | 14m 40s | 2 | 81 | Wants a pilot in Q1 | |
| 3 | R. Tan | Partner | 4 | 12m 05s | 2 | 74 | Intro to their ops lead | |
| 4 | A. Volkov | Buyer | 4 | 9m 58s | 1 | 63 | Returned after the keynote | |
| 5 | L. Meyer | Exhibitor | 3 | 8m 31s | 1 | 58 | Comparing two vendors | |
| 6 | P. Nakamura | Partner | 3 | 7m 12s | 1 | 49 | Budget confirmed | |
| 7 | S. Okafor | Press | 2 | 6m 04s | 0 | 41 | Writing a piece on the show | |
| 8 | D. Fischer | Buyer | 3 | 5m 47s | 1 | 38 | Needs a security review | |
| 9 | C. Almeida | Partner | 2 | 5m 12s | 0 | 34 | Passed to partnerships | |
| 10 | N. Petrov | Exhibitor | 2 | 4m 38s | 0 | 31 | Left contact at the stand | |
| 11 | H. Lindqvist | Buyer | 2 | 4m 02s | 0 | 27 | Follow up next quarter | |
| 12 | G. Rossi | Partner | 2 | 3m 26s | 0 | 23 | Short visit, low intent |
Ranked by real engagement, not by who got scanned. Illustrative names and scores.
Live simulation with sample data. Not a client deployment.
One behavioral layer underneath. Three very different reasons to care.
Every consented visitor ranked by real engagement, the same day. Plus a cost per qualified lead you can take to the budget review.
What exhibitors get →Understand how to price what, see which hall needs staff before the doors close, and redesign next year's layout on evidence.
What organizers get →The same model off the show floor: aisles, concourses and terminals. Trained where behavior is densest, built to travel.
How the model works →Beyond zone performance, the same behavioral layer answers things a headcount never could.
Sample values on real country boundaries.
Your engaged audience skews sharply Investigative and Enterprising. These are people who came to evaluate and do business, not to browse. That one fact reshapes programming and who each exhibitor should meet.
Interest profile built from registration and consented enrichment, never inferred from movement alone. Sample values.
A visitor decides whether to stop at your booth in about three seconds. Nobody measures that moment.
37% of exhibition organizers say their event format needs a full update, and visitor engagement is their joint top priority. UFI Global Exhibition Barometer, 378 companies across 57 countries.
Click → profile → intent → prediction → attribution.
Presence → movement → dwell → intent → nothing.
Heatmaps and dwell time are the starting point, not the output. A dashboard reports last week. The model tells you what to do next.
Movement located in real time, to sub-metre precision. Accurate enough to credit a specific stand, not just a hall.
The model reads the behavior underneath: approach, glance, dwell, hesitation, return, bounce. The whole visit, not one scanpoint.
You get the move, not the chart. Call this lead first. Send staff to the hall that is filling. Price this zone higher next year.
Our first full deployment covered a leading European investor's flagship community event end to end, in real time. The organizer walked in with headcount as their only metric.
They walked out knowing which groups engaged, where, and for how long, then redesigned the next year's layout around it.
Illustrative proportions only.
"Working with TRAKKER at [our flagship event] was a fantastic experience. By categorizing participants and utilizing movement tracking, we gained valuable data-driven insights into attendee engagement. Yuting, Alex and Ahmed provided exceptional support throughout both the planning phase and live on-site execution."
A visitor only shares their journey if they get something better back, so we built the thing they actually want. Live guidance during the event, and a recap worth sharing afterwards.
Visitors already wear a tag, and it has a QR code on it. Scanning opens a page. Nothing to download, one sign-in, done.
Where to go next, who's worth meeting nearby, and which hall is about to fill up. It updates as they walk.
A shareable summary of everything they saw and who to follow up with, which quietly markets your next edition.
No new app to sell. The co-pilot rides inside the loyalty app shoppers already have installed.
Where the item on their list actually is, what's worth knowing about it, and an offer that makes sense right there.
In-store behavior flows back into the loyalty profile, so the next online touchpoint knows what happened offline.
Good morning, Sam
Three things worth your time today
Strong match with your interest in robotics and automation.
Your route
Avoiding the queue at Hall 7 entrance
Ask anything
It knows the floor and your list
Your recap
Everything you covered
You saw more of the floor than 84% of attendees.
However the visitor already reaches you. We don't require a new habit.
No install at all. Scan the tag, get the co-pilot in a browser. The lowest-friction route, and usually the highest opt-in.
For organizers who want the full experience under their own event brand.
Drop it into an app people already open: your event app, or a retail loyalty app. The co-pilot appears where they already are.
Design concept. We're showing the direction before we build it. Sample content throughout. Tap the bottom bar to move between screens.
Hand the tag back. Nothing is recorded.
Movement with no identity attached to it.
Grouped by ticket type, never by person.Recommended baseline
Named, and only ever by explicit opt-in.
Most of the value arrives long before anyone is named.
Identity is never the default. The baseline groups visitors by ticket category and holds no personal records. Naming anyone takes an explicit opt-in that can be withdrawn.
In four steps. Sensors locate movement in real time, to sub-metre precision. The model then reads the whole visit rather than a single scanpoint: approach, dwell, hesitation, return.
From that it infers intent while the event is still running. With consent, the behavior links to a real identity and becomes a ranked lead, a live alert or a layout decision.
Yes, and it's designed that way rather than retrofitted. TRAKKER runs a four-tier consent model.
The recommended baseline is category-level intelligence under legitimate interest. It holds no personal records, and still answers most of what a floor needs to know.
Linking behavior to a named individual sits one tier above, requires explicit opt-in, and can be withdrawn. We are built and operated in Berlin, under EU law.
No. The model reads position and movement, not faces or images.
Where a space already runs cameras, that signal can be ingested as one more input. The intelligence layer never depends on identifying anyone by appearance.
Yes. A REST API for ingest and retrieval, webhooks for live event-driven signals, and full export into your CRM, BI tooling or data warehouse. Data flows both ways, and the data stays yours.
Wherever you want it. We sign a data processing agreement before anything goes live, encrypt what we hold, and run it on the servers of your choice, your own included.
Nothing is sold on and nothing is pooled with another client. You can take an export and walk away at any point.
We charge usage-based, per square metre per day. You pay for the floor you switch on and the days you run it, so a single hall for a single day is a real deployment rather than a year-long bet.
We put the number against your own floor plan in the demo. We are also taking on pilot and design partners this season.
We walk you through the live model on your own floor plan. We are also taking on pilot and design partners this season.
