Skip to content
All case studies

BLP Experiential Studio · Experiential marketing · Brand activations

Proving a brand activation worked, without scrolling Instagram by hand.

BLP Experiential Studio designs trade show booths, pop-ups and festival activations. After every event the client asks what it actually did. I built the tool that answers, and it is worth having because it refuses to count anything it cannot tie to the event.

Client
BLP Experiential Studio
Sector
Experiential marketing
Built
Measurement tool and client-ready reports
Status
Delivered to BLP

The problem

The recap deck had photos. It had no numbers.

An activation can cost a brand tens of thousands of dollars, and the marketing manager who approved it has to defend it upstairs. The recap proved the booth looked good. It did not prove anyone saw it.

Doing it by hand meant opening the accounts of people you remembered being there, scrolling, screenshotting, and missing every post where the brand was on screen but nobody tagged it.

Buying it was worse. The tools that spot a logo in a photo are sold as yearly subscriptions for always-on brand monitoring. An agency needs one report for one weekend in one hall.

What I built

Every platform in, one honest ledger out.

One run takes the brand, the event dates, the hashtags, the venue, a few accounts known to be there and photos of the real booth. What comes back is a private report page the agency can put in front of its client.

  1. Sweep the public posts

    Instagram, LinkedIn, X, Threads, Facebook, YouTube and press coverage. Public posts only, read logged out, no platform logins.

  2. Find the people who were there

    Everyone with evidence of attending gets their own posts checked too, which is how the untagged posts get found.

  3. Look at every picture

    A vision model checks images against photos of the actual booth and packaging. When a picture is the only link to the event, a second model has to agree before it counts.

  4. Apply the counting rules

    A post counts only when the brand is in it and something ties it to this event. Everything else is listed with the reason it was left out, and kept out of the totals.

  5. Write the report

    Reach, engagement, the full ledger with a thumbnail per post, the biggest voices, what people said, and a methodology section that says plainly what the tool cannot see. Every number is labeled measured or estimated.

What changed

The first real run said no to almost everything. That was the point.

The first client run was Shan Foods at Natural Products Expo West in Anaheim, one of the biggest natural products trade shows in the country. Shan is popular all year, so its tagged posts were full of recipe creators who had never been near the booth.

Counting every tagged post would have produced a reach figure in the hundreds of thousands. The tool counted the 2 posts it could tie to the show and listed the rest, each with its reason. One early vision check had passed a big reel because it mistook another company's snack for the client's, which is why a second model now has to agree.

posts read
10,665
inside the event window
2,757
images checked against the booth
1,946
posts counted
2

Under the hood

How it’s put together

The agency site I built
  • Seven public sources: Instagram, LinkedIn, X, Threads, Facebook, YouTube, press
  • Vision checks with a second-opinion model on picture-only matches
  • 140+ automated tests on the counting rules and formulas
  • Every paid call cached, so nothing is ever paid for twice
  • Private, password-protected report page per event
  • No media-value inflation and no invented targets, by rule

More case studies