Flying Fish
Guests answered short questions to redeem, and the answers built a profile of who drinks it.
Service 05 · Software
Audience analytics and ROI
We measure every impression, interaction and conversion and show it in one place, so you can prove what worked.
Ask what we can measure
For brands, agencies and venues
End to end
Every impression, interaction and conversion. One place to see them.
Follow one guest from the screen in the venue to your dashboard. At every step we count something, say how sure we are of it, and send it to Hello Brand OS.
From the screen to the report
Your brand on the screen
Brand Box plays your campaign on the venue's own screens, next to the sport guests came to watch.
- Potential viewsEstimated
- Footfall: people in the venueMeasured
- Dwell timeMeasured
Where it comes from: Brand Box play logs, pod.ai and Lamp.ai devices in the venue (supplied with Moving Tactics), and the audience model for that kind of place.
They take part on their phone
A code on the screen opens the campaign in the phone's browser, with no app to download.
- ScansMeasured
- Sign-ups, with consentVerified
- Game plays and answersMeasured
Where it comes from: the campaign itself. Consent is recorded at sign-up and travels with the data, under POPIA.
They buy and redeem
Rewards are redeemed at the bar or the till, and outlet sales are matched to the campaign dates.
- RedemptionsVerified
- Sales by outletVerified
- Conversion rateMeasured
Where it comes from: reward codes scanned at the bar, till slips, and the outlets' sales before, during and after the campaign.
You see what worked, in one place
Everything above lands on one dashboard in Hello Brand OS, by venue and by week, while the campaign runs.
- Sales before, during and afterVerified
- Reach by region and audienceModelled
- Return on ad spend and brand liftModelled
How sure each figure is: Estimated Modelled Measured Verified What the labels mean
Hello Brand OS
Your brand, in one view.
Brands, agencies and venues look at the same live results, each from their side. These are real dashboards from our venue platform.
Who sees it
Brands
Your brand across every venue and campaign: who saw it, who took part and what it sold.
For brandsAgencies
The same live view, ready to report to your client. We deliver; you keep the client.
For agenciesVenues
Footfall, dwell time and what each campaign brought in, venue by venue.
For venues
What you get
- Impressions and dwell time.
Seen in Carling Black Label
- Engagement rates and conversions.
Seen in Jameson Carling Black Label Budweiser
- Brand lift and recall.
- Reach by region and audience.
Seen in Carling Black Label Jose Cuervo Heineken Vuse
- Campaign performance in real time, with transparent reports.
Seen in Carling Black Label
Brand liftIllustration
- SurveysMeasured
- Social listeningModelled
- Direct feedbackMeasured
Live campaignIllustration
- Sign-upsVerified
- PlaysMeasured
- Potential viewsEstimated
- RedemptionsVerified
Proof of what worked: who took part and what it was worth, venue by venue.
Our platform doesn't just track, it proves. First-party data comes from the people who take part, with their consent, under POPIA. Where a figure is an estimate, the report says so.
How it works
Four steps take a campaign from raw data to a decision about the next one.
From data to ROI
-
01Bring the data together
First-party data from our CRM and mobile platforms, synced across retail, digital and in-store touchpoints, with consent kept.
-
02Measure the same way
Closed-loop sales tracking and consistent metrics: return on ad spend, lift and dwell time.
-
03Engage with content
Games and experiential media on the phone, in venues and in store.
-
04Connect the tech
Online and offline data joined up, so results can be attributed across platforms.
Where the data comes from
Two kinds of data, joined up: what people tell us when they take part, and what the places around a screen tell us about who passes by.
First-party: people who take part
- Scans and sign-ups
- Game plays
- Answers
- Redemptions and sales
Joined up in one place, with consent kept and people counted in groups.
Third-party: the places around a screen
- Mobile location data
- Live traffic
- On-site sensors
- Social posts
First-party
From the people who take part
Sign-ups, game plays, answers and redemptions from our mobile campaigns, across venues, shops and phones. Collected with consent and managed under POPIA.
Third-party
From the places around a screen
Audience signals from location data, traffic and on-site sensors, used to plan and measure media. Anonymised and counted in groups, never as named people.
In the venue, pod.ai and Lamp.ai devices, supplied with Moving Tactics, add footfall and dwell-time data to Hello Brand OS: how many people are present and for how long, not who they are.
| Source | What it tells us | Where it falls short |
|---|---|---|
| Mobile app location data | Which places people visit, so we can group them into audiences and estimate age and gender. It also spots repeat visitors. | Fewer signals at roadside and highway sites, and it depends on how many people carry smartphones. |
| Live traffic | The best count of how many pass a site, using lanes, road type, direction, vehicle types and how many people are in each vehicle. | On its own, it says nothing about who those people are. |
| On-site sensors | Counts at the exact spot: new and repeat visitors, how long they stay and how close they come. Some models also count people and vehicles by camera. | Placed at chosen sites, so they are used mainly to check and tune the model for each kind of place. |
| Social posts | Age and gender, from public posts tagged at that location. | Depends on how many people tag their location there. |
No single source is perfect. So the model first works out what kind of place a screen is in, such as a mall, a busy road or a station, then weights each source for that kind of place. Census, survey and traffic reports act as a check.
How a figure is made
Raw signals are not results. Each number in a report goes through three steps.
Capture
We collect the signals around each site, within about 200 metres, for every hour of the campaign.
Clean
We remove repeat signals that do not mean a real chance to see the advert, so nothing is counted twice.
Weight and report
The sources are combined with the weighting for that kind of place, then turned into reach, views, frequency and audience profiles.
Why cleaning matters. Someone has dinner for two hours at a restaurant with a screen outside. Raw location data would count their phone again and again. Cleaning sees that the phone is not moving and caps the count, so the report does not inflate the audience.
Before, during and after
- Before
Plan on real patterns
Each site's audience from the last three months predicts who will be there on your dates.
- During
Watch it live
Audience data for each site, date and time band is refreshed daily, so we can adjust while the campaign runs.
- After
The full report
Five days after the campaign ends: reach, ad spend, engagement across channels, redemptions and sales, plus proof that it ran.
- Next
Plan the next one
What worked, where and when goes into the next plan, so each campaign starts smarter than the last.
What the words mean
The same terms, used the same way, in every plan and every report.
- Potential views
- The number of times people had the chance to see the advert. The out-of-home version of impressions.
- Unique reach
- How many different people had the chance to see it at least once during the campaign.
- Frequency
- How many times, on average, each of those people had the chance to see it: per day, and over the whole campaign.
- Dwell time
- How long people spend near the screen. Longer dwell means more chances to take the advert in.
- Ad play
- One showing of your advert on a digital screen, for example one 15-second spot in a loop.
- Views per play
- How many people were there for each play, not for the whole hour. hourly views × dwell time in seconds ÷ 3,600
- Share of time
- The part of a screen's airtime that your campaign gets.
- CPM and eCPM
- The cost of a thousand views: campaign cost ÷ views × 1,000.
- Day part
- A slice of the day, such as morning or evening, used to plan and compare.
- Audience segment
- A group with a shared habit. A phone seen at coffee shops at least three times in a month counts as a coffee lover.
- Proof of performance
- Evidence the advert ran: photos, video and, on digital screens, the play logs.
- Conversion
- Someone acting on the advert: a scan, a sign-up, a redemption or a sale.
Worked example: one person, five days
A campaign runs for five days. One person walks past the screen once every day.
- Day 1
- Day 2
- Day 3
- Day 4
- Day 5
- 5Potential views
- 1Unique reach
- 5Campaign frequency
Attention, not just views
We mix numbers with what people say, so you see whether the advert was noticed and whether it worked.
Viewability
Whether the advert was actually seen: how much of it was in view, and for how long.
Engagement rate
How people interact with it: taps, swipes, scans and other signs of active attention.
Dwell time
How long people spend with the advert, which shows how deep their attention went.
Conversion rate
How many acted. Read with attention data, it shows which places turn attention into action.
Brand lift
The rise in interest in your brand after the campaign, from surveys, social listening and direct feedback.
Contextual fit
How well the advert suits the place and the platform it plays on.
What a profile shows
Every scan, play and answer adds to a picture of who your customers are and how they buy.
Profile
Age, gender, location, LSM (living standards measure) and interests.
Lifestyle
Sport, travel, cooking, music, comedy, fashion, health and fitness, pop culture, technology and business.
How they buy
Channels, rewards and loyalty, how often and when they buy, and how likely they are to use a coupon.
What reaches them
Screens with a call to action, reach and frequency, and the app, email, social, web and in-store channels they respond to.
From a real campaign: Flying Fish
- 60.5%of engagements happened at weekends
- 59.86%happened before 17:00
- 27 y 6 maverage age of those taking part
- 58.78%chose female or other
-
Day of the week
Measured60.5% of engagements came from Friday to Sunday.
- 9.3%Mon
- 7.6%Tue
- 8.7%Wed
- 14.0%Thu
- 23.3%Fri
- 25.0%Sat
- 12.2%Sun
-
Age
MeasuredAn even spread, with an average age of 27 years 6 months.
- 18 to 2120.0%
- 22 to 2526.1%
- 26 to 3024.2%
- 31 to 4018.8%
- 41 to 507.9%
- Over 501.8%
- Rather not say1.2%
-
Gender
Measured58.78% chose female or other.
- Female 50.3%
- Other 8.5%
- Male 41.2%
Answers from the guests who took part in the Flying Fish campaign at Tiger's Milk.
Where we have done it
Each project says how far it got: results, results to follow, a concept or work in progress.
-
-
Carling Black Label
A results dashboard for the Varsity Cup 2025 campaign.
-
Tiger's Milk
The venue platform behind the dashboards on this page.
-
Heineken
Daily sales tracked across the outlets during the campaign.
Works with
-
Hello Brand OS
Where every result is collected and reported.
-
Campaign Builder
The campaigns that create the data.
-
For agencies
Reporting you can hand to your client.
Prove what your next campaign did.
Tell us what you want to know. We will show you what we can measure, and how.