“We aren’t Counting Eyeballs… Yet.”

“We aren’t Counting Eyeballs… Yet.”

A primer on how a joke made in 2017 can influence reality to an unexpected degree.

“We aren’t counting eyeballs to measure outdoor impressions… yet”

This used to be the cheeky thing I’d say in a presentation when someone in the audience asked how we counted impressions. I wasn’t really positive that we’d ever make it into this territory… but here we are.

If you’re reading this on the Adder blog, it’s likely that you already know that we work extensively in GPS location and movement data analytics. And if you didn’t know that — well, that’s how we do things. We look at billions of location data points a day, reported from over 50 million mobile devices, to understand how people interact with outdoor ads. Understanding the importance of that is vital for our advertising partners.

50 million devices worth of coordinate data is no small task to process, but measuring 20% of the population’s movement for the most accurate and up to date location data is worth it — we are able to do incredible things with that information. We can tell many, many stories with our data.

But, what about the people that aren’t in our location data stack? How do we count those impressions on OOH ads? How are we supposed to include the people that aren’t connected to the internet a thousand different ways?

There was an obvious solution to our developers:

This was an early test of some of our software and hardware , detecting bodies with realtime image captures.
Body detection was the first step, of many. But DOOH advertisers need even more than that, and we’re looking at the entire analytics picture.

we decided that we should count the eyeballs.

I know, I know — we’re living in Minority Report. This was inevitable, with the recent Cambrian explosion of AI and machine learning hardware, the crypto slump, and the work being done in neural network development.

The best I can offer to assuage any fears is remind everyone that we don’t collect any personally identifiable information, or PII. We understand the weight that such data carries, and frankly, we don’t want to take any chances.

At Adder, we’ve focused heavily on vehicle based advertisements and hardware. Which is to say… we don’t often share too much of our special development project progress. Times are changing, though, and we really can’t help but start leaving a trail of breadcrumbs that lead to our primary development project.

As we’ve moved further and further into spatial awareness and audience location analytics, we’ve found ourselves at the bleeding edge of tech development. We’ve found ourselves at a… final frontier, if you will. 

Ian and Brandon get classified as being located near an outdoor advertisement. Eyeballs = counted.
Here I’m showing Brandon Bush my demonstration version of the body detection/ooh impressions counting neural network.

If you saw my last blog post, you’ll recall that we got early-access to the NVIDIA Jetson NANO devices, which we named George and Gracie.

While some others out there have run outdoor impressions detection on advanced hardware, we’re the first to do it on a $99 piece of hardware.
Now you know one of the many reasons we’re so excited to have those devices, and I’ll share more details about what we’re doing with them, soon.

Since you’re dying to know, I’ll go ahead and say it:

We can count the eyeballs now.

We can count the IP addresses, too. And the GPS coordinates. Can’t forget the Bluetooth Signals. We count everything now. It’s sorta our thing.

That’s all for now — I’ve got a database to optimize with my team — and 30 billion data points worth of compressions, indexes, reads, and writes on the Adder Cloud is no joke!

Ian Gerard, CEO & Director of Robo Consciousness

You don’t have to leave, yet! For more information about our development check out another article, or better yet — check out the out of home analytics portal!

To leverage Adder Analytics data for your out of home advertising inventory, contact our sales team at sales@adder.io


Your email address will not be published. Required fields are marked *

two × 2 =