Nerd out with Ed: Facial Recognition

Facial recognition technology is increasingly being used and incorporated into our daily lives, but what exactly is facial recognition? and can marketers use it to measure the success of their events and activations? We sat down with our resident measurement guru, Ed Myhill, to understand what it is and how marketers can implement this exciting new technology to prove return on investment for their clients.

Q: What is facial recognition?

Facial recognition is a category of biometric software capable of mapping an individual’s facial features.

Q: How is facial recognition being used for marketing and events and how does it work?

Jack Morton is currently using facial recognition in two ways.

Experience enhancement

Apple’s FaceID is a familiar example for experience enhancement, setting the benchmark for facial recognition security. The system works by using the camera to take multiple stills of your face. By measuring distances between more than 80 nodes (points) across your face it generates a unique numerical sequence — creating a face print. When you come to unlock your device, it remeasures and matches your codes, giving you access. This system is incredibly secure, Apple claims the probability an individual can look at your iPhone X and unlock it is one in a million (versus one in fifty thousand for Touch ID).

Although not as accurate, current event technology works in a similar way, at events we take your face print to enhance and individualise your experience comparable to near-field communication (NFC), radio-frequency identification (RFID) or quick response (QR) technology. Some possible uses for experience enhancement include, check in or welcome messages and personalised activations.

Measurement and insights

The technology we use for measurement and insights is fundamentally the same. It’s still recording distances on the face but uses them for a different outcome. Cloud based applications and processes let us compare millions of faces, allowing us to understand what a face is doing and segment it into a category such as age bracket, emotion, gender and demographic. This has become integral for analysing success and gaining insights from events.

Q: What is the best way to measure attendee sentiment?

The best method of analysing sentiment is to associate it with a moment or part of an activation. We recently defined the success of our content by monitoring viewer sentiment combined with the quality of interaction — measured in this instance by time spent. To summarise, if a viewer watched our content for the whole segment with a positive sentiment we gave it a high score. If a person watched with neutral expression and left halfway through we gave that a low score. By looking at all our data we get a quantitative understanding of what’s working and what isn’t. This approach can be reused and adapted during any part of your activation or experience.

Q: Has Jack Morton been using facial recognition for events?

We have recently started using the system for measurement in Australia and our international offices have been successfully implementing this technology in other markets. To enhance attendee experience, our New York office used facial recognition for Royal Caribbean’s Sea Beyond. During registration guests uploaded an image of themselves, then when they arrived at the ship they were welcomed by name, resulting in both a notable and seamless first impression.

Q: How will facial recognition help drive return on investment for marketers?

In events it ’s notoriously hard to prove return on investment. Historically clients have had no interest in combining event data with their CRM data — as such it’s been difficult to accurately understand if the event experience has led to direct sales. However, we’re seeing this shift and as we continue to add depth to our measurements and insights, doors should continue to open.

In the meantime, using our propriety system we can measure the impact of our events, knowing a positive impact is a great way to drive consumers down the sales funnel. By combining the results of facial recognition with numerous other data points proving a good investment. Additional return can be realised when we add insights into our delivery package. When brands understand how different types of audience are behaving they can be more targeted, therefore increasing sales potential, and thus returning on their investment.

How is facial recognition different to existing event tracking technology?

Facial recognition allows us to measure the guest experience without interruption and because it’s less intrusive the data is much more reliable. When you conduct surveys, or entice checks ins with incentives, you hope that guests are with their feedback but there’s no guarantee.

What should marketers look out for when using facial recognition for measurement at events?

Be honest and let people know what you’re doing. Nothing is being stored and no personal or personal data is tied to any pixels. It’s used for measurement and insights to improve the consumer experience.

We’re making sure all our practices follow the European Union’s General Data Protection Regulation to ensure everyone’s privacy is protected.

What is the process to analyse the data gathered through facial recognition?

The program looks for certain indicators in an image that confirm an audience type or emotion. For accuracy, it references thousands of correctly identified images and matches the patterns, allowing an almost real time analysis.

Do you think facial recognition is the future for event registration?

We believe in following this technology closely, seeing how the public takes to uploading their photos alongside their meta data, monitoring accuracy and then choosing the right events to implement.

The tech is available and can improve check in times at kiosks and provide cool welcome messages as seen on previous events. In terms of accuracy, it’s at about 95% when the database of faces is under 1000 but this percentage decreases as the data set increases. That being said as the on-boarding process improves so will the amount of measurements taken and therefore the length and complexity of each face print, optimising the security and reliability of the system.

What type of events are best suited to use facial recognition as a measurement tool?

All events can gain positive and previously unattainable results from using facial recognition for measurement.

What are best uses of facial recognition you’ve seen in any industry?

Disney and Subaru are my two favourites — Disney are using it to measure the quality of their movies and how audiences are responding to certain sample footage. In the automotive industry, Subaru has a car that analyses your face whilst you’re driving and it will alert you if you look tired or better yet, when you’ve dosed off. Stopping needless accidents and keeping people a little bit safer.

If you have any questions about facial recognition or measurement and want to learn more about how we can amplify your next experience, please contact Ed at ed.myhill@jackmorton.com.au