Check out a short documentary about our results!
Psychological and Emotional Targeting
Algernon disruptively improves online targeting by analysing people's emotional and psychological status. It offers a natural complement to the behavioural and contextual ad targeting that is being currently used in online media. Our algorithm derives information from unconscious cursor movements made by users while browsing. With Algernon, clients can target customers based on personality or emotional profiles and improve retargeting and user experience.
How does it contribute to online targeting?
Algernon can radically improve the efficiency and versatility of online targeting. Clients may:
- Transfer personality or emotion based segmentation methods used by offline media advertisers to the online market, e.g. be able to target masculine or self-confident users.
- Predict how users will behave in the sales funnel based on factors other online methods can’t measure, e.g. learn when users get excited about what they read, see or hear in real time.
- Web analysis for publishers and advertisers to understand the dynamics of users' emotional reactions. Providing insight on brand performance.
What can it learn about users?
Algernon's artificial intelligence is able to very accurately identify individuals and to a certain extent infer psychological attributes by analysing movement patterns. This works like — just better than — handwriting analysis.
Algernon classifies users by analysing mouse / touch screen / gyroscope movements along such pre-tested factors as:
- excitement, vigilance, alcoholic influence, masculinity, aggression, distress, sex, age;
- mouse events and other actions carried out on a publisher or e-commerce website that are presumably driven by emotional, psychological or personality factors.
What’s the further potential?
The algorithm is very flexible and versatile and has an amazing array of potential applications beyond psychological and emotional targeting. A few examples:
- Provide secondary user authentication and fraudulent account login alerts to any user registration based service provider: social networks, email providers, online banking, payment and e-commerce sites.
- Quantify the actual number of individual users of a specific shared device, account or software. Providing publishers and advertisers with improved ad reach data.
- Help medical research by improving the diagnosis methods of disorders such as dyslexia.
Technology and Results
Algernon is an innovative algorithm that within seconds identifies user specific biometric features in a digital environment. It analyses multidimensional movement patterns, such as computer mouse movements in real time. The algorithm is based on the finding that gestures made with pointing devices are unique to individuals, just like handwriting or walking.
The algorithm has been developed by a multidisciplinary university research group of 12 high-profile university researchers with 4 PhDs. The scientific algorithm combines mathematics, psychology, physiology, handwriting science and IT.
What are the results?
Algernon’s user authentication feature has been tested on 1,100 users and resulted in
- a security level far exceeding that of competing technologies including fingerprint scanner;
- best mouse identification accuracy at 2.7% crossover error rate;
- digital pen identification accuracy at <0.5% crossover error rate, about 10 times better than rivals.
How does the user classification work?
The tool classifies users following these steps:
- A simple, open source script monitors and records users’ cursor / gyroscope movements and generates an anonymous movement database.
- Algernon’s algorithm finds unique patterns of unconscious movement that for example characterise someone’s handwriting.
- Using complex statistical analysis, the tool eventually learns what patterns to look for and can make the classification / authentication of users in real time or by analysing recorded data.
What input devices does it work with?
Algernon works with and can be specialized for a variety of input devices such as:
- computer mouse, touchpad, touch screen, smartphone gyroscope;
- 2D and 3D digital pen, head and eye tracking devices, GPS.
Our Team
Gergely Hanczar, PhD – research and product development
Gergely is responsible for developing Algernon’s capabilities and products.
He has led the multidisciplinary research group for 10 years that created Algernon’s algorithm. Gergely has also been the Head Professor of the Research Department at the International Business School, Budapest. As an entrepreneur he built a successful book publisher.
Gergely holds 6 degrees from ELTE University of Sciences, including a PhD in Education, a BSc in Physics and a BSc in Descriptive Geometry. He is also a certified handwriting expert.
Tamas Zelczer – business development
Tamas is responsible for commercialising our products.
He is a successful strategy consultant and operational business manager. He has consulted blue chip clients and invested in and mentored start-ups in various sectors. Tamas has worked for a leading financial institution where he led global strategic initiatives.
He holds an MBA from INSEAD and an Msc in Economics from Corvinus University of Budapest.
Bence Golda – software development
Bence is responsible for implementing our Algernon’s applications.
Bence is an outstanding senior software developer and he designed and implemented Algernon’s software. Bence led the development of an online computer vulnerability management system and he built from scratch a graph-based, semantic database system that supports versioning.
He holds an MSc in Computer Science and Economics from Corvinus University of Budapest and an MSc in Computer Science from Budapest University of Technology and Economics.