Turning science into strategy

The attention economy has never been more competitive. Every day, we are confronted by thousands of images and messages, online and offline. Businesses invest heavily in design and marketing in the hope that their products will hook our interest amidst this intense noise.   

Back in 2009, Professor Peter W McOwan (School of Electronic Engineering and Computer Science) was working on software that analysed how humans process visual information. 

A fortunate encounter with a marketer opened his eyes to the value of the research for designers. Together with Dr Hamit Soyel, he translated their fundamental know-how into a support tool for designers, Dragonfly AI. Since its launch in 2018, creative professionals have used the tool to create designs that appeal to consumers, instantly. 

Research

Dragonfly AI’s story begins in 2009 when McOwan developed 'spot the difference' software to analyse how humans process visual information. Together with Soyel in 2011, McOwan first applied this fundamental research to robots to give them human-inspired visual skills. 

But, after meeting a marketer at their Science Museum presentation in 2013, the team learned their work could have a very different outlet: visual communications.  

McOwan, who sadly passed away in 2019, and Soyel built on their earlier research to refine the algorithm that now underpins Dragonfly AI. The algorithm reveals what people will notice most within a design, and what they will miss. Extensive testing with users ensures the algorithm produces reliable, actionable insights.  

Stronger designs, quicker and cheaper 

McOwan and Soyel co-founded a spinout business in 2018 to market Dragonfly AI. The tool is driven by a patented algorithm based on their knowledge of how humans process visual information.  

Dragonfly AI’s scientific basis takes the guesswork out of design processes, allowing designers to create, with confidence, clear and digestible visual communications that stand out on the shop shelf or webpage. 

It provides instant feedback on designs and guides decisions on: 

• the size of visual elements  

• how to arrange elements 

• which colours to use 

• what (not) to put in the background 

• the most effective typography 

Feedback from Dragonfly AI’s customers reveals that the tool’s instant insights cut creative production and testing cycles by up to 50%. Traditional design processes can take 4-6 months. 

They further cut the costs of testing designs by up to 40% by reducing the need for expensive studies, such as focus group consultations and eye-tracking tests. A traditional testing cycle costs $30,000 -70,000. 

Driving product sales  

Creatives and marketing teams in over 50 of the world’s biggest consumer product companies, including Unilever, Danone and Coca-Cola, use Dragonfly AI to steer their packaging and advert designs. 

Feedback from Dragonfly AI’s customers reveals that consumers now pay more attention to their products on shop shelves and e-commerce platforms like Amazon. One global frozen food brand reports a 45% increase in on-shelf attention across their top three product lines after using the design support tool. 

And the increased attention translates into stronger sales. One global household brand saw a $300,000+ sales uplift by improving their visual presence on the shop shelf. Another forecasts a sales lift of $440,000-500,000 on Amazon after tailoring their packaging design for the platform. 

Dragonfly AI’s basis in fundamental human visual processing makes it particularly valuable for global brands as its insights apply across different cultures and markets.  

From healthcare to education: future uses of Dragonfly AI 

For Soyel, his foundational research in human visual processing, and the algorithm that followed, has immense potential beyond marketing. 

Future uses could include: 

• Healthcare: clinicians could diagnose disease from medical imaging faster and more accurately by prioritising critical visual cues.   

• Education: students’ comprehension and retention could improve if learning materials were redesigned to capture and retain attention effectively.  

• Transport safety: aligning vehicle control panels with natural visual attention patterns would reduce human error. 

• Public policy and urban design: safety signs, public information campaigns and urban layouts could all be enhanced for positive societal outcomes. 

• Gaming and entertainment: designers could refine virtual and augmented reality experiences through a better understanding of what users will notice and interact with. 

Key takeaways

• Bridging science and impact: The Queen Mary research behind Dragonfly AI shows how fundamental science can translate into tangible economic, global impacts.   

• Transforming the creative economy: The creative sector is a major contributor to GDP worldwide, but it faces industry challenges including inefficiency, subjectivity in decision-making and slow testing cycles. Dragonfly AI demonstrates how integrating AI into creative sectors can effectively address these issues. 

• Beyond marketing: while the tool has established its value in marketing, it has broad potential and offers solutions to a host of social challenges.