In today’s modern setting, our proclivity to notice something (even subconsciously) – especially commercial products ­– is affected by what we now call visual saliency. Put simply, if something is visually salient then it is visually prominent or easy to spot. Our brains have a limit to how much visual information we can process simultaneously. As a result, we have learned to prioritize where we direct our attention based on key, visual cues such as colour, size and shape. We are especially sensitive to these cues when we are actively looking for something, like when we are shopping.

In the research paper, Decisive visual saliency and consumers’ in-store decisions, authors J. Clement, J. Aarstrup and S. C. Forsberg report that although “it has been argued that many consumers shop by routine or by simple rules and justification techniques when they purchase daily commodities …, they make a majority of decisions in the shop, and … are affected by the visual stimuli in store.”  Furthermore: “Our everyday purchase of food products … relies on fast and clear interpretation of the information and cues in the Point-of-Purchase (POP) situation. Parameters such as product packaging … [and] product placement in shelf … affect which products consumers will put in the basket.” The research also suggests that the “relative placement of products in shelves and the use of signage affect the visual saliency … as well as the final choices of consumers”. With all these various cues vying for our visual attention, then, the design of a product’s packaging and its placement in-store are clearly invaluable to its likelihood of being noticed and purchased.

One highly efficient way to improve the visual saliency of your product is to use a predictive attention tool powered by artificial intelligence (AI). Predictive attention heatmaps and metrics help to improve the visual saliency of your product by showing you exactly where a customer is likely to look, and which specific features of your design garner most attention. AI gives you pure data, unbiased by human emotion. Although different AI tools exist, what Dragonfly AI specifically does is help you with data-informed design over data-driven design. The term ‘data-driven’ can sometimes make graphic designers nervous, perhaps implying that the AI is steeling the scope for creative input in the project. However, at Dragonfly AI, we like to use the term ‘data-informed’ as we believe that the optimal situation is when our AI helps to inspire a discussion within a design team, rather than dictate to it. We like to think of our AI as being in a symbiotic, constructive relationship with the designers, helping to co-generate creativity.

To demystify any negative associations that designers might have with AI, here are the three main types of visual saliency detection reports that Dragonfly AI can provide to help you and your design team create the most salient possible content:

  1. Qualitative analysis

We create predictive eye-tracking, attention heatmaps that illustrate clearly where viewer attention is distributed across any design. A heatmap visualizes precisely where attention is going, how much is going there, and what the hierarchy of saliency is between the various different features within that design. For example, is your logo receiving too much attention in comparison to the product image on your box? Physical packaging can be especially unforgiving if the design isn’t right. With a website, you can always tweak your design and simply update a given page on a daily basis. With physical packaging, however, it might already have taken an entire one-year design process and employed multiple focus groups. Once your packaging goes live at the printer or factory, you can’t realistically tweak the design without incurring significant extra costs. Dragonfly AI helps you to save time and money by getting it right the first time.

2. Hotspot analysis

We identify the highest-scoring areas within a design, showing you clearly which are your most attention-grabbing features. We highlight the things that most people look at, and rate them relative to each other in order of attention (not time). We can also measure average saliency across a given scene.  You can then reconvene with your design team to check if the results are what they intended. If they aren’t, we can help you to achieve the optimal saliency hierarchy between all of your various features, ensuring that you deliver the truest, most efficient design possible. From solving a simple quandary like whether your logo should be larger or smaller, to whether the lid of your jar should be red, green, or blue, we can help you make those decisions. In a sense, Dragonfly AI provides you with the guard rails within which your design team can experiment with and optimize their creativity.

3. Regional analysis

This process involves the design team deciding which regions they intend to be most visually salient (with our input if appropriate).  Then Dragonfly AI focuses on those regions and scores them accordingly. We start with pre-determined, intended regions, and then we confirm if your specific design features are indeed in the saliency hierarchy that you require. Is your current design really hitting the benchmarks you aimed for? Dragonfly AI gives you an unbiased, data-informed answer. We help you to tweak and improve your design until it’s certain that viewers will always notice your highest-value content. Dragonfly AI can do more than analyze just a standalone product in a studio; we can also analyze your product ‘out in the wild,’ so to speak, rubbing shoulders with competitors on a supermarket shelf. Is it receiving enough of a share of attention on that shelf? How about in an aisle, an end-of-aisle promotion, a pop-up promotion, as a thumbnail photo on a website, or in an Amazon results page? We can analyze all of these situations –  from physical shelf to digital aisle, across mobile, desktop, email, social, video, and eCommerce – and give you an unbiased answer every time because Dragonfly AI will always tell you – in literally every situation – exactly where people will look.

Interested in finding out more about how our technology can help improve your design?

Similar Articles


How to improve the shopper experience with predictive visual analytics

The use of predictive visual analytics processed by Artificial Intelligence (AI) allows brands and retailers Read More

8 September 2021


All you need to know about the science behind Dragonfly AI

For many businesses, the choice of heatmapping tools available, providing predictive performance insight across their Read More

1 September 2021


3 ways to optimise your digital fundraising

To suggest that 2020 was a challenging year for charities is something of an understatement. Read More

20 August 2021