The Art & Science of the Sale: Decoding Brand Perception at Scale

You’re probably familiar with the statement “perception is reality.” It’s used everywhere: in politics, conversations about relationships and psychology, and very often in discussions concerning marketing. Regardless of the setting, the premise is that although how we view the world and everything in it varies from one person to the next, everyone’s personal view is more important to them than any absolute, objective truth.?

?The value of understanding customer perception is tremendous and indisputable. Many brands invest millions of dollars in packaging design and pursue celebrity endorsements to influence how their company and products are perceived. And it’s the reason advertising agencies and R&D groups don’t simply trust their expert judgment but instead supplement it with extensive market research and focus groups.?

?Two factors make perception especially important for our business of selling wine online, as well as for our mission of opening the alcohol industry for business and advancing the science of selling:?

1. The overabundance of choice — Whether you buy wine online or in a store and buy it often or sporadically, selecting a bottle can be intimidating. With over 110,000 new wines hitting the market last year, even the most knowledgeable and discerning drinkers can feel overwhelmed.?
2. The importance of the intangible?— Unlike many other consumables where the choice mostly comes down to the ingredients, flavor, and nutritional value, we also look for a story when buying wine. We want the wine to match the occasion, make us feel a certain way, or convey the right message.?Attributes such as when and where the wine was made or how many points it scored play a role, but so does the impression that you get from the bottle, the label, and the language of its description. This individualized brand perception plays a key role in whether you want to buy a bottle or not, what price you deem a bottle to be worth, and ultimately whether you put it in your cart.?

Introducing Perception AI?

At DRINKS, we know that to put the right wine in front of the right consumer at the right time, we need to capture the distinctive personality of each bottle fully, and that to do that, we need to dig beyond “tangibles” like color, varietal, region, vintage, or alcohol content. That’s why our Label AI and Language AI technology supplements this baseline knowledge by leveraging Computer Vision, Machine Learning, and Natural Language Processing to:

• Analyze product packaging and branding??
• Define and study wine label properties??
• Explore connections and associations between wine descriptions?

Top colors detected in labels on two Wine Insiders Red Blend bottles from California (numbers indicate the relative amount of each color)?

A quick look at the example above is enough to draw conclusions about the two labels. The wine inside each bottle is similar – both being California, old vine red blends – but they “feel” considerably different.?

One may feel more appropriate for a dinner party with close friends, and the other might seem like a better choice for celebrating someone’s promotion. One is likely to be perceived as more fun and vibrant, the other more formal.?

This is brand perception.?

Decoding signals from images, descriptions, and transactions to understand how such perceptions are formed, to predict how a given label may “feel,” and ultimately to use this insight to enhance the customer experience – that’s our patented Perception AI platform.?

It may not surprise you that colors play a significant impact on brand perception:?

• Want to create a friendly, approachable brand? Use botanical color palates – they’re 300% more likely to be described as cheerful?
• Creating a quirky brand? Use vivid colors, which are 90% more often quirky than romantic?
• Elegant and premium wine? Use light-colored labels to contrast with rugged and bold brands that use dark labels 50% more frequently?

However, human visual perception extends far beyond just colors. Shoppers subconsciously evaluate details such as the shape and size of the label, amount of white space, word count, distribution of text, type and frequency of objects, the intricacy of design, and the style of artwork and decoration.?

This leads to some of the following insights when it comes to designing new labels:?

• Signal an elegant wine with small fonts and lots of white space?
• Bold brand? Cover more of the label with text and use fewer words?
• Cheerful, quirky, and bold brands should use large labels; elegant brands should use small labels?

Thanks to computer vision, we can efficiently and automatically decode visual clues like this for any number of new wines. Perception AI’s Machine Learning models can then be swiftly deployed to predict how likely each new wine will be perceived by consumers and within demographic groups and regions.??

Here is what Perception AI says about our two California Red Blends:?

Perception traits of wine labels as predicted by DRINKS Perception AI?

Our human intuition suggests that these two labels “feel different,” and they are quite distinct in the “eyes” of the models as well. The algorithms predict both labels are likely to be seen as moderately rugged but assign a much higher probability of also being bold, quirky, cheerful, and romantic to the wine on the left (Label A) while at the same time predicting that the wine on the right (Label B) is more likely to be perceived as more delicate, traditional, elegant, and premium.?

Capturing these “intuitive” perception signals as part of our ever-expanding universe of wine data enables DRINKS to enhance our understanding of consumer preferences and behaviors and improve customer experiences. Predicting how a label will be perceived is just one of the unique ways our DRINKS Recommend platform can personalize experiences for each shopper, identifying which wines they prefer – every single time.?

With Perception AI, DRINKS has taken a significant step toward further understanding the science of selling. As the saying goes, perception is reality.?