Marketing & PR

How to leverage Visual Insights for marketing

presentation

AI-powered image recognition is the latest breakthrough in modern digital marketing. The ability to identify logos or their various iterations on user-generated photos — which contain no text whatsoever — means new opportunities for the brand to communicate with their consumers. Everyone understands the fun of this feature, but only a select few know more than a couple of basic use cases.

Tetiana Kutsa

Tetiana Kutsa

Chief Marketing Officer

23 July 2018

Image Description

AI-powered image recognition is the latest breakthrough in modern digital marketing. The ability to identify logos or their various iterations on user-generated photos — which contain no text whatsoever — means new opportunities for the brand to communicate with their consumers. Everyone understands the fun of this feature, but only a select few know more than a couple of basic use cases.

How it works

To begin, let's break down how this works. Users don't always add hashtags or captions to their published images. In order to avoid having to sort through millions of user images manually every day, new software was developed to analyze images on the pixel level and identify the ones relevant to your brand.

Adidas Visual Indights

Adidas' logo visual setting. One of the most common objects to appear alongside this brand is the smartphone.

In order to analyze mentions with the brand's logo, YouScan creates a topic stream that collects all visual mentions, which you can see under the Gallery tab. The logo will be highlighted with a bright orange square.

To identify the context of these photos, you can sort the Gallery by object or scene, or view any other analytics report with criteria of your choice.

Which business objectives are fulfilled by Visual Insights?

Get unique insights about your users

Visual analysis of the context surrounding the identified logo allows you to get useful insights to inform your social media communications strategy.

Based on the information about the objects that appear on the photo (people, pets, buildings), the setting (street, office, bar, outdoors), and the analytics report, you can understand key factors in various use cases of your products by your brand's audience.

You can apply these insights by correcting or resetting your audience targeting. The Visual Insights feature also allows you to catch and better understand the audience's impressions of your brand, and use this new knowledge to reposition your marketing efforts or maybe even plan a brand refresh.

We recently caught an unexpected insight like that and haven't decided what to do with it yet. So we're going to share it with you here.

Before integrating the Visual Insights image recognition function into YouScan, our data science specialists tested out the algorithm images of cats and dogs (the AI learned this way). When it was time to develop the logo recognition feature, they noticed a peculiar connection between brands and animals: For example, "smart" gadget brands, especially Apple, often appear in numerous user photos along with cats, while sports brands (like Adidas) are more "drawn" toward dogs.

Apple CatApple Cat 2

Сat photos with Apple products (identified brand objects are highlighted with a bright orange square)

We even made a ranking system of brands that appear on the same photos as cats. Apple came in first by a large margin.

Cats

A likely reason for this is the cats' affinity to lounge on laptops, and delighted owners love taking photos of them and share those photos on social media. And, alongside an adorable photo of their pet, they're inadvertently advertising Apple products. We made a similar ranking for brands and dogs, where Adidas came out on top.

Dogs

This also makes sense: Owners often walk their dogs while dressed in athletic wear, often Nike or Adidas brands. Since we were conducting this "research" during the 2018 World Cup, social media feeds also featured a lot of photos of dogs in hoodies, shirts and hats.

So there you go: new insights. Sometimes a picture tells a better story than words.

Measure the impacts of your ad campaigns

Your Gallery of analyzed images will quickly show whether the actual use of your product matches the intended use outlined in your ad campaign.

The screenshot below shows a hypothetical situation we built with one of the brands in our system. Imagine that your brand is sponsoring a sporting event or another public event of that scale, or acting as a sponsor to a sports team. Your logo is featured on the uniforms of players, referees, coaches, fans, displayed on stadium banners, and numerous other places. You're aiming to increase brand recognition, since it will be seen not only by fans attending in-person, but also millions of people watching online.

In order to objectively measure the impact of this sponsorship activity, it's important to understand which audience demographics are seeing your logo, in what context they're seeing it, what kind of engagement the posts with your logo are getting online, and what is the overall content and tone of the publications that feature your brand.

Age and Gender

If you run a major consumer brand, then it is imperative for you to understand where, how, and in what context your brand appears in your consumers' real life scenarios in order to conduct effective ad campaigns.

No market research or focus group surveys will give you the kind of information you can get from a social media image analysis.

How often are your products used at home? On the street? Outdoors in nature? Which other objects often appear alongside your product? Do these surroundings change over time? How do your ad campaigns influence these use cases, if at all? There is no limit to the number of insights you can draw from Visual Insights.

Age and Gender

Soccer match featuring the sponsored team

Catch implicit brand mentions

Users don't always explicitly mention your brand in their online reviews — sometimes, they simply post an image of your product. With Visual Insights, you possess the superpower to find these "implicit" mentions of your brand, and analyze them in the full analytical capacity of our monitoring system.

Track negative sentiment or sarcasm toward your brand

As we already mentioned, a picture's worth a thousand words. Social media users can post photos of your brand in a negative context, modify branded images, or publish photos of counterfeit or pseudo-counterfeit products. Finding and mitigating these types of publications is especially important for online brand protection.

So you should be happy if the only non-verbal humour at the expense of your brand is limited to these kinds of images.

Sarcasm

Identifying thought leaders

The logo recognition function allows gives a greater competitive advantage to those who possess it—namely, by removing the need for explicit advertising. Integrating ads into the users' native feeds by sponsoring popular bloggers' content allows you to skip cumbersome practices like branded hashtags or special campaign trackers—AI will do the tracking for you.

The logo recognition function will track all of the blogger's sponsored posts and shares of that post, and all the negative comments on such posts can be removed.

Social media influencers won't have to mention your brand anywhere in the caption, therefore avoiding the explicit association with paid ads; and the brand manager, in turn, will only have to track the photos of a recognized thought leader.

For example, let's look at this Instagram post from Wanda Nara, an Argentine media personality, football agent:

Wanda Nara

Wanda's photo contains an Adidas logo, but the caption doesn't tag or mention the brand. If this mention was organic, Adidas will likely want to know about it as soon as possible. And if Adidas is already collaborating with Wanda, our Visual Insights technology would have helped the brand track how and when the celebrity represents it.

Image analysis gives brands access to the whole picture of their consumers' online conversations. And logo recognition is just the beginning! The ability to identify objects, settings, faces and actions in photos allows brands to analyze images the same way as text. Every day, users post more photos and videos on their social media feeds, and some 3 billion images are uploaded to the internet daily. This creates a huge amount of data for researchers and marketers to tap into, in order to find new tools for effective communication with their consumers.

Stay tuned for more remarkable news from YouScan! We are solidifying our position as a leader in the market for social media monitoring in the CIS, expanding our services to foreign markets – all while continuing to enhance our product for customers.

Subscribe to our Blog