7 Ideas For Image Analytics In Marketing

7 Ideas For Image Analytics In Marketing

Great visual content doesn’t come cheap these days, does it? So why not make sure it works to the fullest by adding a little data sauce? Visual Content Analytics tools like DeepBrand can do amazing things finding some less-obvious facts about your brand-related content that could not otherwise be discovered just by measuring standard marketing KPIs. From simple monitoring to content optimisation, there are many ways businesses can benefit from applying computer vision to their marketing strategies. Here are the 7 ideas on how visual content analytics can be the rocket fuel to your digital marketing strategy. Let’s dive in.

1. Monitoring Brand Presence

Some people use hashtags. Influencers tend to use them to mark the brand that sponsored their Instagram post. Brands use hashtags to easily track posts related to specific subjects, e.g. contests or events. But what about those of us who don’t use hashtags? Or those who don’t even mention a brand in text? How can you find a picture about your brand if the brand is not mentioned anywhere in the post description? You can’t. Or, to be precise, you couldn’t – until now.

DeepBrand makes it easy to discover brand-related images that were not tagged with a brand name. It is possible thanks to our AI which can recognize a brand logo in the picture. This way you can monitor the actual brand exposure outside the channels under your control. Also, you can find your organic ambassadors who are sharing the love for your brand.

2. Competitive Reviews

If there is any ground truth about working in brand marketing, it’s the overwhelming urge to compare yourself with the competition. If this is your case, DeepBrand has a bunch of features to ease your pain. You can define and analyse competing brands in DeepBrand’s My Category section. Right there, not only can you keep track of you rival brands’ exposure, but also you are able to analyse the Share of Voice for each of the defined brands over time.

Another interesting feature in DeepBrand is the relative context analysis. DeepBrand can compare the top contexts of each brand defined in the category with its rivals. For us, the contexts are derived from other generic objects found in every image along with a brand logo. On the charts, you will clearly see which contexts are highly specific for a given brand. If other brands did not post many images that look similar, there is no major context overlap. This type of analysis lets you find your brand’s distinct characteristics or identify areas that are common for the whole category. Learning that may help you better differenciate from your competitors – and we all know how crucial that is!

3. Influencer Profiling

Meet Joey who wanted to find influencers to help him promote his new organic coffee. Joey found four of them who were willing to create some content for him, but couldn’t tell if what they promised will be of the right quality. Also, he wanted to know if they have ever posted anything about coffee. Joey didn’t have time to check all the social profiles by himself, so he decided to use DeepBrand.

With DeepBrand, Joey found that two of his prospective influencers have already cooperated with major coffee brands. Those companies were his competition – so he ruled them out. Another person didn’t have any coffee-related content, but his Instagram profile was already saturated with lots of other brands. This made way too much clutter and practically compromised his credibility as an opinion leader. The last Instagram profile turned out to be a hit. The author has already published photos associated with coffee – it was one of the most popular motifs on her profile. In addition, this influencer did not cooperate with any coffee or cafe brand yet, and her photos were not overloaded with sponsored content. Joey decided to cooperate with the last person on his list. After that, he also used DeepBrand to generate a post-campaign report that excactly measured his brand exposure.

To wrap up, influencers can make a great impact when you wish to promote your brand in an authentic and convincing way. But it doesn’t mean any exposure is always good exposure. With DeepBrand, you will be able to quickly evaluate social profiles and run reports on your influencer marketing activities.

4. Creative Research

How do you decide what type of content works best for your audience? To solve this, agencies usually have a variety of tactics to choose from in the ad toolbox. There is attitude or interest mapping tools like Kantar’s Target Group Index, personas, trendscouting or the good ol’ fashioned desk research (asking people around at the coffee machine doesn’t count). As the communication strategy takes shape, it is nice to have some data to back up your line of thought. You’ll want to use it when showing your cards to the client.

With DeepBrand, you will be able to recognize and classify visual content that is focused on any topic. Our tool also helps you understand the relation between the content posted and effect it generated. DeepBrand provides insight for the creative and social media teams to create visuals that work. And instead of spending long hours doing research, you will have it done in no time (which means the world when you’re under deadline).

5. Content Audit

Let’s say you are an agency and your client asked you to prepare a content audit. He needs whatever had been posted on the brand’s official online channels over the course of last year. Remark: you already know what the client actually meant was to find an argument to rid you from managing their Facebook profile. He’s once stumbled upon several of those pitiful memes that were expected to boost engagement but everyone feels it took a toll on company’s image. So ideally, you’d need to go through everything and prove to the client that the impact of kittens is a bit exaggerated and that the fan page is still dominated by the safe brand-related stuff.

DeepBrand can help you with content analysis on your brand channel. All you have to do is to see the leading topics revolving around your brand. You can do it by mapping the images posted on the channel. You could also search for specific objects in the images and see if they make a noticeable piece of the pie. Lastly, you can also determine the social impact of the chosen topics and find out if they were among the most engaging ones.

6. Finding Generic Objects

Marketing-related image analysis is not always about the brand. Sometimes you may want to quickly identify posts or articles that show images with a specific object – which, in turn, sets the right context for your brand or product to shine.

For instance, if you sell bicycle accessories, you may want to find bicycle aficionados, to personally engage them in a direct conversation. If their love for bikes is pure and genuine, they’re guaranteed to post photos that have a bicycle. The other case might be if you intended to quickly identify channels about bikes that have the most engaged audience. So how do you find your potential customers by content search?

DeepBrand lets you search visual content for a specfic object – like a bicycle – and find all the image posts or articles with it. Once you get all the picture results, you can filter out those that are outdated or have the least engagement. This way, you will be able to identify key channels for your direct marketing activities.

Let me make one important notion here. No website or a social media channel have a homogenous target audience. The articles are massively discovered via search engines and Facebook posts’ visibility is heavily dependent on personalization algorithms. Just like with long-tail SEO, it’s better to follow very niche, specific content because it’s usually just more relevant to the user. And that’s where DeepBrand’s content search comes in handy.

7. SEO Optimisation

Until recently, optimising content for SEO used to mean having cleared the right keyword saturation, fixing the <h1>-<h6> tags and making sure if every blog post contained alt-texted images, to help Google and improve user experience. But now, since the search giant’s doubling down efforts to extend the importance of visual search, you might want to really consider optimising website content to get more traffic. Running an on-site image analysis can help you cross-check relevance with the text content and create a proper, descriptive list of image file names and alt attributes that will all get you an improved keyword optimisation.


So, there you have it. I hope you enjoyed reading the article. If you feel like sharing your ideas on how to leverage image and video analytics with a broader audience – send me a hint at dominik@deepbrand.ai so I can feature you here on DeepBrand blog!

Dominik Komar

Dominik is the CEO of DeepBrand. Passionate about technology. Always curious about what's next.