Are marketers getting data from your selfies?

Smile! Click. And upload.

Most of the time, when posting photos on your Instagram or Facebook accounts, the biggest concern on your mind is which filter to use. But what might just be a simple click of a fleeting memory for you, could mean a mosaic of information for marketers.

Creepy, right?

Well, there are several photo mining start-ups out there today which are actively sieving through consumers’ public selfies or images to help marketers and advertisers have a better understanding of consumer profiles and the data out there.

One such photo mining startup is Ditto Labs which mines publicly available images on Instagram, Twitter and Tumblr. In a conversation with Marketing, David Rose, CEO of Ditto Labs said social media is becoming increasingly visual but brands aren’t able to “listen to photos” – this is where the company steps in. It fills the gaps with proprietary image recognition tools that tags brands, scenes and objects in the daily torrent of photos people share on social media. This is then transcribed into coherent customer insights and engagement strategy for Ditto Labs’ clients.

“There's a famous phrase a picture tells a thousand words. In a single photo there can be a number of unique buying signals. For instance, a marketer can tell where the image came from, what brands are present in the image and the person's interests,” Rose tells Marketing.


Today the world is moving to a point where personal stories are told through social images. Mingle that with the growing smartphone penetration and dropping price range, image-based storytelling will sooner than later, dominate.

“Millions of new people are going to be walking around with a digital camera in their pocket. Couple this with people's preference for image-first social networks like Instagram, and you have a new way for marketers to understand customers' behaviors and preferences,” he added.

Photosharing Visual

(Data courtesy of: Ditto Blogs)

Without naming clients, Rose added that in several instances privacy concerns were raised by potential clients. However, he reiterated that his team only analyses publicly available photos – not very much different from how marketers mine for data from text messages.

“Admittedly, some people do express privacy concerns when they first see the platform. However, this initial reaction is typically due to seeing something novel, and not because there's a sense that people's privacy is invaded. I think social photo analysis is where social text analysis was a few years ago. Nowadays people understand that if they tweet about a brand, that text is searchable. Soon people will adopt the same understanding about the photos they share,” Rose said.

Data mining from selfies an evolution to text mining?

However, disagreeing with Rose is long time digital marketer Damien Cummings, global head of digital marketing at Standard Chartered. He explained the two are very different as consumers post images in a bid to share it with their loved ones or friends.

Cummings said:

You can’t have a personal relationships or friendships with brands. So corporations don’t have a right to hijack selfies for their own marketing interest. I suspect there would be a huge backlash from consumers if found out that their selfies were being used without their permission, even for simple data gathering.

In his view, text based content are usually publicly posted and there’s already an assumption that all of this content (whether it appears on Facebook, Twitter or a blog) is going to be seen publicly and as fair game.

Prantik Mazumdar, managing partner of Happy Marketer however, agrees with Rose that this could possibly be an evolution of text based data mining.

“I think once any piece of content, irrespective of format, which is on the public domain, is available for anyone to see, re-hash, mine, analyse, re-produce unless of course it is copyright protected. So I don't see why data mining for consumer pictures should be any different from data mining for text based posts or data gathered from campaigns,” Mazumdar said.

Whilst the data is on the public domain and open for mining, what would be nice is if the brand chooses to be proactive and open about the data mining initiative, added Mazumdar.

Is Asia ready for data mining from images?

While the service might be taking off on the US, most of the marketers Marketing reached out to in APAC said that at this point of time, they were not data mining when it came to images. Cummings echoed these marketers and said that while there is a massive interest in moving to insight and data-driven marketing, there's still only a small percentage of marketers that have embraced it.

Even when it comes to mining for data based from their own brand campaigns, most marketers are simply content to have customers engage with their brand, or share their experience of that brand with a friend.

“The idea of mining data from selfies is an interesting one but for what end? Unless a brand has a team of analytics to mine insights, or a robust digital marketing operations capability that can take advantage of these insights, then the effort isn’t worth it,” Cummings said.

Of course, there is also the very thorny issue of privacy.

“In the short term, there aren’t that many benefits. The risk of upsetting customers by stealing data from their photos appears to massively outweigh any benefits,” he added.

Don Anderson, regional managing director of social media agency We are Social added that currently there is still very little guidance and regulation in the industry for how researchers should or shouldn’t use social media data.

“For research, our use of such data tends to remain at an aggregate audience / community level to gain insight into consumer behaviour and trends, and not really at a personal level to target individuals,” said Anderson.

On a whole however, he added that this is probably no different than ‘traditional’ social media listening.