This Startup Uses Machine Learning To Create Interactive Content With 90% Engagement Rates

Whichit developed an innovative interactive advertisement unit that’s embedded into web content pages. The advertisement unit presents a poll or survey with four preset answers in the form of text superimposed onto an image, in a four quadrant format. Based on the gamified nature of the advertisement, user engagement clocks in at 90% to automatically generate a gateway to a conversion opportunity. How is this done? Their machine learning algorithm can profile every viewer, recording and processing user details to understand their online DNA and use that information to provide the most accurate commercial incentives, in real time, with the highest chances of conversions.

Originally started as a project in 2012 to source the opinion of the one’s network in response to questions starting with “which”, the technology proved promising to engage internet surfers on a viral level. A year and half later, Whichit was officially launched.

[gallery link="none" columns="1" size="full" ids="2313"]

Today, Whichit allows users to create unique rich media ad units in the form of polls, quizzes, trivia, surveys and more, called ‘Whichits’. Consumers can vote on their preferences via these interactive ad units and those preferences are collected and analysed to allow for real-time profiling of consumers. “Our customer campaigns have reached as high as an almost unheard of 90% Engagement Rate. Click-Through Rates too are 40x higher than traditional display advertising” explains Jonathan Gan, CEO and co-founder.

[gallery link="none" columns="1" size="full" ids="2314"]

Started by Jonathan Gan, Galit Gan and Yarden Jacobson, Whichit is the product of Israeli technical minds. Jonathan (CEO) is a former Israeli Air Force Major in the anti-aircraft company; Yarden (CTO) is a former technical consultant for SAP and the lead developer for Ness Technologies; and Galit, an experienced designer in UI/UX and 3D Imaging.

In recent news, Microsoft’s new collaborative 365^x Business Accelerator has chosen Whichit to join their inaugural class, located in Kfar Saba, Israel. Whichit has been in the spotlight before, winning ‘startup fo the year’ from Facebook in 2015, and soon thereafter drawing in a $250K grant after winning the UK R&D Funding Award to support the development of their innovative machine learning algorithms used in the user profiling process. The final algorithm employs several AI techniques, such as neural networks, clustering, hidden markov models and naive bayes classifiers.

[caption id="attachment_2315" align="alignright" width="525"] Whichit's advertising unit for the Birchbox campaign[/caption]

A Whichit advertisement posits a question with 4 multiple choice answers, typically represented by images with textual overlay. The questions often take the form of polls, trivia questions, or quizzes and the goal is to induce an answer - user engagement. After nearly two years in the market, Whichit has acquired a massive trove of user data. They captured more than 1.26 billion interactions, amounting to over 450 gigabytes of behavioural, analytical and classification data.

With Whichit’s machine learning technology, businesses can answer the questions: what are their users searching for; how the user plans to make a purchase; and what content type results in the highest conversion.

After one completed form, Whichit presents several more, strategically chosen advertisements that attempt to identify a new profile detail of the user. With all advertisements completed, and all answers submitted, Whichit then has enough data to generate their user profile.

The Whichit technology is based on a two-pronged approach using semi-supervised machine learning, natural language processing, and computer vision. Using natural language processing and computer vision, Whichit’s text and image content are classified and extracted from the webpage. Once the Whichit advertisement is posted, a machine learning algorithm clusters, categorizes and predicts user preferences, which are used to derive user activity and engagement trends. Ultimately, they create a digital DNA of user preferences to further enhance the user’s online experience, adjusting their interactive content (to each individual color pixel) and providing laser-aligned commercial incentives.

[caption id="attachment_2311" align="alignnone" width="1000"]
Whichit's audience engagement and user profiling technology, credit: Whichit[/caption]

Whichit is generating revenues (growing monthly), raised a total $1 million in funding, and is soon gearing up for their next A round. They’ve worked with major global brands, such as MailOnline, Birchbox and Mckinsey&Company, and have recently unveiled their Whichit for Enterprise offering. As an innovator in the machine learning based content marketing landscape, Whichit is making big moves to shake up the industry.