How deep tech startup Turing Analytics is helping retailers step into the age of visual commerce

0
81

Turing has developed a visual search and recommendation engine that runs on artificial intelligence and machine learning algorithms, and helps shoppers search for products using photos and videos.

At a glance

Startup: Turing Analytics

Founders: Divyesh Patel, Aditya Patadia

Founded in: 2015

Based out of: Bengaluru

Services: AI-based visual search, visual product recommendations and trend analytics for online and offline retailers

Sector: Deep tech

Funding raised: Bootstrapped

It is the age of analytics, of data-based learning and decision-making, of seamless buying and improved customer experiences. It is also the age of visuals — images, videos, GIFs, motifs, virtual reality, and more. The internet is an increasingly visual medium, and it is imperative that businesses stay on par.

The founders of Turing Analytics — a deep tech startup named after Alan Turing (regarded as the Father of Artificial Intelligence) — realised the potential of visual commerce back in 2015, when the term wasn’t as ubiquitous as it is now. Today, it is a buzzword.

According to a 2017 eMarketer study, 75 percent of online shoppers in the US “mostly or always” search visually prior to making a purchase, and only three percent never do. Nearly 74 percent of consumers claim that text-based keyword searches are inefficient in helping them find the right product online, according to a separate report by Slyce (a visual search firm).

Global market research firm Markets and Markets estimates the image- recognition market will grow to $25.65 billion by 2019, a rise of 216 percent from $9.65 billion in 2014. Hence, visual commerce is here to stay.

Visual search in visual commerce

What really is ‘visual commerce’?

The process of visuals aiding ecommerce — and retail in general — is visual commerce, and it is revolutionising the way customers shop. Amazon, which leads product searches on the internet, has been a pioneer in implementing visual search tools. Closer home, fashion e-tailer Myntra has done the same, shortly after Amazon brought visual search to India. World over, scores of other retailers have done it too.

READ  Customers expect AI-powered engagement — can you deliver? (VB Live)

This is significant because the growth and acceptance of Turing Analytics began only after Amazon.in (and later Myntra) enabled visual search for Indian shoppers. They inculcated a new habit, brought about a new convenience, and eventually established the need for visual search – that allows customers to look for products using photos instead of keying in text in the search bar.

Turing Analytics has developed an image-based product discovery and recommendation engine for ecommerce sites that enables shoppers to upload a photo and look for similar-looking items across categories. The visual search engine is powered by AI and deep learning algorithms that can scan millions of images in a few milliseconds and display the most relevant products to users.

The engine is built on “deep neural networks” – the same structures that give humans their visual recognition abilities. Essentially, Turing’s technology processes swathes of unstructured data (images, videos, even text), and identifies, catalogues and organises them like any human would. But, because these are machines, the process is much faster, smoother, more methodical, and largely accurate (90-92 percent, claims the company).

So, how did all of this come about?

Origins of Turing

Not very long ago, Turing Analytics co-founder Aditya Patadia was a product manager at homegrown e-tailer Shopclues, where a part of his job involved scanning customer reviews, analysing them, and identifying shopping trends.

“Product reviews were unstructured data. We were analysing about 150,000 reviews a month. There was a need to moderate and structure them, and it could be done using machine learning,” Aditya tells YourStory.

Structuring the data would not only give retailers a better understanding of buying behaviour, but also help in cataloguing and identifying products that sold and those that didn’t. But, Aditya wanted to do more with the data generated by customers.

Around this time, machine learning, predictive analytics, and AI were picking up in the West. Fashion and furniture retailers had realised the need to offer cutting-edge shopping experiences to savvy, new-age customers. Aditya teamed up with Divyesh Patel, who was a summer intern at Kimberly-Clark, and the two started Turing Analytics in Bengaluru in October 2015.

READ  Japan’s Hayabusa 2 mission lands on the surface of a distant asteroid

Aditya says, “Machine learning was picking up, and we thought let’s indigenously build an analytics company for retailers.” Interestingly, both the founders’ past employers — Shopclues and Kimberly-Clark — came on board as clients on Day One.

Clients and impact

In about three years since then, Turing Analytics has worked with Tata Trends, Tata Chemicals, and other divisions of the Tata Group. It has recently signed up with Future Group’s C&D Labs, a Rs 100-crore accelerator fund for startups. Turing will offer visual intelligence, similar product reccos, trend analytics, etc., to Future Group’s multiple retail brands.

Aditya says,

“We are one of the seven startups that are a part of Future Group’s Retail 3.0 strategy. We’re looking to improve the customer experience both in-store as well as online by focusing on image and video analytics.”

The startup  claims that its visual intelligence solutions help retailers improve product discovery, boost customer engagement, increase conversion rates, and grow repeat visits. The more customers search using photos or videos, the better the product recommendations get because the AI engine keeps growing smarter.

While Turing doesn’t share any measurable data, an eMarketer global report indicates that visual commerce helps increase product views by 48 percent, and grow retailer revenues by 11 percent. Turing concurs that the impact of its solution is on the retailer’s topline. “We are adding to the cart size,” Aditya says.

How Amazon triggered growth

Turing’s visual engine can currently identify more than 500 different types of products across fashion, lifestyle, home and furniture catalogues. Over 30,000 visual searches are being recorded on a daily basis, and they are rising 30-40 percent. The growth has particularly taken off in the past six or seven months, notes Aditya.

He explains,

“Earlier, there was no awareness about visual search. But, the moment Amazon put a camera icon on its search bar, people started looking for items using photos. When consumers became aware, businesses also became aware. Soon, even Myntra allowed visual search, and now there is a demand for the feature. So, Amazon helped scale the whole thing. For us, onboarding clients became much simpler.”

Expansion and funding

Now, that the initial adoption challenges have been overcome, Turing has its eyes set on international markets, where “almost all retailers” are enabling visual commerce.

“We’ve run pilots in Southeast Asia and Australia,” reveals Aditya, “and we look forward to developing the product for those markets where customer needs may be different from India’s.”

The startup has been bootstrapped so far, sustaining its 10-member team on the family finances of founders and with the help of a “consistent cash flow” from clients. It claims to have been profitable from Day One, but there was no way to ascertain this. “Frugality helped us sustain,” says Aditya.

READ  SpaceX’s Starlink aims to put over a thousand of its communications satellites in super-low orbit

However, the startup is now looking at funding. Aditya reveals, “We’ve had conversations with investors in the last 3-4 months. They have loved the product, and for our international expansion, we might need the funds.” Nothing is on paper yet, but Turing is hopeful about closing a round this year.

Source: Yourstory

To Read Our Daily News Updates, Please Visit Inventiva Or Subscribe Our Newsletter & Push.

Comments

comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.