How ICONIC uses data analytics to help customers buy what they really want

To fully engage with customers, online retailers have to bring the human touch of an in-store experience to online shopping.

Many companies rely on data analytics to better understand their customers. Online clothing giant THE ICONIC is a prime example with data analytics intrinsically woven into its marketing and merchandising strategies.

“At THE ICONIC, decisions are not necessarily made from the gut,” says Kshira Saagar, Group Director of Data Science and Analytics. Instead, most of the decision-making involving customers, internal stakeholders, operations and marketing relies solely on data analytics.

Helping customers buy what they really want

Through data analytics, THE ICONIC helps customers make more informed purchasing decisions. “Our aim is to make it easier for our customers to shop for products that they actually want,” Mr Saagar says.

Through its ‘snap to shop’ feature, for example, customers are able to upload a photo of their desired item of clothing or accessories onto the website. An algorithm then compares it against the existing product range and produces a populated feed of similar products across various brands.

“Think of how a customer would operate in a normal real-world environment when they want to purchase something that they have seen someone else wear. The ‘snap to shop’ feature allows just that and translates the real-world requirement into the digital world,” Mr Saagar says.

Since its release, the ‘snap to shop’ feature has been a resounding success as time-poor customers increasingly value efficiency in shopping online.

THE ICONIC has also released an augmented reality feature called ‘Visualise AR’ as part of its app. The new feature was designed to bridge the online and in-store shopping experience and offers customers the opportunity to virtually try on a pair of sneakers.

“You can open up the app, compare and pick the shoes that you want, and you will immediately see how they look on you,” Mr Saagar says.

From a customer point of view, this makes it easier for them to shop and inspires them to discover products that are aligned with their preferences.

How can data influence consumer behaviour?

THE ICONIC records datasets based on customers’ historical preferences and captures two sets of data: what customers buy and what they browse.

This process involves looking for similarities in products and identifying what customers like to buy and whether this will fit in with their price range and brand preferences.

“When we put these aspects together, we can uncover what the customer’s preferences are and this enables us to know which products will suit them best,” Mr Saagar says.

Long-term planning for the right merchandise

For Mr Saagar, choosing what merchandise to sell on the site is still a creative decision.

“Buyers and planners decide what they want to bring on site. But how much of it do we sell? While a human being can do this task really well, we definitely need algorithms for high volume SKUs in order to be more efficient with our resources,” he says.

Mr Saagar’s team builds algorithms based on historical data. Such algorithms can then be used to forecast the demand for any particular product in the next few months. The level of investment in a product is also adjusted accordingly in order to meet the forecasted demand.

Professor Jack Cadeaux, Head of the School of Marketing at UNSW Business School, observes that the algorithm acts as an economically efficient process to assist customers.

This creative yet analytical process depends heavily on specifics that are linked to the organisation.

Furthermore, cross-functional learning, interdependency and a development process across the whole organisation are key characteristics to the broader success of THE ICONIC, says Professor Cadeaux.

Mr Saagar adds that the above factors were supported and enhanced by an in-house algorithm designed to inform the business at large.

Data sharing enables better decision-making across the organisation

“A big milestone that we have achieved in the past couple of years at THE ICONIC was to make data available to everyone,” says Mr Saagar. “All employees can see how different divisions across the company are performing. In fact, more than 800 people in the business can access over 500 dashboards at any point in time.”

With readily available data, all teams across THE ICONIC can make better business decisions.

For example, the buying team can see what the marketing team has planned for the next three months, while the marketing team can also find out what new brands are in the pipeline in order to help build a more informed marketing strategy.

On this note, Professor Cadeaux observes that “people from the marketing, creative and merchandising divisions, are not blocked off in separate silos from the data analytics team”, he says. “They understand each other’s language, expectations and goals and this is why this model works.”

How to measure performance and success?

To assess how successful a strategy is, Mr Saagar says the ideal expectation of an algorithm or any smart piece of code is to improve efficiencies.

“Whenever we apply an algorithm, we also measure how much it has improved efficiency. You can also see the improvement in terms of dollars, when you see an efficiency gain,” he says.

Similarly, when it comes to refining marketing strategy Mr Saagar examines what the media mix might look like and which channels are operating on a real-time basis. This helps identify which channels are working and which campaigns are not.

“Because we don’t have to wait 30 to 90 days to measure performance, we can instantly make a decision to change a low-performing strategy and reassess the performance instantly,” he says.

“We have never been more in the ‘failing-fast approach’. If something doesn’t work, we just keep moving forward. Resilience is built in and you can recover quickly.”

How can THE ICONIC identify what customers want?

Understanding what a customer wants across multiple touchpoints is essential for any business in the retail industry.

“We are notified when a customer opens an email or a text message from us – even if they are not on the site. We can also identify when customers come to the site through different marketing media,” Mr Saagar says.

How can product clusters assist with repeat purchases?

At THE ICONIC, the data team aims to cluster products into groups based on their brand and product category combination.

“You may be a customer who likes to wear Adidas shirts but Nike shoes,” Mr Saagar says. “This kind of combination comes up in product clusters. We would then use a finite set of these clusters and make a decision based on a unique combination of five attributes – the brand, the product category, the subcategory, the fashionable level and the price point level for a particular customer.”

As this algorithm sits at the customer level, it is able to treat each customer as an individual journey.

Continuous improvement is important at THE ICONIC, and Mr Saagar says a future goal of the business is to bring the human touch of an in-store shopping experience to its app.

However, the lingering question for many online shoppers remains: what would the purchase look like on me? While many businesses use virtual mirrors, Mr Saagar says that the most effective marketing tool would be to upload a digital version of a customer online, which would then automatically populate and personalise the whole catalogue as customers browse for their next purchase.

/Public Release. This material comes from the originating organization/author(s)and may be of a point-in-time nature, edited for clarity, style and length. The views and opinions expressed are those of the author(s).View in full here.