Retail Insider Awards – Meet the ‘Most Intelligent Use of Data’ shortlist

Retail Insider recently announced the shortlist for its ‘Transforming Retail Awards’ and over the next few weeks we will be focusing on each of the six categories in the awards.  For each we will be providing a profile of each of the shortlisted entries.

Most Intelligent Use of Data (sponsored by Clipper Group)

1. Shop Direct ‘Very Assistant’

Shop Direct became the first UK-based retailer to use conversational user interface (CUI) technology for customer service by launching a fully-automated platform to allow customers to quickly find answers for their questions in a WhatsApp-style chat environment.

The ‘Very Assistant’ is available through the Very app and makes the user journey even simpler for customer service questions by allowing shoppers to interact with speedily and easily in a format that they are adept at using day-to-day. The platform is automated, which means customer queries are answered quickly and accurately.

Very Assistant works by asking the app user if they need any help before they are presented with a sequence of questions and multiple action options, which the customer taps within the chat environment. The customer’s responses enable the platform to instantly serve up the information they are looking for.

Customers can use Very Assistant to track an order, make a payment on their account, confirm that recent payments have been processed, check their payment dates and request a reminder of their account number.

The new technology, which was developed in-house, has been introduced in response to customer research, which showed that people wanted to interact with the group’s brands in a chat environment.

The research also analysed the most common queries customers ask, those questions that are best suited to the Very Assistant platform, and how the new technology fits into the company’s overall customer care journey.

The technology is the first step towards a ‘natural language’ solution. Shop Direct is working with IBM Watson to develop and introduce during 2017 an AI-fuelled CUI platform. This will allow customers to ask questions in their own words within a chat environment, with the AI technology serving up the answers they are looking for – making their experience even more personalised.

Alex Baldock, group CEO at Shop Direct, said: “Our customers want to chat to us as they do their friends on WhatsApp – it’s what they’re used to. Very Assistant is our response to that desire and it will make it even easier for our customers to shop. But this is only the start. AI will change the game and we’re backing it in a big way. It’ll allow us to offer a personalised, ‘natural language’ CUI experience for service queries in 2017 – which will be massive for our customers.

Jonathan Wall, eCommerce director at Shop Direct, stated: “This fully native platform is squarely focused on what our customers need. It’s delivered through our app because that’s where they want to have questions answered. It’s also the best place for us to collect feedback and constantly improve Very Assistant.

“We think this new technology will simplify our user journey, improve satisfaction, and help to boost efficiency in our customer service operation. It’s also the first step towards ‘natural language’, AI-driven CUI – which is something we’re hugely excited about.”


2. Geoblink

Geoblink is a two-year old Spanish SaaS-based geospatial business intelligence product whose mission is to become the global solution for retail companies regarding the analysis of their physical network of stores, so that they can take optimal strategic decisions in relation to their network of points of sale, Network design (understand why some stores perform better than others) and expansion strategy (where to open the next point of sale).

According to PwC’s Global Data & Analytics Survey 2014, 80% of retail executives base their most business critical decisions on intuition and previous experience alone. This behaviour leads to incorrectly made business decisions that strongly and negatively impact ROI.

This problem is especially important in decisions about their network of stores. Retailers need to figure out:

1) Where to open new stores

2) How external factors influence their network performance (weather, pedestrian traffic, wealth, competition, cannibalization, etc)

3) Where to do their marketing campaigns.

Most of the Retailers are not sophisticated, use intuition or pay someone else a lot of money to figure it out for them. Geoblink integrates multiple sources of data, combining it with internal data, along with advanced analytics and an easy-to-use beautifully designed tool, with the goal of providing them with means to improve their decisions.

Geoblink is based on three pillars:

1) Ready to use rich datasets from multiple private and public sources, both small and big data (High variety of traditional data (sociodemographics, socioeconomics, …) and non-traditional (competition, social media, traffic, properties for rent, tourism, climatology, …). We also build our own data indicators (eg, we do advanced inferences to make data as granular as street level).

2) Geoanalytics and advanced analytics techniques, (including machine learning techniques): build influence areas, search an area with specific characteristics, compare two different areas, search twin areas, discover potential cannibalization effects, sales prediction before opening a store, etc).

3) A powerful visualisation layer including business intelligence and map capabilities: Key information is beautifully presented and quickly accessible. Easy to use. Interactive and dynamic. Accessible anytime and from any device.

The business model is based on an annual license fee that depends on client’s characteristics, such as number of stores, number of users or number of countries. It therefore combines the robustness of a subscription model with the potential of scaling the complexity of our customers.

In short, Geoblink, with more than 50 paying clients, helps retailers to make inform decisions based on objective data, saving ultimately time and money.


3. Blue Yonder / Otto

Software from Blue Yonder is helping a variety of retailers including Morrison’s and German-based Otto significantly improve their inventory management by revolutionising their forecasting capabilities.

The capabilities of Blue Yonder is built on research undertaken at the CERN laboratory in Geneva and which enables it to analyse billions of transactions and 200 different variables in the retail environment including past sales, web searches, and other criteria such as weather.

The result of this enables it to predict for its retail clients what their customers will buy in the near future. It has proved extremely effective – with an incredible 90% accuracy in predicting what will be sold within a 30-day time-frame.

This enables the automated replenishment of stock that has driven significant reduction of out-of-stocks and end of being frequently limited to offering customers long delivery times. Whereas previously customers often had to wait seven days for delivery this has been reduced to only two days.

Although Otto only initially used Blue Yonder as a forecasting tool its confidence in its outputs resulted in it allowing the automated replenishment of goods. It now automatically purchases 200,000 items a month.

Dealing with the vast data sets on which Blue Yonder makes decisions in real-time for its clients would not be possible by humans. It is only possible through artificial intelligence powered machine learning.

The power of the Blue Yonder analytics has enabled Otto to reduce its surplus stock by 20%. It is now frequently the case that goods automatically replenished do not actually go into storage but are sent directly to the customer.

Morrison’s recently started using the Blue Yonder solution also to optimise replenishment and for the automated ordering of some items. It is enabling the company use intelligent analysis on the 13 million ordering decisions it makes every single day. This capability has been helping the supermarket chain reduce the levels of out-of-stocks on its shelves – by 30% at this stage.