With customer behaviour continually changing, how can retailers ensure the quality and relevance of each customer's experience? The UK-based omni-channel fashion retailer J D Williams has made a start by capturing some 65Gb of online customer data using the Celebrus platform every month. Here we examine the whole process that led to the company building stronger-than-ever customer engagement. This article is copyright 2015 The Best Customer Guide.

By feeding all of the online customer data into a Teradata data warehouse holding significant customer history and transaction data, the company is tackling attribution (Marketing spend), enhancing website functionality, improving segmentation, creating behavioural personas and using predictive modelling to deliver a relevant and personal customer experience across every channel.

The ways in which customers interact with retailers have changed fundamentally since J D Williams was founded in 1875. A leading UK omni-channel retailer, operating over 20 successful brands, the company not only generates over 58% of its sales online but also offers customers the chance to carry shopping bags across its websites - from Jacamo to Home Essentials and Simply Be for instance - and complete the checkout on any site.

Increasingly these customers are visiting using multiple devices - with mobile traffic (Smartphone and Tablet) now accounting for 50% of all sessions. Two years ago this figure was 16%. This creates new challenges in the quality of customer experience, but the company's analysis of customer behaviour has revealed that those using multiple devices are significantly more valuable to the business through greater overall sales. As Gareth Powell, Head of Web Analytics, J D Williams, says, "In this changing marketplace, any company not willing to use online customer data will get left behind."

J D Williams has recognised the value of customer data for many years - over the past 20 years it has created a single repository for all trading and customer data in a Teradata data warehouse. The quantity of online customer information now captured is extraordinary - since 2010 the company has captured and retained every single web site click, search, basket add, purchase etc within Celebrus, creating some 65GB of customer data every month. Whilst that's a huge amount of data, the company keeps it all because it gets enormous value from having that level of detail, plus this is an enabler to creating a more complete picture of the customer and their life-stage.

Data-Driven Model
The data-driven technology transformed J D Williams' understanding of the customer journey. With 62 tables of granular data, the retailer can visualise a raft of customer activities at an individual level, from the way the customer is sorting products, to the time spent viewing a single page, the entry method to the site (such as paid search), the products added - and products removed, the use of filters (such as size or price), and the searches used externally - such as 'plus size maxi-dress'.

Additionally, tracking exit pages, pathing customer journeys or tracking abandoned products are all important ways J D Williams can assess the way the customers are responding to the overall experience. As Powell confirmed, "The beauty of Celebrus is that we can allocate a customer account number to around 50% of our traffic, which means we can build up a really good picture of what each individual is doing on the web site in one session - and then stitch that together over multiple sessions to get a highly detailed single customer view."

Individual Customer Insight
Combining offline and online data provides the company with deep customer insight, including contact, payment and order history, as well as exposure to marketing campaigns - building an accurate picture of lifetime value and the creation of a profit score for every individual.

In addition to using this web data within all marketing campaigns - from email to outbound telemarketing - J D Williams is using this depth of web insight to deliver a far more personalised overall experience. The retailer will be using Celebrus data to drive forward with future behavioural emails such as 'browse not bought' and 'abandoned basket'. Mailing selections are informed by product preference or the way a customer sorts on site - indicating price sensitivity (for example).

The retailer is becoming increasingly sophisticated in its use of customer data, for example using predictive modelling to understand the likelihood of a customer making a purchase. J D Williams has over 50 predictive models which, before having web data, were based on transactional and payment insights. Enriching this with six months of Celebrus web data, including browsing information which reveals a greater intent, has significantly improved the accuracy of the models from 75% to 93%.

The retailer has also created a number of behavioural personas - such as value hunters, frequent abandoners and on trend customers - in order to create a more relevant and personal experience as customers arrive on site. Powell explained, "For value hunters - those who consistently visit the sales area or sort by price on the navigation - we have a series of triggers that we will be able to action via website personalisation".

By contrast, frequent abandoners could be identified and offered incentives to encourage them to complete the purchase or increase the number of products in the bag; while on trend customers - those always looking for new products - can be targeted with the latest items and aspiration emails. Powell added, "These personas build on the customer segmentations we have traditionally created using transactional data by adding a real understanding of intent and wants through the behaviour exhibited by individual customers online."

Enhancing the customer experience
The data also plays an essential role within the company's multi-variate testing (MVT) by identifying the downstream customer impact of MVT activity, including standard cosmetic tests, from colours and placement to call to actions buttons. In addition, the data highlights friction points on the web site - enabling the company to rapidly highlight and address issues that could affect the overall customer experience.

"From a highly practical customer service perspective, if we have downtime on site, we can track the people who hit a holding page and follow up with an email," Powell said. The result has been a 7% improvement in sales per customer contacted - the company has recovered the potentially lost demand and added incremental sales.

Indeed, the company is now using this customer experience data to inform on-going web site development. This has been one of the most valuable benefits the organisation has realised from the Celebrus data. For example, in the last year J D Williams has focused heavily on the smartphone experience, including enhancements to the sign-in and checkout process, improving its Account Management functionality and improvements to Search and Navigation - all opportunities highlighted and supported by the Celebrus data. "Traditionally the ecommerce development roadmap was based on what was happening in the retail market and industry best practices. Now we have a third, key (and arguably the most vital) ingredient which is the voice of our own customer" Powell added.

J D Williams plans to further exploit its data to drive customer personalisation and is looking closely at the way consumers go online and encouraging more customers toward a multi-device interaction. Furthermore, the company is executing a strategy around retail stores in the UK as well as further developing its brand footprint internationally. This will boost overall brand awareness. The company is also focusing on creating a seamless cross channel experience that embraces call centre, web and store. "At the heart of this is the ability to perform deep dive analytics to gain a better understanding of customers' needs across every channel," Powell concluded.