How to do personalisation at scale

Personalisation has become a common goal for marketers specialising in ecommerce. Tailoring services to customers’ preferences and behaviours often enables small businesses to build a strong following and encourage repeat purchases. Yet as companies expand, invest more in marketing or develop new product lines, how can they ensure that their ecommerce site remains attuned to the needs of its customer base? Maintaining the personal touch is a tricky task for fast-growing companies.

The Chapar, a personalised online styling service for men, is tackling this challenge by investing in its back-end staff. The three-year-old company has tapped into a growing ecommerce trend to offer hand-picked products based on the data that customers supply.

Men create a profile on The Chapar’s website and then receive a call from a stylist to find out about their fashion tastes. Based on this information, the business selects and sends a trunk of premium clothes to the customer, who decides what to keep and what to send back. The Chapar collects the unwanted items and the customer only pays for those they have chosen to keep.

READ MORE: How brands can use on-site data to avoid abandoned baskets

The company, which was set up by former Levi’s executive Joe Middleton and his son Sam, sells 60 brands to 10,000 customers and employs 25 personal stylists. In a sign of its growth, The Chapar last month launched its first major advertising campaign via a 60-second spot in branches of Picturehouse Cinemas.

Chief executive Sam Middleton says one of his core aims is to maintain the personalised nature of the service as it grows. He explains that one-third of new customers arrive at the site via word of mouth – a statistic he calls “a great validation” for the brand’s personal approach. The Chapar works to a model of employing one stylist for every 500 customers and plans to scale up accordingly as more members join.

“If we scaled without growing the number of stylists, it would dilute the offering because it would become less personal,” says Middleton. “They would be too rushed and wouldn’t be able to get to know their customers [or understand] what works and what doesn’t. It’s very important that we maintain the customer proposition that we started off with.”

The Chapar has invested in the Saleforce customer relationship management platform to enable staff to manage its growing repository of customer data. However, Middleton claims that the personal contact that the stylists have with customers is just as important to delivering personalisation as any algorithm.

“A lot of services claim to do personalisation, when actually it’s just a machine behind it doing all the work. Often [those services] don’t deliver what they claim,” he argues. “We’re selling premium brands and when we get it right, the average spend is quite strong, which means we can dedicate the time and resources to each customer because the economics add up for us.”

Technology investments

Men’s outfitter Enclothed employs a similar business model, also delivering hand-picked clothes to registered customers. The business appeared on BBC programme Dragons’ Den last year and went on to raise £459,000 from crowdfunding website Crowdcube (see Scaling up personalisation, below).

Co-founder Levi Young says the company saw a huge surge in the demand for its service following its appearance on TV, which presented the challenge of “delivering a truly personalised service to a mass audience”. The company invested in back-end technology to support its growth, including new inventory management software to allow it to make selections for customers at greater speed and scale.

“As a startup, we had always grown quite organically and been able to build with demand and spend ages thinking about each box [of clothes] and each customer,” says Young. “Suddenly, when we got a load [more customers] overnight it became evident that we needed to integrate a lot more of our systems and software in order to help us make personal decisions quickly and leverage that data.”

READ MORE: How data can drive awareness, not just sales

Other small businesses are using personalisation tools to react to user behaviour on their websites. The Guitar Hut, a startup providing guitar repair and cleaning services, has calibrated its website to display different messages depending on the location of the visitor and their frequency of visit. The company used the website builder product offered by 123-reg, a web domain registrar, to set up its ecommerce site earlier this year.

This included incorporating various ‘insites’ – commands that change the display of a website according to different inputs. The Guitar Hut uses a ‘local customer insite’ that informs visitors within the vicinity of its East Midlands-based workshop that they are nearby. The message encourages these visitors to pay the shop a visit and provides them with special discounts and promotions. In addition, the site presents frequently returning visitors with a pop-up message encouraging them to like the company’s Facebook page so that it can keep in touch with them more directly.

READ MORE: How to innovate with incentives

“I have seen a steady flow of traffic to my site since personalising it and once I am able to add more to the site, traffic will increase,” says founder Luke Turner. “I have also seen an uptick in social media engagement since using the website builder as I’ve been able to draw attention to my Facebook page and newsletter sign-up option more easily.”

Testing and refining

Meanwhile, for larger corporations personalisation is an ongoing process of refinement that involves testing website functionalities at different stages of the customer journey. Mickael Paris, head of digital marketing at savings and investment company Standard Life, frequently tests new online personalisation features with target customers before setting them live.

He says: “As part of getting to know our customers and creating the experience they want, we use a wide range of techniques such as customer co-creation sessions, online feedback, fast and flexible consumer testing and more.”

In addition, Standard Life uses Oracle’s online testing platform Maxymiser to test a range of features in real time once they have gone live. This includes A/B testing where conversion and click-through rates are measured according to different webpage layouts and adjustments, as well as personalisation elements such as displays tailored to the users’ demographics, location and browsing behaviour.

By comparing users’ actual behaviour with their reported attitudes prior to going live, Standard Life aims to get a more complete picture of its customers.

“We know that it’s one thing to ask customers what they like but it is another to analyse how they actually use our site,” says Paris. “For instance, just because you tell us your favourite colour is blue, it doesn’t necessarily mean it is the colour you are most likely to click on.”

Standard Life is also using the data gathered from its ecommerce platform to build audience segments automatically and to improve the level of personalisation offered across different devices.

“In some ways, we are at the start of this journey,” explains Paris. “Our platforms are consistent across devices but we plan to do much more around personalisation on mobile.”

He reports that the business is seeing a strong return on investment through its use of personalisation tools. For example, Standard Life recently conducted a conversion rate optimisation campaign that included changing webpage layouts and delivering more personalised messages. According to Paris, this resulted in a 217% uplift in click-through rates and a 22% improvement in people engaging with the brand’s ISA saving product message.

Paris believes that as the business grows, the opportunities to deploy new technologies that drive return on investment will expand too.

“Overall, Standard Life is able to deliver more personalised experiences through online explicit and implicit data, mobile usage, our content strategy and phone interactions [with customers],” he says.