March 15, 2012, 9:58 AM

Personalization boosts conversion by 60% for a U.K. e-retailer

Jeweler Astley Clarke introduced software from PredictiveIntent last fall.

Lead Photo

Luxury British jeweler Astley Clarke says the percentage of visits to its web site that end up in purchases has increased 60% since November when it introduced personalization technology and behavioral search tools from PredictiveIntent.

The retailer says PredictiveIntent’s PersonalMerchant and PersonalSearch tools also have boosted the conversion rate rate by 35% among consumers who search on the site. The software takes note of visitor behavior, including what they view, search for and put in a cart, and use that information to tailor what the site presents to each visitor.

Dan Coleman, I.T. and web development manager at Astley Clarke, says the personalization platform has helped the company bring its in-store personal shopper experience to the web, by showing visitors items they’re likely to find appealing, based on their behavior.

“Although we can never hope to entirely emulate the personal service and expertise of our personal shoppers through algorithms alone, the ability to learn customer preferences and use this to guide them to their ideal product gets us significantly closer to our goal of replicating the personal shopping experience online,” he says.

The jeweler, which recently launched a U.S. web site at, retails its own collections online, as well as offering a custom design service and showcasing designs from other high-end jewelers.

James Doman, marketing manager for PredictiveIntent, a company based in the U.K. with an office in New York, says the firm’s software scrutinizes behavioural data to hone in on a consumer’s interests.

“We watch a visitor as they use a web site, and because we understand the products they are looking at from the retailer’s data we can guess affinities,” he says. “So if you look at 60% pearl, 20% diamond and 20% emerald products, then we know you like pearl, diamond and emerald.”

If the client is a shoe retailer, the system notes the shoe size when the consumer adds an item to her shopping cart. After that, Doman says, “We will never recommend a product that isn’t in stock in their size.”

“For a clothing retailer, within three clicks we can tell whether a user is shopping for a male or a female. If they search for a generic term, like ‘jeans’, we will surface men’s jeans above females.  So our software knows that this visitor is male, their favourite brand is D&G and that they prefer black.”

Doman says PredictiveIntent charges retailers a monthly subscription fee based on the number of queries for suggestions made into its system. “Our pricing starts from just $500 per month for up to 200,000 queries, which we estimate would be suitable for retailers with up to 400,000 page views per month.”

He says the software not only drives sales, but saves retailers time. For example, Astley Clarke previously used its e-commerce platform’s built-in system for recommending products based on what the visitor was viewing.

“So two people were going through the products weekly, changing rules and overriding the automatic feature to ensure relevance,” Doman says. “Now all product recommendations are dynamically driven by PersonalMerchant.”

Coleman says Astley Clarke sees a lot of untapped potential in the technology.

“We’re looking to plug PersonalMerchant deeper into our web site to dynamically merchandise categories and to influence promotional content displayed on site,” he says. “We are also working with PredictiveIntent to identify opportunities that will allow us to provide a more relevant experience across multiple touch points including e-mail and third-party web sites.”

comments powered by Disqus




From The IR Blog


Anna Johansson / E-Commerce

Why is social proof big for niche brands?

A small online retailer that lacks brand recognition can get a big boost from high ...


Donn Davis / E-Commerce

Technology takeover: The fashion industry is next

We are now entering the third decade of the Amazon effect, and it is just ...

Research Guides