NETTING IT OUT

Since 2009, PredictiveIntent has been serving 53 small-medium retail clients operating 82 sites, many of which are multi-country.

PredictiveIntent delivers recommendations capabilities to retailers in the English-speaking world and to their customers in any language. PredictiveIntent’s IntentPredictionServer is designed to seamlessly select and deliver content to any channel, to manage customer profiles across channels and visits, and to personalize the customer experience by adapting content, navigation, and menus. PersonalMerchant and PersonalSearch are the solutions built on IntentPredictionServer.

PredictiveIntent is able to use Facebook data to select more relevant content. Users explicitly allow this use by logging in; users who don’t like behavior-based personalization can opt-out and have their behavior information deleted. This makes PredictiveIntent a strong partner for retailers concerned about complying with the new “cookie laws.”

PredictiveIntent’s PersonalMerchant should be on the short list of omni-channel merchants—especially those anxious to experiment with and leverage Facebook data and functions.

OVERVIEW OF PREDICTIVEINTENT

PredictiveIntent

PredictiveIntent was founded in 2009 by Neil Hamilton and Stuart Swift, who have considerable experience in the development and sale of mobile internet technology. Neil has held executive positions at Qualcomm, weComm, Celtick, and elata, firms involved in delivering mobile services. Stuart also worked at Qualcomm as well as at financial services and high technology firms. PredictiveIntent’s IntentPredictionServer began as a mobile WAP portal personalization technology for use by mobile carriers, system integrators, and third-party portal providers. As smartphones and the “app” model became more popular, PredictiveIntent specialized the IntentPredictionServer for the ecommerce market and branded it PersonalMerchant. PredictiveIntent has 53 clients operating 82 sites. Current headcount is 11 employees, and the company is recruiting support staff as well as J2EE, Linux, MySQL, and web software engineers.

Eighty percent of PredictiveIntent’s clients are retailers, spread across all categories. The remaining clients are in online services, travel, B2B, and media. Eighty percent of clients are based in the UK, 15 percent in North America, and the remainder in Australia and New Zealand. The PredictiveIntent console is deployed solely in English.

PredictiveIntent released the beta of IntentPredictionServer in 1Q2009, and Release 1 was launched in 4Q2009. Subsequent releases delivered profiling, PersonalMerchant, and clickstream support in 2010 and configuration, reporting, and search support in 2011.

Pricing ranges from $500 to $6000 per month with a standard 12 month contract term; pricing is based on blocks. A block is 100,000 recommendation requests.

PredictiveIntent Offerings

PredictiveIntent has these three major offerings:

• IntentPredictionServer (IPS). The IntentPredictionServer is PredictiveIntent’s core framework. It allows a variety of inputs (such as visitor data, visitor behavior, product data) to be processed using a variety of algorithms, filters, and pre-conditions; output is returned in any format. Using IPS, a third party can insert, customize, or suggest relevant content to individual users according to their explicit and implicit data. It is primarily available as SaaS middleware. A comprehensive REST API web service is provided.
• PersonalMerchant. PersonalMerchant is an ecommerce-specific offering. It includes pre-integrations for popular ecommerce platforms, including AspDotNetStorefront, Visualsoft, Magento, and Interspire. These integrations connect a retailer’s online store to the IntentPredictionServer (IPS). PersonalMerchant is built on the IntentPredictionServer and allows online retailers to apply IPS’s 160+ algorithms, unlimited filters, real-time awareness, and minimum latency architecture.
• PersonalSearch. PersonalSearch allows retailers to implement behavior-driven, personalized search. Results can be skewed (ranked and filtered) by the individual visitor’s behavior, using their real-time or profiled preferences.

All clients use IPS and PersonalMerchant; 15 percent also use PersonalSearch. All clients use web recommendations; one fifth also use APIs, and 15 percent use email recommendations.

EXAMPLE: ASTLEY CLARKE

Astley Clarke is a luxury online jeweler featuring imaginative designers that attracts a celebrity clientele. Astley Clarke believes that jewelry purchases are intensely personal, and that making the right recommendation is both critical and difficult. The London retail store employs Personal Shoppers, who guide customers through the designer collections to reach a purchase decision. For Astleyclarke.com to be successful, it needed to deliver a similar level of service. Search accuracy and relevance, and recommendation relevance, had to be top shelf.

Astley Clarke uses PredictiveIntent’s PersonalMerchant to achieve its goals for personalized search and recommendations. PersonalMerchant selects and ranks search results and recommendations based on customer profile and behavior. Astley Clarke uses rules to control which of those selections are displayed to the customer. For example, customers shopping by collection or designer are not interested in products from other designers. By setting rules, Astley Clarke can limit product suggestions to the correct designers.

In our example, a visitor arrives from a Google search for “gold jewelry” and then searches for “rings.” PredictiveIntent technology prioritizes gold rings, applying the visitor’s behavior to estimate context and intent. See Illustration 1.

Finding Astley Clarke Jewelry

(Click on images to enlarge.)

© 2012 Patricia Seybold Group Inc. and Astley Clarke Limited

Astley Clarke uses PredictiveIntent’s PersonalMerchant to achieve its goals for personalized search and recommendations....