Technology
ChoiceStream has been providing high-impact, revenue generating personalized recommendations for major brands for more than a decade.
CONNECT’s recommendation engine relies on consumer behavior – computing all types of consumer activity to build rich, personalized consumer profiles that assure your consumers receive recommendations that are meaningful and elicit action.
We use a lightweight API to monitor consumer activity and our testing and optimization tools are world-class. Further, CONNECT offers all the tools you need to manage recommendations on an ad hoc basis while providing the daily, weekly, quarterly and annual reports that you require to monitor the solution’s performance and its impact on your bottom-line.
ChoiceStream CONNECT’s technology:
- Seamlessly integrates with industry standard tools to deliver behavior-based personalization
- Provides comprehensive catalog/inventory analysis to inform intelligent recommendations
- Brings simple solutions to your complex problems
- Delivers real-time optimization of customer behavioral trends
Easy Integration – Built On SaaS
CONNECT integrates within any consumer touchpoint quickly and easily with minimal effort. CONNECT provides a flexible recommendation solution to meet even the most demanding and complex requirements.
Whether you’re a billion-dollar retailer, provide video-on-demand content to millions of consumers via a streaming player, or are an online pure-play, ChoiceStream offers world-class recommendations via a software-as-a-service (SaaS) solution that takes the headaches and additional costs of maintenance off your internal resources. We support increased performance for the life of the partnership with regular upgrades and enhancements and we back up our technology with engineering, implementation and analytics personnel who are an extension of your team.
Both client-to-server and server-to-server integrations are supported by CONNECT. With standard integration, getting up and running with ChoiceStream CONNECT consists of two key integration points:
First, real-time catalog updates allow for near real-time inventory adjustments based on consumer behavior or trending sales. With nearly limitless options for managing your control over your inventory, you can be assured that the right products are recommended at the right time. Catalog updates follow a standard data exchange format using standard schema definitions and data integrity verifications to ensure a seamless and stable inventory update process.
Second, our JavaScript API is simple to implement, easy to understand and flexible enough for the most complex recommendation paradigms. The API is used to share non-PII (Personally Identifiable Information) about consumer activity on your site and to display recommendations in real-time. Because the recommendations are delivered via a high-speed, light-weight RESTful interface, you can be assured you have total control over how recommendations appear. In cases where sensitive customer profile data must be shared in order to maximize effectiveness, our secure and compliant platform requires that data to be hashed or encrypted to protect you and your customers.
ALGORITHM
After more than a decade of applying automated personalized recommendations to a wide range of use cases, business models, content and product types, perhaps the most important lesson we’ve learned is that there is no single model or technique that meets the demands of real-world personalization.
We’ve also learned that context is vital – as much as is past purchase and other historical behavior in determining which recommenders are best to deploy in any given case. Especially in the case of new consumers where there is little to no history, context is crucial.
Using a blended algorithmic approach, the ChoiceStream CONNECT recommendation algorithm uses rules-based, behavioral-based and collaborative filtering in concert to create a nuanced recommendation engine that satisfies the needs of each of our clients. The key data sources we utilize to power our recommendation engine include:
- Past purchase, views, search and download history – by touchpoint
- Current in-session behavior
- Ratings, reviews, social media inputs, and other explicit and implicit data
- Marketing data including demographic, margin and segment information
This blending of techniques drives our ability to accurately predict what the consumer may be in market for next. The result is an increase in lift in more situations for more businesses seeking to generate higher revenues and increased customer engagement through multiple touch-points.
API
Both client-to-server and server-to-server integrations are supported by CONNECT. CONNECT’s Application Programming Interface (API) is lightweight and easy-to-use and integrate into your website, mobile environment or wherever the consumers you’re trying to reach connect to your content and product catalog.
The API was built specifically to improve the efficiency, performance and power of personalized recommendations for CONNECT’s clients. The result is easy integration and a more impactful user experience that gives you total control over what is recommended and how it is presented to the visitors to your site.
The CONNECT JavaScript API supports all popular browsers, including the most recent versions of Internet Explorer, Chrome, Safari, Firefox, Android and iOS which are increasingly accessed by consumers via mobile devices including smart phones and tablets. The JavaScript API is also extensible with all major third-party JavaScript libraries and requires no custom code for recommendations to be displayed on the client-side. Using the JavaScript API ChoiceStream can rapidly deploy recommendations in even the most complex customer environment.
REAL OPTIMIZATION
ChoiceStream’s patent-pending RealOptimization™ (RO) is multivariate testing on steroids. This technology delivers testing and optimization techniques that eliminate standard time-consuming A/B testing and replace it with automated, scalable testing.
Using RealOptimization, CONNECT’s analysts simultaneously compare several recommendation strategies to see which optimizes clicks, purchases or actions in a specific retail or media consumption setting. By understanding how consumers are reacting to various recommendations, the CONNECT algorithm can be adjusted to assure higher efficiency and success…and Real Optimization allows this to be accomplished very quickly.
The knowledge and insight gained from RealOptimization allows our analysts, account managers and program managers to examine how recommendations are driving conversions – or any metric you are targeting.
