User Profile Service Overview

The User Profile Service (UPS) handles all user interaction events. It gives you a full, 360-degree, omnichannel view of your customer with real-time updates on user interaction on any instrumented channel. It also stores historical behavior information which can be supplemented with user attributes, segments, and preferences.

The User Profile Service (UPS) allows retailers to identify a user and store information about that user's behavior as the interact with the retailer's site, then offer personalized recommendations based on that behavior.  The more you know about shoppers' behavior, the better your recommendations can become.

Unlimited amount of storage (no more reliance on cookies) paired with a combined view of consumer behavior across devices means a complete view of the consumer. This results in improved personalization, generating more engagement and overall lift in sales.

Without the need for cookies, Algonomy can integrate our personalization technology with the retailers’ cloud directly (also referred as Server-side integration) without any limitations: an important requirement for many retailers as they want to centralize the delivery of content to consumers in their own clouds.

Instead of relying on cookies, Algonomy has built a real time enterprise data bus that stores consumer behavior as it happens across the globe in our 11 data centers. Every view, click, purchase is stored in our cloud and is available within a second for our personalization engine to make decisions on the next page/screen the consumer travels to in their shopping journey.

What is the User Profile Service? 

Personalization solution providers use web cookies to store what the consumer has viewed/clicked/searched/added to cart in the current web session. This information is then used to personalize product and content recommendations within that shopping session. Cookies have limitations:

  • You can only store a small amount of data in cookies.

  • Legislation across the globe has forced browser manufactures to disable 3rd party cookies altogether. 

  • Cookies only store the consumer’s behavior on a single device.  If a consumer switches between their laptop, table and smart phone, the cookies are not shared.

The net effect is an incomplete view of the consumer’s shopping journey.

Why is it important? 

Unlimited amount of storage (no more reliance on cookies) paired with a combined view of consumer behavior across devices means a more complete view of the consumer. This results in improved personalization, generating more engagement and overall lift in sales.

Without the need for cookies, {rr} can integrate our personalization technology with the retailers’ cloud directly (also referred as Server-side integration) without any limitations: an important requirement for many retailers as they want to centralize the delivery of content to consumers in their own clouds.

How do Retailers use the UPS? 

Retailers don’t have to make any changes to their current Algonomy implementations. Every implementation already includes the two "keys" (SessionID and UserID) used by the UPS to keep track of consumer behavior in our cloud.  When available, the User ID assigned by the retailer (since that’s a stronger signal of the identity of the consumer) will be used. If retailer can’t provide that (consumer is new, hasn’t logged in and retailer doesn’t have soft login implemented), we will use the session ID (a shorter living ID usually attached to a single shopping session) to track the user.  

In addition to storing the behavior attached to a session or user ID, when the following scenarios happen, we will merge the behavior across those sessions in our cloud:

  • two or more session IDs coming from the same browser

  • a user ID accompanying a session ID mid-way through a session.

  • the same user ID from multiple devices

These techniques will further enrich the depth of information we have for each consumer in our cloud.

User Data

User data from UPS is used by Algonomy products to facilitate engagement personalization with the user. The data is also available directly by the customer through the Profiles API.

Data is loaded into UPS in multiple ways:

  • Instrumentation: This is when real time user interaction on your web and/or mobile channels is captured and sent to Algonomy through instrumentation code embedded for that channel.

  • Batch Uploads: This is when you want to update information for multiple users at the same time in an offline fashion. For example, adding user segments based on some offline data clustering.

  • Offline Order Feeds: This mechanism allows in-store / non-instrumented channels to enrich the user data stored in the User Profile Service through a regular feed process.

Instrumentation and user identification

Deciding when to instrument the user ID will have an impact on the quality of personalized recommendations and is an in-house decision which will likely consider your corporate stance on customer privacy, tracking, and the use of cookies. The two extremes are as follows:

Best-case scenario: you set a dedicated, persistent cookie as soon as you can identify a user. The value of that cookie is instrumented as R3_COMMON.setUserId('USER_ID’); until cookies are manually cleared in that browser. Aside from a few edge cases this approach best enables Algonomy to form high quality personalized recommendations.

Worst-case scenario: you never send a user ID. Recommendations will still be based on wisdom of the crowd but will be sub-optimal at best.

The instrumentation should not pass a userid (that is leave this field empty) until the user has logged in or identified. So, in the scenario where a user is shopping on a site for the first time and is not logged in, the instrumentation should pass a sessionid but no userid. Once the user logs in (for example because he/she makes a first purchase), or is a known user visiting the site, then the instrumentation will continue to pass the same sessionid but will now also provide a userid which is the ID of the user in the e-commerce system.

User Linking

In the ideal scenario, you would know each user across all your touch points and be able to assign the same user identifier in each channel. However, the reality is that very often different channels use different systems making it hard to recognize and assign users the same user identifiers in real time. In some cases, users can be matched up doing some post-transactional analysis. In others, it is a question of the cost and effort of replacing legacy systems. This results in a disjointed picture of a single user, split into multiple different profiles that cannot be connected in a coherent fashion.

The User Linking feature of the User Profile service makes this task as simple as loading a file into our platform.