Products

Zoral Behavioral Data Warehouse

December 2, 2016
“The problem with the internet is that you can’t see your customers. Behavioral data fixes that...”

Behavioral data can be highly predictive. It improves customer visibility and helps you make better decisions. It allows you to adapt your digital customer journey to each customer, improve conversions, increase automation, reduce cost, risk, fraud, defaults and much more.

The more you know about your customer, from the first point of contact onwards, the better you can respond. Companies who use behavioral data efficiently are more profitable. They react more quickly and appropriately to customer and market demands.

However, behavioral data is hard to gather and control. Plus it is difficult to build such systems from scratch. Also, not all behavioral data is predictive. Answering such questions as... “Which data items to collect? Which are predictive and when?” is complex, time consuming and often requires much trial and error.

Zoral BDW solves this problem. BDW is a fully featured, Behavioral Data Warehouse with a comprehensive toolset. It automates the collection, storage and quality control of your behavioral data. BDW contains many 100’s of man years’ research into the commercial application of behavioral data. It knows which features are predictive, of what and when. So, there’s no need for trial and error.

By using BDW, your organization can rapidly implement and benefit from a fully functional, predictive behavioral data warehouse at a fraction of the time and cost it would take to build such a system in house.

BDW allows your team to focus on using behavioral data rather than the time-consuming tasks of trying to work out what is predictive, acquiring, testing and managing it.

What is behavioral data?

Behavioral data is derived from human, device or system behavior, as opposed to structured, application data.

For example, a user might be completing an on-line application for a product via your website. In this example, behavioral data might include: were bank details entered manually or using cut/ paste? when pasting, did they use the mouse or the control keys? how fast did they type? what device did they use e.g. tablet, phone, PC, iOT device, etc.? What version of the operating system does the device have? etc.

Typically, during a 1 minute, on-line session, between 5,000 - 12,000 items of behavioral data/features will be collected. Similarly, voice data might be captured via a service desk, digitized, extracted and combined in the BDW. Using BDW we can build a comprehensive “picture” of the user, which can be highly predictive in a wide range of applications.

BDW captures a wide range of predictive features such as...

  • Click data (e.g. button click, URL click, keyboard usage, type speed)
  • Clipboard usage
  • Device data (e.g. device id, device type, cookies, web history, cache)
  • Environment data (e.g. environment ids, platform type, browser type, screen resolution, language)
  • Browser data (e.g. browser version, window size, window position, scroll position, orientation)
  • GeoLocation data (e.g. longitude, latitude, address)
  • Errors
  • Page elements interactions,( e.g. navigation between elements of the web page,focus type by key or by mouse, lost focus, duration of interaction with UI element etc.)
  • Value data (e.g. values entered by the user, granular history of changes of the values, including time stamps and speed of changes)
  • Text Deletion data (e.g. original value, position of deleted text, aux value, etc.)
  • Web-data (e.g. page load times, window focus time)
  • Marketing campaigns data (e.g. tracks marketing campaigns and sources)
  • Visits and Sessions (e.g. tracks sessions, visits per session, page views per visit)
  • Session data (e.g. client IP address, session start and end time stamps, number of visits, user identification)
  • Mouse data (e.g. mouse track, movement speed, mouse speed deviation)
  • Social network logins (e.g. Facebook, Twitter, Google)

How is behavioral data used?

Behavioral data allows you to build a 360° picture of your clients. It can be used for far more than just simple scoring, (e.g. credit scores etc.). For example, by the way people type, the way they press the keys and use their mouse etc., we can accurately predict age, sex, level of education, identity and many other items.

Typically, predictive behavior data is combined with “conventional” data such as, application data, credit bureau data. This can significantly boost the accuracy and range of predictive results. This should come as no surprise. Banks and Insurance companies used to operate via branches and agents. They would see and get to know their clients, observing 100’s of behavioral characteristics, which they would use in making sales and service decisions.

Zoral behavioral data warehouse, KPI optimization

For example, an insurance agent, calling at a home might notice that there is a new car in the drive, children’s toys in the garden and that the house has recently been repainted. These observations would immediately “tune” his sales approach and views on risk, choice of product, pricing etc.

With e-commerce, this contact has been lost. However, using Zoral BDW it can be replaced and improved. Having good quality, behavioral data allows better understanding and prediction of likely behavior. However, it can help beyond simply improving or boosting the performance of existing, predictive models. It enables you to “see” previously hidden detail.

BDW makes a big difference

For example, Graph 1 shows an actual loan portfolio categorized, using credit bureau data, into 10 risk score categories, (highest risk of default on the left, 4 categories of default, etc.). Whilst it is reasonably efficient at discerning the extremes of “good” and “bad” risk, you can see that it becomes flat and “confused” around the middle range.

This is significant as, based on this, many “good” risk clients will be rejected or “bad” risk ones accepted. In reality, many companies simply avoid this part of the market, or approach it with great caution. However, if we add BDW behavioral data to the same portfolio analysis, the impact can clearly be seen, as shown below in the Graph 2.

Bureau based score
Bureau and behavior data based score

It is now much clearer, with the distinctions in the mid- range exposed. Thus, BDW not only increases the accuracy of decisions, but also widens your market reach. For some clients/market sectors this translates to a 15%- 20% increase in addressable market. So, BDW can make a big difference.

BDW is easily installed and integrated

BDW data can be easily integrated with other predictive data sources. Using this combination, you can achieve a significantly higher degree of accuracy than using conventional data sources on their own.

Zoral decision engine, custom behavioral features

BDW also improves decision speed. It is gathered in real time, so can be used for instant decisioning, whilst your customer is interacting with the system. It can perform real-time data enrichment so your system can make automated, intelligent, next-best-action decisions. Also unlike many other data sources, behavioral data is free.

You can optimize many functions by using BDW data in decision workflows and models. For example,

Zoral behavioral data warehouse, customer touchpoints
Zoral behavioral data warehouse, fraud optimization
  • digital marketing
  • robo-underwriting
  • dynamic customer segmentation
  • customer lifecycle management
  • conversion optimization
  • fraud detection/prevention
  • retention management
  • risk adjusted pricing
  • churn prevention
  • collections optimization
  • marketing campaign optimization
  • cross-sell, up-sell
  • user identification
  • KYC
  • user digital fingerprint
  • next-best-action
  • mass personalization
  • dynamic customer journey
  • affordability prediction
  • claims prediction
  • and many others.

BDW...

  • is easy to install and integrate
  • handles very large, real-time behavior data streams
  • is a fully configurable and customizable, enterprise-level product
  • comes with “out of the box,” pre-configured data capture for a large range of KPI’s

zoral

Zoral is one of the world’s leading companies in the collection and use of behavioral, unstructured, MNO and social data for predictive analytics.

With over 100 data scientists and ai/ml engineers, Zoral operates one the largest Artificial Intelligence/Machine Learning labs in Europe.

Zoral is recognized as a R&D pioneer in this field. It has a global client base, (Europe, USA, Canada, South America, Asia, Australasia).

zoral knows

It takes a huge amount of research to know which combinations of data features are predictive and when. With BDW you instantly gain the benefit of Zoral’s many 100’s of years of R&D from one of the most advanced research centers in this field.

Zoral knows which behavioral items to collect and when/how they are useful. With BDW, you will too.

zoral impact

With improved customer visibility, you can tune each customer journey in real time and mass personalize your digital products, across all devices and customer touch points.

BDW data can be used in exactly the same way as other predictive data sources. So you can easily and rapidly enhance predictive modeling and improve business KPI’s.

Zoral BDW, non-intrusively captures predictive behavior data, generating ”Big Data” volumes. It is simple to integrate with your existing decisioning or other Zoral modules such as Decision Engine (ZDE) and Zoral Standard Model Library (ZML).

15%
Reduction
in Fraud Rate
(Consumer Credit Australasia)
33%
Increase in Acceptances
(On-Line, consumer credit UK)
22%
Increase in renewal rate
(Consumer Credit UK)
18%
Decrease in defaults
(On-Line, consumer credit USA)

Rapid results

Zoral BDW can be used stand-alone, or in collaboration with 3rd party decision engines, data sources and analytical software. It can be rapidly put into production with minimum disruption and cost. It is highly configurable and equipped with a simple, high performance integration toolset. So, unlike many, conventional data warehouse projects, benefits and return on investment can be achieved within months.


Features List

Nine questions

  1. Are you using behavioral data to model, predict and optimize each step of the customer lifecycle?
  2. If so, how much do you use?
  3. Could you use more?
  4. What is the data quality? How is measured?
  5. How much better could your results be if 1-4 were optimized?
  6. How much more could you automate your operations if behavioral data were applied optimally?
  7. How does your organization compare to market norms and best practice in this field?
  8. What would it add to the “bottom line” if Behavioral Data modeling were applied to each, applicable area of your business?
  9. To answer the above where should you start?

The answers

The good news is that the bulk of the above questions can be answered in as little as 4 weeks. Contact Zoral to find out how.
Email: sales@zorallabs.com

This document is provided by Zoral Limited and its affiliated companies (“Zoral”) for informational purposes only, without representation or warranty of any kind. Zoral shall not be liable for errors or omissions with respect the information contained in this document. Product Specifications are subject to change without notice. The only warranties for Zoral products and services are those that are set forth in the express warranty statements in Zoral’s standard contracts for such products and services, if any. Nothing herein should be construed as constituting an additional warranty.

© Zoral Limited 2017