Security Normalization

Overview

Security normalization is a process that identifies and gathers complete information associated with a security. The process uses the information aggregated from a consumer's account and compares it against Yodlee’s proprietary reference data. The comparison is performed by a complex set of algorithms that can handle multilevel matching using the security’s data attributes. The following pieces form the core of security normalization:

• Aggregated data
• Yodlee security master (i.e., reference data)
• Normalization algorithm

Aggregated Data

Yodlee syncs and aggregates data from more than 3,000 sites in the wealth domain. After a consumer links an account with Yodlee, all the information related to the positions, transactions, and investment options from the account are aggregated. The position data is then run through the normalization process to accurately identify securities as shown in the image that follows.

Yodlee Security Master

Yodlee maintains a proprietary security master that has more than 4 million securities across 80+ exchanges and a dozen asset classes. The master database stores all important information about the security including ticker symbol, CUSIP (Committee on Uniform Security Identification Procedures), ISIN (International Securities Identification Number), SEDOL (Stock Exchange Daily Official List), traded exchange, currency, latest trade price, security type, etc. Third-party data providers such as Standard & Poor's (S&P), Envestnet, and Xignite are used to keep the security master updated with the latest information about public securities from around the world. Yodlee also provides support for private securities including hedge funds and private equity/venture capital investments.

Normalization Algorithm

The security normalization algorithm uses information from the aggregated data and attempts to identify the exact security by matching aggregated security data against the Yodlee security master reference data. When financial institutions do not provide complete security data, the normalization process helps identify the correct security. Using the algorithm, all positions aggregated from the consumer's account are normalized, thereby updating and enriching the missing information to provide a more complete picture of the consumer's portfolio.

The security normalization algorithm performs the following steps:

1. Security Identification – The identification process keeps running through the scenarios in the following order until single security is identified from the security master:
   1. Match using the CUSIP, ISIN, SEDOL, and VALOREN
   2. Match using the ticker symbol
   3. Machine learning-based scientific approach to match holdings using the security description.
   4. Machine learning-based price model used as the selection criteria for securities.
2. Security Elimination – If the security identification process identifies multiple securities, then the security elimination or refining process is performed using exchange, currency, and domicile/region.
3. Security Verification – If single security is successfully identified after the security identification or elimination process is complete, the verification process is done to ensure the accuracy of the security identified. The verification process compares multiple data elements from the aggregated security with the security master data. Key fields such as CUSIP, ISIN, SEDOL, VALOREN, exchange, currency, ticker symbol, trade price, etc., are used for the security verification comparison process.

Normalization 2.1 Benefits

Normalization 2.1 is our latest version. The benefits of Normalization 2.1 are as follows:

• Machine learning to identify securities
• More complete coverage of security types and account types
• Centralized logic for security type and account type identification
• Comprehensive security master

Security Types