CATEGORY: Data Quality

REVIEWER: Bruce Richards, information technology director for Harte-Hanks Ltd.

BACKGROUND: Harte-Hanks Ltd. is a European division of the U.S.-based marketing giant Harte- Hanks, Inc. Headquartered in London, we specialize in providing customer relationship management (CRM) services, database and e-business solutions, and other marketing-related services to clients across many industries. Harte-Hanks also produces and distributes the award-winning Trillium Software System, a complete CRM data quality product. Currently, the majority of our customers are in the UK, but we are rapidly making inroads into mainland Europe.

PLATFORMS: SyncSort UNIX runs on all major UNIX platforms. Harte-Hanks Ltd. uses a Solaris platform on a variety of Sun enterprise-class hardware.

PROBLEM SOLVED: Creating single customer views is an important early step in the CRM solutions that we provide for our clients. Some clients have many customer databases arranged around different products or institutions they've acquired, so they can't easily identify all the information for a single customer across all lines of their business. Typically, we receive a number of multigigabyte flat files from a client. Each file contains differing fields, layouts and record lengths. To identify the key fields that define a single customer, we sort the files numerous ways and compare them in great detail on many different keys. Harte-Hanks' own Trillium Software System is used for much of the process, and for high-performance sorting we rely on SyncSort UNIX. A single sort job for us can involve tens of millions of 3,000-byte records that need to be sorted many times.

PRODUCT FUNCTIONALITY: At Harte-Hanks Ltd., we use SyncSort's basic sort functionalities. For high-performance sorting, this product is faster and more versatile than anything else we've found on the market. In the near future, we anticipate using SyncSort to presort files for rapid bulk loading into relational databases. We also plan to leverage SyncSort's parallel processing capabilities. SyncSort UNIX can run many sorts simultaneously with no degradation in performance.

STRENGTHS: Sorting, especially on multiple keys, is one of the most resource-intensive data processing functions, so maximizing sorting speed and efficiency is critical. SyncSort's greatest strength is the sheer performance of the sort, grounded on a well-established, proprietary sorting algorithm. SyncSort UNIX easily integrates into our process flows and performs its job in a very low-key fashion. We experience no problems with it although we use it every day.

WEAKNESSES: Though SyncSort's sort utilities are remarkably easy to use, other more elaborate features such as advanced reformatting can initially seem somewhat more complicated. However, this is to be expected given the depth and versatility of the overall SyncSort UNIX product, which has a wide range of data manipulation functions that extend well beyond simple sorting.

SELECTION CRITERIA: When we made the decision to run this business on a UNIX platform, we had many issues to address; and we didn't wish to tackle the sorting problem. With our huge data volumes, we needed a sorting tool that was going to be industrial-strength, would do the job efficiently and would give us few headaches. We looked at SyncSort UNIX based on recommendations and favorable experiences we had with Syncsort's mainframe product. SyncSort UNIX absolutely does the job.

DELIVERABLES: Although SyncSort is a key component for creating single customer views, it is also used for other processes such as data enhancement. For example, a Harte- Hanks client might have customer data in three separate files that are in three different formats. After identifying the unique customer keys, we use SyncSort to sort the files on those keys, making it easy to merge the data and to identify data discrepancies among the files.

VENDOR SUPPORT: We have only occasionally needed to call Syncsort's help desk. When we have, we've found the support to be helpful and the support personnel to be knowledgeable and responsive.

DOCUMENTATION: SyncSort UNIX arrives with a comprehensive user manual and several overview booklets. All documentation is accurate, complete, well-indexed and useful.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access