About 6 years back we were talking through the data cleansing and quality management options of BusinessObjects Data Services. We ran some small projects and assisted clients in deploying information quality management solutions. Recently I had the pleasure of working with a fantastic team at Fonterra to upgrade their data quality systems on SAP using the latest version of Information Steward. A whole new Data Capability team and strategy was delivered and the extent of the Information Steward product deployed over the organisations global operations. The following 5 points highlight the key success factors in delivering this technology on such a scale.
1. Take the time to understand the whole product offering. Information Steward has multiple facets, and enormous benefits can be gained from integrated use of the complete tool. Although Data Quality can be used effectively alone, the Data Quality Rules can also be linked back to the Metapedia content, and the Metadata Manage content can both guide the developers to the right columns and table to test.
2. Model the outcome desired clearly up front. What are you looking to achieve? Often competing goals drive you away from the best outcome and derail some of the most valuable parts of the project.
3. Are your clients looking for “data extracts” using the tool because they cannot get simple reports easily? If so, reassess your BI strategy. As a BusinessObjects customer, your BI reports should be quick and cheap to develop in-house, and a back door is not needed to get data.
4. Do not underestimate the volumes of requests you will receive. The master data records will give you a small clue on how many data quality artefacts you may need to build.
5. If like many progressive and leading organisations you have distributed the ownership of data to business owners, remember this is a tech tool so dashboards and reports are not ideal for business consumers. A universe designed over the repository with a simple datamart to allow bridging between modules can help deliver reports and dashboards to business users. Using conditional formatting reports can quickly highlight problems.
Those were the biggest five however we encountered a host of interesting and fun challenges in delivery of the project. The difference here is that instead of a bunch of disconnected tools SAP has now brought them all together in a comprehensive suite. If you are a Business Objects Data Services customer, you also can still use the data cleansing packages so overall, a great solution to a strategic problem.