Enterprise resource planning (ERP) is more than merely a new application; it is a new way of doing business as each ERP module installed makes demands on how the business is run, forcing best practices methodologies and processes on the organisation.
A recent study by The Conference Board painted a poor picture of ERP installations. The study reported its respondents suffered from a series of complications when deploying ERP solutions ranging from difficulty in developing a business case, realising value and staying under budget, to soaring maintenance costs and a constantly moving target for implementers to track.
Having the right data for a specific purpose delivered to the right place when needed is key.
Julian Field, MD, CenterField Software
Yet corporations still undertake ERP projects with the expectation they will achieve a rate of return exceeding the cost of implementation. Delivering business value as quickly as possible while minimising the expense normally associated with deploying and maintaining these applications are important goals in every corporation installing a new or even an upgraded ERP solution.
To achieve these goals, data must be a top-of-mind issue for project leaders. Having the right data for a specific purpose delivered to the right place when needed is key. Careful management of where data is sourced from, how it is combined and its accuracy verified and how it is finally moved into the new ERP database can speed project time and reduce future problems.
Using purpose-built tools for these tasks can help businesses accelerate time-to-value and contribute to a reduced total cost of ownership of business applications. They also cut down on the amount of coding needed to get data to and from source data stores, leading to lower staffing as well as lower maintenance costs.
There are seven building blocks to efficient data management. These steps relate to the extraction, transformation and loading of data from multiple sources into a final, single database for use by the ERP system.
Building blocks
1. Connecting to all possible sources: Data can come from a variety of sources and the data collection process should easily enable connectivity to these sources via a simple graphical interface that requires no technical programming knowledge to operate.
2. Profiling data: Data from these sources needs to be profiled to enable the users to understand what data there is and how it interrelates. In other words, developing an understanding of the data content: quality, completeness, validity, compliance with business rules, etc.
3. Transformation normalisation: Data needs to be converted from the various formats it is stored in into a common format consistent with the rules of the business.
4. Data quality: This process consists of analysing and developing rules to correct specific data quality problems that are critical to the company. For example, incorrect values, duplication of data and inconsistent data content needs to be corrected and then verified as an accurate version of that data before being accepted into the ERP solution. The cleansed data should also be returned to the original source to ensure all corporate systems use the most up-to-date information available.
5. Interim store: Automated extraction, transformation and loading (ETL) procedures need to be followed to reduce reliance on people manually coding transformation routines - which will reduce the overall cost. An interim data staging location should be used to bring all the extracted information together in one place and in one format. From there, final validation rules will be run against it to ensure there are no duplicates or bad records.
6. Running in parallel: To support the integration effort and overall project in a timely manner, the data tools used must be able to run their ETL processes on multiple sources at the same time to speed the process.
7. Central repository: Since an ERP implementation process takes time, managing all the integration metadata in a central repository makes for an easier process the next time data needs to be read to or from the source data stores.
There are many complex steps involved in the process described above and various products designed for each of them, such as metadata management tools, data quality tools or ETL tools. These can be bought from different vendors, but for optimal results they should be handled from a single, integrated platform.
The ETL process is meant to smooth the ERP implementation process, not complicate it by forcing IT to waste time, resources and effort integrating a disparate set of integration tools. Compared to manual coding, project time can be cut by 40% on average with the correct, integrated ETL tool. The most cost-effective approach to data integration over the life of an ERP deployment that extracts data from a variety of data sources is to use an integrated platform for data profiling, discovery, quality assurance, preparation, transformation and movement procedures.
The bottom line benefit for the organisation is not only lower total cost of ownership through cost reductions, but also an expedited realisation of the benefits because the solution gets to production more quickly than via traditional methods of data integration.
Share