The Impact of Political Wrangling in the Implementation - Research Paper Example

Data Warehousing and the Political Dimension (First sur no qualifications like Dr. College Position in organisation Place of author Abstract The research is on the very real problem of reducing the impact of political wrangling in the implementation and running of a good data warehouse. There are a variety of areas to be investigated from what politics in a company means to what a data warehouse is, its design, how it is implemented, and to why politics should come into something clearly set up to help, everybody who needs decent data within a company or organisation. Politics in a Company. Politics, it is said, is a blood sport. Robinson (2003). Politics stems from people having competing viewpoints and strategies in a company. There are many reasons why politics can be a difficulty. There is the problem of profitability, who is losing money and why, protection of interests, the need for power, empire building, dislike of someone offering an alternative argument. To some the actual item being discussed barely matters as the political attitudes are very deeply entrenched in the psyche of individuals such that they will always argue; as a knee jerk reaction. Dyche (2000). Politics is always divisive and difficult for senior managers who may in fact be players in such situations or even encourage others to fall out (divide and conquer). There are always going to be some politics in any organisation, simply because there are always conflicting needs. E.G. HR holds all the employment records but does not know about staff salaries and they would very much like to. But the salaries are held by finance who will allow nobody outside their department to have such data as they have held it forever and the manger says that security is the problem. So the managers of both departments have a problem with the other. Such a situation is clearly understood and accepted, until something appears that could shift the power, such as a new IT system, and both managers feel the attack in different ways. HR will feel they can now ask for the finance department databases to be in the warehouse, meanwhile finance are going to block such a move any way they can. Including scuppering or delaying the new system altogether as they are in charge of the budgets, and this is a simple example. Hansen (2008). An area that will need to be investigated is that of existing processes, often there are well known routes through the company system to find out any data that can be allowed to be disseminated allowing for privacy and security issues. It is important in this area to be aware of stepping on managers’ toes. They will not welcome changes unless there is a net gain to their department and work. If managers are happy with their present systems it can be very difficult to get them to change to something new. Robinson (2008). The Concept of Data Warehousing. The development of data warehousing was to be able to mine various existing databases for queries about different aspects E.G. An HR manager may want to know how many people in the company have dental protection and how many do not. But may be driven by the fact that someone in engineering is suing the company for having a tooth knocked out. So the HR manager wants to know how many staff working in engineering have all their teeth left and visit the companies dentist regularly. But the present HR database may not be able to fulfil the entire query, whereas if linked to a second or third database elsewhere in the company such a query becomes possible. And this is a relatively simple query to answer. Combining databases is the start of data warehousing; although for most companies there can be dozens of databases to be combined. This means that to get the fullest possible use of a system there is a need to be able to “see” all the data, no matter where it is. It is a requirement that once set up properly, allows a user to drill down into the data to find precise information to prove an argument or theory, often aimed at added efficiency or cost savings in their own area of the company. There are always more efficiencies and a chance to gain that all-important competitive edge. Greenfield ², (n.d.). Types of Data Warehouse. Dyche (2000) states. The premise behind Occams Razor says that the simplest solution is most likely the best. But overkill technologies, numerous staff resources, and padded budget money are the dictum of the day for many data warehousing projects that should be straightforward and easily executed. The types of warehouse structure used is either the “star” or the “snowflake” type, both have their own particular uses. The star system has a central data warehouse with a few outer blocks coming straight off it, whereas the snowflake concept has the central star system, but has then got child entities coming from each of the separate blocks. Figure 1: The Star Schema The simplest type of data warehouse is the star where there is one set of hierarchies feeding into and out of the warehouse useful these days in an individual department, but not great across a complex company. The problems come from who wants what in the company, if it is seen by managers that they can have themselves as a separate entity they will want it. For all sorts of political reasons they may not like or want to be lumped together in a simple star. Figure 2: The Snowflake Schema A snowflake adds to the star schema by having children (shown in red above) on the entities so breaking down each function into component parts making a more complex requirement for the data warehouse, but also giving a better ability to reflect the needs of the department. It is also possible it may help with the politics among the users, as they will see themselves as separate from others in their requirements and with some input to their own needs. Today a data warehouse is often a part of a corporate information factory (CIF) as in Figure 3. Which includes ERP, eCommerce and CRM. Its architecture includes data marts, exploration warehousing and operational data stores (ODS), to name a few according to Bill Inmon, the “father of data warehousing”, while to others there is a preference for Ralph Kimball’s concepts. The very people who design data warehouses are open to political argument it seems! Inmon (2007). Figure 3: A Modern Warehouse Environment, Inmon (2007). Implementing a Data Warehouse. From 1keydata (2011) in italics. There are a series of stages to go through to implement a warehouse. They are: 1. Requirement Gathering. At the end of the project, what do the users want from the system? This is the start of that decision-making. 2. Set up the physical environment. Getting the servers and applications set up. 3. Data modelling. Building the data, ensuring it exists and then creating a physical data model from the logical data from the user requirements. 4. ETL. Extraction, transformation, and loading the data, this takes the longest time, but it should not be rushed, it can take up to 50% of the project time; but has to be done properly. 5. OLAP Cube design. Online analytical processing, the place where the queries and reports are decided. Users are not always sure of exactly what they want; this is the point where such decisions are made. 6. Front End Development. The actual display users will see, it should be web based as this means there is no need for IT support to have to load it on each PC, and it should help users with their queries and reports by making it simple to do. 7. Report development. Important to the users, this is what they want from the system and comes from the requirement-gathering phase. It involves customising the reports allowing correct access for users and delivery of the reports. 8. Performance tuning. Trying to get slow system areas to speed up, such as at the ETL phase which can be slow at loading. 9. Query optimization. SQL queries can be a problem as they can slow down the application if to long or complex. 10. Quality assurance. A team from the company who attempt to break the system after the developers are done to find any problems. 11. Rolling out to production. This is where the data warehouse goes live. 12. Production maintenance. This is where such things as regular back ups and to see how much users are actually utilizing the system to clarify the return on investment. 13. Incremental enhancements. This is where tiny changes are made that better reflect the business of the company on a regular basis. The Politics in Implementing Data Warehousing. There can be problems with the implementation unless there is firm support from the highest managers in the company. At the start of implementation there has to be agreement about what is to be aggregated, such that there is the best use of an expensive new resource. Again though, certain people will try to rule; it is possible to see some folk with demands for their departments needs above those of others. If the department is sufficiently powerful in the company then others will lose out to them. Robinson (2003). Managers of any company or organisation today recognise that such battles as the examples above are counter-productive and most will try to find a positive outcome wherever possible. Especially when trying to get something as influential as a data warehouse set up, it can be an opportunity to actually get conflicting groups working together if presented and carried out properly. Greenfield ², (n.d.). All the stages are fraught with problems and so any stage could take longer than anticipated. The political angle is complex with an implementation there can be disputes between the different IS teams as to who is responsible for what. Problems between IS and the developers often again, about who should be controlling what area of the implementation. Should the IS team responsible for the implementation also be responsible for the production maintenance and incremental changes, after all they do know the systems internal workings better then the IT departments’ staff. The frustration caused amongst users can be considerable with any delays adding to their disbelief in the system ever coming online and then of course the user-to-user disputes can become heated especially if someone feels ignored at the requirements stage or at the report development stage. An awareness of the politics that can occur before starting an expensive and difficult transition is important too much of the time such projects fail because people ignore the political angle, where as with some careful thought and planning at the start such problems can be dealt with by experienced managers. Demarest (1997). 1. 2. 3. 4. References 1keydata. (2011) Data Warehousing. 1keydata. 5. Web page: http://www.1keydata.com/datawarehousing/datawarehouse.html 6. 7. Demarest, Marc. (June 1997) The Politics of Data Warehousing. 8. Web page: http://www.noumenal.com/marc/dwpoly.html. 9. 10. Dyche, Jill. (8th February, 2000) The Politics of Data Warehousing. DSstar. Vol 4. No 6. 11. Web page: http://www.tgc.com/dsstar/00/0208/101300.html. 12. Greenfield ¹, Larry, (n.d) The Case Against Data Warehousing. LDI Systems Inc. Web page: http://www.dwinfocenter.org/against.html Greenfield ², Larry, (n.d.) The Case for Data Warehousing. LGI Systems Incorporated. Web page: http://www.dwinfocenter.org/casefor.html. 13. 14. Hansen, Gloria. (18th February 2008). How to Benefit Playing Company Politics. 15. Ezine Articles. 16. Web page: http://ezinearticles.com/?How-To-Benefit-Playing-Company-Politics&id=993291 17. 18. Inmon, William (2007) Corporate Information Factory. Inmon Consulting Services. 19. Web page: http://www.inmoncif.com/library/cif/. 20. Robinson, Scott. (11th March 2003) Tread Carefully Through Company Politics. Builderau.com.au. Web page: http://www.builderau.com.au/strategy/businessmanagement/soa/Tread-carefully-through-company-politics/0,339028271,320272759,00.htm. Read More