THE MISSION
5 people are working within the Marketing Intelligence Department (MID). They are reading, synthesing, translating and copying every day a paper press review broadcasted to managers together with synthesis. Too big, not customised for end users ? Press reviews are not read by end users. Even so, this task is heavily time consuming.
Word synthesis are available for everybody throughout the Outlook client in the dedicated competitor surveillance Public folders. Still, information is available, but quiet impossible to find out. No search engine, no queries, of course in Outlook.
Making information retrievable, available and customised for everybody.
1 Business processes
In this kind of technical mission, systems or software are no big deal. The human factor is the critical matter.
1st step identification of the actual business processes
2nd step needs identification
3rd step identification of the non relevant tasks
4th step identification of the sensitive matters. This particular step consists in identifying tasks you are quiet sure you get into trouble if you try to change them. Why ? Many reasons. Cultural, social or simply educational and of course power (information is power !).
Two points may determine if you go ahead or not :
- Level of support you get from hierarchy
- Necessary time to train, evangelise people towards new business processes
5th step is validation of the first results. Compulsory for consultants who wants to be paid ! Information technology involves so many data, cross functional processes, people, for the success of the mission it is needed to go STEP BY STEP hardly ever for huge systems like Datawarehouse.
2 Achived Tasks
2.1 Paper document
Transfer paper document to the existing Intranet owing to powerful tools available on the market. Setting up of a EDM / WORKFLOW process working on a FILENET WATERMARK / @MEZZANINE plateform. For the hardware we used the existing and huge) scanning system already used by Customer Care Department. Unfortunately the implemented search engine is Microsoft Index Server. But it works for basic search by theme, author, media, and keywords as well.
At this particular step, I act as an interface with IT Department. It is not my responsability to require a specific software or system, but I used largely the precise and orientated specifications books I wrote, in order to get the expected system and then results.
2.2 Electronic document / Excel sheets
On Excel sheets are gathered all market analysis (prices, market share, product and services specs…). Available also on the Public Folders, same situation like for Word synthesis. People ( ie sales Sales Department) need to have an answer to queries like "TELEDANMARK subscription price sold in a Debitel shop with a NOKIA 3110 handset". Impossible with Outlook, of course".
I recommend to transfer those information on a database together with a CGI gateway in order for the database to be queried through the intranet.
3 Results (see results)
Information is retrievable and available.
Within the MID section on the Intranet, people can retrieve, query on a single uniform interface document as well as data. Operational Departments are using this system and helps them to react on market evolution. They regularly ask for update. MID get recognition among operational Department. Automation of business processes allows MID to devote on real added value tasks like analyse, prediction, enlargement of sources…As information make a real jump towards availability MID don’t lose their time anymore answering to punctual and personal queries.
Forms had been created in order for everybody to update market information with their own sources. Within a company, many people get their own competitive information. Commercial people, technical people. Actually everybody is intended to bring some interesting substance. The key matter is to describe a procedure which encourage people to report and makes possible to capitalise information.
Conditions
Form on the intranet may be open Money, substantial prices is still the best way to encourage people to report Anonymity may be available Validation procedure must be describe before 4 Next step
Will be to customise information through profile and group definition throughout the intranet. It will make possible to make PUSH in formation towards a dedicated group. Like for Pointcast on internet, many dedicated tools for PUSH exists on the intranet. As a very interesting article on June 1998 INTERNET PROFESSIONNEL warns, PUSH information can be very "bandwidth consuming", hardware and network may be design to afford this throughput. June 1998 INTERNET PROFESSIONNEL reports "in 1997 POINTCAST used 18% of the whole http bandwidth…"
5 Datawarehouse Project
Specificity
As the Intranet, the Datawarehouse is a strategic matter. It involves all the operational departments for inputs as well for outputs. DatawareHouse is most of time very expensive (15 millions FF project is quiet common) and heavily time consuming (and juicy business for consultants…). And like Intranet it is compulsory to describe precisely the expected results to afford such costs.
How to crash a Datawarehouse project :
- Give the project management to IT people. Once again technical matters are important but not critical. Many Datawarehouse projects failed because nice server, huge network and expensive tools are implemented and delivered to users with no idea what results they’re expecting with it.
- No repositories. Making big queries on operational databases can dramatically lower operational tasks (billing, customer care )
- No datamodel definition. Datamodel describes relation between a concept ie. "after how many time a customer activate his subscription "or "how many 60’ subscriptions sold in the shops last month" and logical interpretation through the company’s databases. No validated results without datamodels definition even if they may be thousands in a single company. The Datawarehouse has to be build on datamodel. Very useful anyway to implement business processes. Every operational people has to agree on signification of datamodels ie what is a sale ? When salesman sigh the order ? When the goods are delivered ? Billed ? Paid?
- No validation, no steps. Most of time Datawarehouse projects are huge one. Waiting the whole implementation before any validation from end users and after 15 millions FF expenses, you can be sure the project will fail before it ends.
Read this very interesting "everything you always want to know about Datawarehouse" edited by IBM (725ko). Ok, they promote for their servers but this the state of the art.