The “Hudson’s Bay Company”:http://www.hbc.com/ (HBC) is in the process of deploying a common business intelligence (BI) sales reporting system across its major retail The Bay department stores, “Zellers”:http://www.zellers.com/ discount stores and “Home Outfitters”:http://www.homeoutfitters.com/. In this article on ITbusiness.ca, they openly share their implementation difficulties for this large integrated project, which involves over 5,500 end-users.
Kudos to HBC for sharing. I suggest you “read the full article”:http://www.itbusiness.ca/it/client/en/home/News.asp?id=40609, but here are my three major takeaways.
The first takeaway is that the days of defining functional requirements and handing these over the wall to the Information Technology (IT) department are gone. For complex, integrated projects, business people need to bring IT into the loop early in discussion so that the tradeoffs between business and IT requirements are worked together, not as salvos fired back and forth about what business can and can’t have. A major cultural change on both parts, this is easier said than done.
The second any analytical or reporting project needs to take future, unanticipated requirements as a given. Market competition means change will happen, faster than ever. Scope creep, anathema to IT, is therefore a given. So build flexibility. If you don’t, the system won’t be used. HBC analysts reportedly want to hang onto a deep level of data in their new reports, even though they don’t need it right now, because they might need it later. This means two things – (1) the analysts aren’t confident that the system will provide the necessary data in the future, and (2) they probably have their own analytical tools on their desktops to distill the data they need right now, defeating the purpose of a central warehouse with consistent sales data presentation.
Third, for large complex projects, run pilots. Pilots may appear to slow down the launch date of the project, but it’s better to fail on a small scale and find out the gaps early. HBC acknowledges they probably should have run pilots. Terminology (e.g. what’s a “sale”?) at a detailed level is being clarified. Some senior managers are refusing to use the dashboard reports. Isn’t it preferable to discover problems such as these when only a few reports have been built, so that the issues can be addressed and fixed before a broader rollout?