Managing BI success

Why do we manage projects? So that the expected deliverables will successfully be created within set timeframes and budgets. Key is the "expected deliverables". And isn't that so often the difficult part to define? The pot of money is often well defined, as are the timeframes. Each stakeholder, however, has a different definition of success; so what is successful for one party does not necessarily mean the same for someone else.

Let's focus on the two words "expected" and "deliverables" as they need to be managed differently.

You're on your way to a meeting with a client, but you're stuck in traffic, and you know it's going to be tough to be there on time. There are two options:

* The one most of us would choose: we drive like maniacs and try to be in time. Speeding, staying in the slipstream of taxis on emergency lanes, and jumping a couple of "orange" robots are some of the means at your disposal. In the process, you also cut a couple of other cars off. Finally you arrive, but you're still 10 minutes late, and you're stressed. And best of all: it appears your client was in the same traffic jam and happened to be the driver of one of the cars you cut off. What a great way to start the meeting!
* The second option is one most of us don't consider: stay calm and call the client's offices to let them know you're 15 minutes late because of the traffic. You find the client happens to be in the same traffic jam, so it's no problem at all that you're late. Problem solved: how simple was that?

There is a generic model behind this. Going for the first option means you're trying your utmost to align deliverables with the expectations (being in time for the appointment); the second aims at changing the expectations around these deliverables (I'll be in a bit later).

But how about the every-day challenges we face in business intelligence (BI) projects? An excellent method to obtain a clear understanding of the various expectations of people involved is to conduct a stakeholder analysis.

Preferably this analysis takes place at the start of a project, but it also pays off when conducted at a later stage in the project. Having a short talk with each stakeholder separately about personal success criteria for the project will be very useful. Quite often you will find stakeholders are simply looking for a success story, in other words: "Will I be considered successful in my organisation when this project comes to an end?" When entering this type of discussion with direct clients, you're actively managing their expectations.

Also think of situations where you work extreme hours to get the requested five years of history imported into the data warehouse. Did you ever speak directly to the person that requested this? Maybe two years was more than enough if you addressed this with the right person, and this gives the opportunity to be less stressed and focus more on the quality of the work.

Doing BI projects means the organisation is in discovery mode.

Erwin Bisschops is senior solutions architect at Harvey Jones Solutions.

When it comes to managing deliverables, what is key to being successful? The answer is what Woody Allen once said: "Create ideas that are 15 minutes ahead, not light years ahead." In this context, the success of BI solutions is directly related to user acceptance of the BI deliverables. And this acceptance won't take place by creating an overly difficult solution, but by creating relatively easy solutions that people actually understand and, therefore, use.

Too often I've trained end-users from various organisations on their own cubes with 20 dimensions and 40 measures, and these end-users battle to understand what they're looking at.

Doing BI projects means the organisation is in discovery mode. The BI project team puts business processes, business rules, definition issues, as well as the (quality of) existing data assets under a magnifying glass. The business knowledge gathered can be huge and sometimes overwhelming. Therefore, the BI team can be light years ahead of the actual end-user community.

Reporting requirements in a project team environment are often overstated, meaning a lot of energy can be wasted on topics like slowly changing dimensions or the ideal partitioning strategy. But quite often the simple requirements are overlooked; for instance, having a simple one-click YTD sales report with verified data. The following picture illustrates this.

Hard work is good, but hard work done in a clever way is better. Successful management of BI projects includes both managing the deliverables, and managing the expectations around those deliverables.

* Erwin Bisschops is senior solutions architect at Harvey Jones Solutions.