Today was the second day of a Naked Data Bootcamp (basically a paid Proof of Value) at a client site. Even though I have been involved in many of these, it is always an extremely satisfying moment when you watch a client finally “getting it”, and understanding the range of flexibility and power that Jedox (and your solution you are building them) offers.
For me, getting the client’s penny to drop is and end game of a delicate process – we want to really deliver what they are looking for but at the same time steer them to a design and a model that will be robust and long lasting. The Sweet Spot.
The client wanted to be able to analyse down to a detailed level of granularity but the level they wanted in the model just did not make sense in a multi-dimensional world of a Jedox cube. It was just way too flat and transactional. I knew that if I built the model the way that the client was demanding, they would be left with a model that was sub-optimal in performance.
I convinced them that we should build a model with summary data (still with 30k plus elements in some dimensions), with drill through capability to their relational database for the flatter, transactional data. Although unsure at first, once they saw it in action, the client quickly agreed that this would more than satisfy their requirement.
There was a fair bit of time white boarding and convincing the client that this was the way to go. Why bother? We could have easily built the model and finished in half the time if we initially agreed to their demands. From experience, however, managing expectations and communicating your rationale is critical in the long term success of a model and in forming a great relationship with your client. It basically is an exercise in building trust.
Sometimes you need to risk the relationship by backing yourself and your experience. Initial engagements mean that people come to the party with their own expectations and baggage. It takes them a while to understand that you may be offering them a new way of looking at their old problems.