User Input in Business Intelligence projects

Recently, I read Simon Sabin’s blog on this importance of intelligence in BI, which prompted me to revisit the idea of user inclusiveness in the process. I originally posted a blog as part of the Microsoft ‘Industry Insiders‘ series a while ago, and as I continue on my BI journey, I’ve updated my thoughts on the importance of users in BI projects as a result of going through different project experiences.

It can often be easy to underestimate users, and I’ve even heard a perspective that users should be completely insulated from the process because they are not ‘technical’. However, users can sometimes surprise you when ‘water cooler’ conversations reveal interesting things about their business. I produced some geospatial mapping graphs recently using Tableau, and I must say that I was very pleased with the results! Being based in the UK, I had imported a series of UK postcodes along with their corresponding longitude and latitude for use in the graphical representation of the data. A ‘water cooler’ discussion revealed that I didn’t need to do this; in fact, a distinct work stream at a completely separate part of the business involved on delivering GIS software, so this gave me another opportunity to use the best possible data source for the display of the data.

Users can be very creative at generating the data that they need to do their everyday jobs, and it can be useful to harness this creativity to generate proper data warehouses. User creativity can be seen in situations where there are lots of Excel spreadsheets or Access books lurking around the business, individually maintained and updated by users in the absence of a central data source that meets their requirements. Additionally, this creativity can mean that they find workarounds in manipulating their systems to do what they need. When these situations arise, it is important to collude with users to ensure that all the necessary data is in the data warehouse, in the format that they need. Thus, the business teams more likely to adopt the solution, and the business benefits overall because the data is centrally stored, backed up, and has a proper lineage. Don’t get me wrong, I love Excel; but I don’t like the situation where there are lots of Excel spreadsheets floating around, unchecked, and not validated enough to be considered a proper authentic data source.

The initial blog grew out of a discussion I’d had, where a number of people were debating the need to have a business representative or business-focused Project Manager involved in the design stages of a data warehouse. My own view is that business users should have input at every stage of the development. User buy-in is absolutely critical to the success of a business intelligence project, since they are the ultimate consumers of the information. If they don’t like it, they won’t use it – it’s as simple as that! Users need to be involved right from the start. They can save you a lot of additional, unforeseen work since they will tell you directly if anything is missing from the warehouse. I came across an example of this, where the report consumers were not involved at all until the training. During training, the users noted with a great deal of disappointment that quite a few of their key criteria were simply missing from the warehouse. This led to discouragement and a loss of user confidence in the system. The project was ultimately leading to failure; since the users could obtain these key criteria outside of the warehouse, they saw no need to use the new warehouse.


My brief was to turn this situation around to ensure successful delivery of the project. The first step in doing so was ensuring that the users all had their say in the content of the data warehouse. Since these omissions had been discovered during the training phase, the project had been at a mature stage of development at that point of time. Thus, there was a considerable amount of re-work at every stage of the project to ensure that the warehouse contained the business users’ requirements to support their decision making processes. Needless to say, if the business users had been involved at the beginning, there would have been no need for the extensive re-work.
When building a data warehouse, it can be tempting to exclude users and to ultimately dictate a solution to them. This is particularly the case where users have never been exposed to a business intelligence solution and have a long journey ahead of them to understand the difference between a data warehouse, and a database. This situation is compounded by the fact that developers are not always the most gregarious people, and may possibly even be naturally disinclined to speak with customers.

I appreciate that it can be difficult to explain complex concepts, such as cubes, to business users. One customer told me that a ‘grey fog of confusion’ descended on his brain whenever he heard the word ‘cube’! However, I have found that using Microsoft BI products have helped business users to really see, understand and use their data have the ‘Aha!’ moment and to really understand. For example, showing business users the contents of a cube in Excel or in the Analysis Services browser supports their learning journey to the ‘Aha!’ moment.
The users’ journey to understanding and confidence can be supported by actively involving them at different stages. One useful way of doing this is to get the user’s input when creating cubes. I achieved this by listing out the possible hierarchies and attributes in an Excel spreadsheet, asking the business team to organise them to reflect their business structures. Since they knew the business well, they did this easily. When they saw the completed cube, it was a real confidence boost for them to see that they had had direct input into the creation of the cube and that it was a correct translation of the data into their business needs. The business users’ confidence levels were increased by the fact that they could interact with the cube in the comfort zone of Excel. Ultimately, involving the customer at all stages directly helped them to accept the system. Getting business users to accept, understand and interact with a business intelligence solution is often one of the hardest parts of a BI project. However, in my experience as a BI practitioner, the Microsoft BI stack of products makes this easier since it is directly aimed at this group of users who will ultimately determine the success – or otherwise – of the Business Intelligence project. What’s the point of delivering a technically perfect solution if nobody will use it or can understand it? Solutions like
Tableau, XLCubed are an excellent addition to existing solutions since they enhance the analysis of the underlying data, and so they’re also worth considering.

Leave a Reply