Each industry is different.  Some have huge margins but low volumes (e.g. Software) while others have huge volumes but very low margins (e.g. QSR).  Needless to say, the approach within each industry has to be different.  There are several areas the clients need to consider when building a business intelligence and Big Data infrastructure in the QSR industry.  These comments are based on actual field experience while being engaged at Burger King Corporation in their BI initiatives.  A note on why BI and Big Data are talked about together in this article … in our opinion, Big Data Analytics are an extension of BI.  One can say that Big Data Analytics is the new BI (i.e. BI 2.0).  Back to the considerations when building BI infrastructure:

  1. Business Intelligence is worth everything:  It’s not about the data, it’s about the intelligence that the data will provide!  All too often, IT departments get caught up in the details of technology.  The real value of technology and data is what you can derive out of it.  In and of itself, data and technology is useless.  This point is obvious but most (I repeat most) BI implementations are technology-driven.  So, the key is to make sure that the *business* drives the BI initiative.
  2. Clarity of the outcome:  Clarity is power!  Know your metrics.  Knowing what the “industry” metrics that you need to monitor (such as average check, traffic, sales per store, etc.) and specific metrics to your organization are key to building the BI strategy.  The clearer you are, the faster you can implement.   This clarity also ensures that IT will implement in a critical path.  Therefore, always implement top-down (business->IT) and not the other way around.  The ideal IT team is the one that sits with the client and develops the solution.
  3. Data Volume:  the data deluge is real for all industries including QSR.  Therefore, it’s very important to know exactly how much data you need to keep.  It’s important to know how the importance of your data tapers off as it decays.  So, if 90% of all BI is done on data that is 30-90 days old, then don’t invest a lot of time for the other 10%.  Know thy intelligence.  A lot of IT effort is wasted on infrastructure (data and technology) that does not provide value.
  4. Control rampant requirements:  If I were to ask an end-user how much data they would like to keep, their most likely answer would be “Everything!”.  This is a recipe for disaster.  Business users need to make 100% sure that they are clear on what they want and why.  Vague requirements create a lot of scope creep and wastage within the BI infrastructure.  It’s easy to just build it and feel like you are making progress.  However, cleanup efforts can cause a lot more than building it.  This can be seen when old systems are “sunset”.  The cleanup effort can cost up to 10x the build cost.  So, net net, ensure that you are very clear about what you are building.
  5. Data Modeling:  To build a powerful BI infrastructure, you need a powerful data model that represents the business accurately.  So, during the build phase, make sure you invest a lot in getting the data model right.  A good data model is also very fast.
  6. Project Approach:  Agile. Agile. Agile.  There is no other way.  Teams have to run in “lean” mode and the whole BI strategy should be made up of “user stories”.  User stories are capabilities that key business people need in order to create better intelligence.  For example a user story for a marketing manager could be something like “As the marketing manager, I need the capability to get feedback for my coupon campaigns within 8 hours of the campaign being initiated so that I can evaluate its effectiveness and change it if necessary.”
  7. Technology – Data Layer:  Less is more.  Using an appliance approach (e.g. Netezza) vs a full-blown database approach (e.g. Oracle) simplifies the BI infrastructure and allows it to become much more nimble.  As with the project approach, the Technology aspect of BI should be very lean.
  8. Technology – BI Layer:  With respect to reporting technology it is important that all deployment of information is done over the web.  With HTML5, a rich experience can be garnered for the end-user.  The BI tool should be able to present all interactivity over the web and through smartphones and tablets.  It should be assumed that users will expect to get information anytime and anywhere.
  9. Data Acquisition:  Understanding source data and getting access to it (i.e. If it’s not native within the organization) can be a big unknown.  Data acquisition, if not properly controlled, can become an never-ending project!  Therefore, it’s very important to have clarity of the outcomes of the BI system.  This will help weed through the “noise” in the data and start creating value instantly.
  10. Personalities – We believe that success comes from people and not technology.  The best people create the best solutions.  Therefore, this is one of the most important aspects of success in your BI initiative.  Getting the right personalities and skills within the BI team will create success quickly.  Choose wisely and make sure you select players that have the skills and experience to deliver.