Let’s Put Analysis Back into Analytics

It’s all the rage.  Everyone is doing it.  But what exactly is everyone really doing?

Simply stated, analytics should be about analyzing something.  Similarly, reporting should be about reporting something.  When we use an analytics tool to present last quarter’s KPI’s (or measures or metrics or whatever you want to call them) and we let our executives simply view the report (sorry folks, a dashboard is just a report when you use it this way), we are just straight-up reporting – not analyzing anything.

Now don’t get me wrong.  There are some terrific analytic solutions out there that really help a talented business analyst do his or her job.  Have a look at SAP’s Rapid Deployment Solutions for Analytics for example –analytic solutions are available for many different Industries and Lines of Business, which can implemented in only a few weeks.  But to effectively use these analytic solutions for analysis you need to understand the analytic process.  There are numerous variations on the theme – just Google PDCA (Plan-Do-Check-Act), Shewhart Cycle, OODA (Observe-Orient-Decide-Act) loop, DMAIC (Define-Measure-Analyze-Improve-Control) from Six Sigma, Analytic Hierarchy Processes, and you’ll get your fill – but they all basically boil down to a few key concepts:

1. Analytics are iterative – as soon as your analysis is complete a new question arises.
2. Analytics focus on an addressable problem – more than a simple retrospective question (How many units did we sell?), the problem must relate an action and a result (more later).
3. Analytics begin with a hypothesis – If I do this, then that will happen.
4. Analytics rely on data – How will you support or refute your hypothesis?

Now, I for one, believe that here is where Performance Measurement (what I call Reporting) and Performance Management (Analytics) diverge a little bit.  I believe that the focus of reporting is packaging your analytics into a repeatable process.  Twenty years ago, all you needed (well – all you got) was a P&L or a Sales Report, or a Headcount Report.  Today, you get a cube or a graph (sorry – a visualization) or a dashboard.  You’d get it monthly, weekly, daily or maybe even near real-time.  Regardless – it still tells you what you did without an expectation (necessarily) of an action. 

Most of these “analytics” started life with lofty goals of being “game changers” (LOVE that word J), but for the most part they have become simple scorecards that use the web and cool graphics (oh yeah – and iPad!) to fundamentally just replace the hundreds of pages of “green-bar” reports that I used to pour through back in the day.  (By the way, your new hires have never heard of green-bar reports).  But still, if it walks like a duck and quacks like a duck, it’s probably a duck.  Sorry, but your shiny new dashboard is really just a duck…er…a report!!

So is it the tool or the technology’s fault that we are simply Measuring Performance with our Analytics rather than Managing Performance?  Of course not!  The latest Analytic Applications offer some real alternatives to passive report reading – if only you understand what really makes them Analytics!

Analytics help you to focus on Performance Management (OK, EPM, CPM, BPM, PM – there, I said it…).  But there are expectations that come along with analytics for Performance Management.

1. True analytic apps don’t focus entirely on historic results.  They really focus on performance drivers.  Use regular reports to highlight the impact of certain drivers on performance over time.  Use analytics to manage that performance.
2. Develop KPI’s and Metrics that highlight how specific cost & performance levers drive the performance of your activities.
3. Remember that real Analytics are Dynamic and Interactive.  Expect that your analytic app will require some “care and feeding”.  The real value of an analytic app is in the hands of the analyst. If you are setting up an analytic app expecting to run the same way forever, you are setting up a report that can be tended by a clerical worker.  If the problem never changes or evolves or your business never changes or evolves, you need a reporting tool – not an analytic app. A number of SAP rapid-deployment solutions (such as Performance Analytics for Engineering, Construction & Operations) include SAP Business Objects Explorer to meet this demand.
4. Demand that your analytics permit what-if analysis.  This can be as simple as having a little dial or slider or input box built into your analytic app that shows what happens to my results if one of my cost drivers changes up or down X%.
5. Focus more on Predictive Analytics.  To me, it’s less about the algorithm than the result.  Find the WHY and then the WHAT NEXT.
6. It sounds trite…but it’s true…analytics must actively enable you to turn Insight into Action.

So in summary, how do you put Analysis back into Analytics?  It’s simple – and it is a process:

1. Start with an addressable Business Problem.
2. Generate a hypothesis as to how you can address the problem.  State a targeted outcome.
3. Use a true Analytic Application to iterate through your data and challenge your hypothesis:
   • Visualize your data.  Assess the validity of the data by visualizing your data. (This will also help you to identify outliers or extreme randomness in your data that may cause you to refine your hypothesis!)
   • Identify the performance drivers.
   • Manipulate the performance drivers to see how/if they impact results using a predictive analytic tool using sensitivity analysis.
4. Make a recommendation based on your results!  Change something.  Grow something. Reduce something. Add something.  For an analytic tool to have business value – it needs to be used to effect a positive change on your business.
5. Learn from your analysis.
6. Repeat.

The value of putting Analysis back into Analytics is tangible and real in terms of increased profits, reduced costs and improved efficiency.  The way to do it is to use analytic applications smartly to model and test hypotheses that address business problems.

Anything else is just reporting.

For more info about rapid-deployment solutions for analytics, check out my other blogs:

• Increasing the IQ of your Rapid Deployment Solution
• Keep Calm and Carry On
• A Key to Rapidly Solving the Puzzle of Project & Program Management – Take a Lifecycle Approach

Follow SAP Rapid Deployment Solutions on Twitter: @SAPRDS