James Taylor, one of the IT analyst thought leaders of the broadening scope of business intelligence, has written a thoughtful piece titled <a href=http://jtonedm.com/2009/07/31/analytics-simplify-data-to-amplify-its-value-2" target=_new>Analytics simplify data to amplify its value. In his article James states that he likes his article’s title as a succinct description of analytics. He states:
(Simplifying data to amplify its value) always struck me as going to the core of analytics – the power of analytics to turn huge volumes of data into a much smaller amount of information and insight. … In every case the analytics are simplifying the data (a picture, a graph, an equation not thousands of rows of data) and yet amplifying its value by showing a data consumer what the data means. … IT people need to educate themselves on the role of these different kinds of analytics and their potential. … . Your systems store and manage data so something that makes that data more valuable makes your systems more valuable.
There is some danger when people talk about the keep-it-simple-stupid KISS rule. They forget its corollary LOVE rule – leave-out-virtually-everything. Too simple may be insufficient to draw correct conclusions. James understands this balance, and he has been an advocate of good decision making regardless of all the buzzwords the IT community may use. Many organizations are drowning in data but often starving for information. I’m with James. Performance management methodologies imbedded with analytics of all flavors unleashes the potential power for good decisions buried in data from transactional systems.
For KPIs to influence performance they must satisfy a number of criteria:
KPIs must be derived from an operational business model; set up using the Dupont analysis method so the changes in ROFE are linked to changes in operational KPIs.
The KPI model must reflect the working structure of the organisation. People need to be able to see the effect of a different decsision or work method on the departmental output KPI.
A KPI, by definition reduces the clutter of data, and provides real information.
When a KPI is tracked graphically real information is provided.
Performance improvement is free, when these criteria are used in the information provided to the people. People will naturally follow the path that optimises their results.
In my lexicon a KPI is an operating ratio that has high leverage on ROFE.
For more on these thoughts, readers could check out www.bizlearn.biz.
- I have visited your website and generally like your thinking. The BSC community needs more like you. I am not sure that I fully agree that a strategy map’s KPIs must be linked to the Dupont decomposition model of financial statements, but you have made me think that may be an appropriate “anchor” of some kind. (“Green and sustainability” environmental KPIs should reside too.)
The use of the Dupont disaggregation technique forces real rigour into the modeling in that every aspect of operational revenue and cost must be included in the model. The search for drivers of revenue or cost is the vital thing and I find it important to insist that the relationship is defined by a specific algorithm in the model. We used to call them “plumbing charts”: there must be no leaks.
There can only be one accounting system in a business, and I believe it should enable many views of the accounting data. If it does, it will feed data to a model that can deal with product segments, market segments, production, or service, even contracting issues.
I do not pretend that only operational KPIs ($/quantity ratios) are important.
Quality, service and environmental KPIs are also vital. The difficulty is to define an algorithm that enables them to be fitted into a simple business model that people can understand at the coal face.
Some years ago I addressed the service and quality KPI issue for a service business, and we found that a target such as “85% of jobs completed within 48 hours of pickup”, was important in building market share. Customer complaints were low when the target was achived, and rose steeply when it was not. We established this by correlation, and it seems we go it right because the business was very successful. We were never able to devise an algorithm to define the relationship.
I suspect the way to deal with green and sustainability KPIs is to use a probability based modeling technique such as @Risk.
For the moment the power of the Dupont based KPI model is its ability to demystify the accounting statements and provide a visual display that is meaningful to business operators. It is a superb teaching tool. Once the process is understood it is never forgotten.
If you look at the archived bizlearn newsletters you will find a number on this topic. There are also whitepapers as a free download that explain it to members.