Several years ago, I wrote a newsletter article called "Adding Value" Means "Improving Information". Its premise was this – moving up each level in the hierarchy from Data to Wisdom involves more summarization, abstraction, or processing from a previous level.
Levels of Value
Moving up a level adds value – each move requires changing and improving the context and usefulness of the information, going from raw data to something increasingly valuable – as illustrated in the following examples:
|4 – Wisdom||Strategic Plan||Interpretation|
|3 – Knowledge||Variance Explanation||Executive Summary|
|2 – Information||General Ledger||Report|
|1 – Data||Transactions||Isolated Facts|
I've recently seen these same ideas play out in two client projects – one a data warehouse and executive dashboard, and the other on inventory reduction.
Data Warehouse and Executive Dashboard
At its lowest level, the data warehouse and executive dashboard for our property development and leasing client was merely a collection of a large number of facts – transactions and account balances from the ledger, clauses abstracted from hundreds of lease documents, schematics of the properties, and information from each property's manager and tenants.
A project like this requires an unprecedented level of consistency among the various data elements. For example, a property might be coded "xxx-001" in one system" and "xxx-01" or "xxx 001" in other systems. ALL the incoming items for the property had to have the "xxx-001" code – or be translatable into the "xxx-001" format 100% of the time. What happened if the code were "xxx-00a"? It would be kicked out as an error, and then researched and resolved. What happened if there were no property code? The same thing – it would be held up until the problem was fixed.
The benefit of this level of consistency? Suddenly, comparing information across various properties became much easier. Complex analyses which had been too difficult to develop using typical reporting writing tools became practical. And questions from the users? They became more sophisticated, looking for additional clues – exceptions and trends – among the data.
Visual Trends & Exceptions
Since most people have a hard time seeing trends or exceptions among tables of data, a large number of charts were developed, making it easier to spot issues and opportunities. Once something is spotted, it is easy to start from the high-level graphical presentation and drill-down to find the root cause.
All of the charts and tables were presented to users via an executive dashboard – and only the information relevant to that person's specific responsibilities was shown. The dashboard presented an unprecedented amount of knowledge to personnel of all levels. Once people became familiar with the initial dashboard content, they requested more forward-looking information be added and presented, which we did.
What did the users – especially the executives and key managers – do with all this knowledge? They changed the way they did detailed monthly reviews – looking at a property's historical performance, its performance against its peers, its tenants' performances in other properties and across their peer groups, and projected future performance – of selected properties. Our client moved from a series of manually-compiled paper-based reports to using the dashboard almost exclusively. In short, these detailed reviews now require very little preparation effort, and capital expenditures and strategic decisions are now made more easily and confidently.
"Inventory" means something different for product-based companies versus services companies. For the former, inventory consists of the goods available for sale, while inventory for the latter reflects unbilled amounts for services already performed. Our client in the medical services industry began a fiscal year with a manageable – though large – inventory of unbilled services. For a variety of reasons, this inventory grew eight-fold in ten months. By that time, the inventory was both too large and unmanageable – it was like a boat taking on water more quickly than it could be bailed out.
This excess inventory was attacked in two ways. First, a large team was assembled of both internal and temporary personnel, who devoted many hours over three months to work down the inventory – in other words, to bail water faster. This team was successful, because the inventory dropped 35% during this period.
The problem? The inventory reduction had plateaued – inventory settled at a new level lower than its previous peak, but much higher than its desired level.
Technology & Process Improvements
The second attack was then launched. Initially, we developed and deployed technology and process improvements, sending items research upstream to the medical technicians more familiar with the circumstances around the unbilled services. Not only were the technicians able to research and resolve the exceptions more quickly than headquarters personnel – they also began to see why the errors occurred in the first place.
We are currently developing and readying for deployment the next salvo of technology and process improvements, all aimed at reducing error creation. We know certain procedures used by the technicians require duplicated effort and introduce the possibility of inconsistent information across systems – our task is to develop software to simplify and error-proof these procedures.
How did we find the errors to be addressed?
- Some were well-known among client personnel – these were the easy ones.
- Others were spotted when we were asked by managers at our client to look at some unusual errors.
As an example of the latter, our associate noticed one thing a group of errors had in common – they were all billed to Tricare, the health care program for active and retired military personnel and their families. Another commonality – virtually all the errors occurred in two offices, and both were near large military bases. Armed with these two insights, it was then easy to determine the root cause, address it in the exceptions at hand, and incorporate these insights into the software being developed.
Where are the "levels" in this second case?
- The first prong of attack – the large error cleanup project – addressed the problem at the first two levels, the inventory items themselves (the Data level), and reports or queries of the items to be worked (the Information level).
- The second prong of attack – the error prevention project – addressed the third and fourth levels, in addition to the first two. Being able to see patterns and correlations among the data was the Knowledge level, and analyzing and determining root causes was the fourth and highest level, Wisdom.
In the example I described, we were fortunate to spot the correlation between Tricare and two locations, quickly leading to identifying the root cause. Neither we nor our client prefer to rely on good fortune to identify correlations, so that led us to another recommendation – build a data warehouse of the unbilled inventory items, and then use data mining algorithms to search for patterns.
Data Mining Algorithms: Correlation or Causation?
An important note – "pattern" and "correlation" do not guarantee "causation." Two items can be statistically correlated, yet neither causes the other. As a purely hypothetical example – perhaps a declining percentage of women in the US wearing dresses in the workplace tracks closely with the annual percentage increase of average temperature in New Zealand. Pattern? Yes. Causation? Hardly.
Data mining algorithms can identify patterns and correlations more quickly and more thoroughly than people, yet the items found still require a person to review the results and assess whether a cause-and-effect relationship exists.
What's the Value?
Software can help with the first three levels – Data, Information, and Knowledge – by processing, summarizing, and analyzing large sets of items. Nonetheless, the final jump to the Wisdom level still remains a distinctly human endeavor.
What is the value associated with each level of the hierarchy? Admittedly, that's hard to quantify. Yet attempts can be made.
Reduce Monthly Meeting Prep Time and Improve Decision-Making
In the first project, the reduced preparation time for executives and managers provides benefits every month, on an ongoing basis. Better information via the executive dashboard does result in better decision-making, so an issue or opportunity may be spotted sooner, while it is still small or new – the benefits could be lower repair costs, or more profitable lease renewals.
Stop Errors from Occurring
In the second project, the software development costs to reduce errors causing unbilled inventory, as well as the cost of designing and developing the data warehouse and data mining capabilities, are estimated to be LESS than the entire one-time cleanup project. Improved processes embedded in software and improved data access and analysis provide a recurring benefit – error elimination – versus a cleanup project yielding a one-time benefit, error remediation.
The Value of Better Information?
While the value may seem tenuous or not appear obvious, at least at first, the value becomes much more apparent over time. Indeed, information's ultimate value can eclipse the higher costs or lost opportunities of the "do nothing" alternative – especially since a one-time investment typically provides recurring benefits.
If you could make a single investment and buy a one-year payback period and large ongoing annual benefits, would you do it? Hopefully so, since that's a smart decision by any standard.
Todd L. Herman