Kathy Gibson reports from SAPInsider in Vienna – Sometimes innovation can come from the most unexpected places – but organisations might not stumble across it without the help of data analytics.
Jayne Landry, global vice-president and GM at SAP Business Intelligence, cites the example of a hospital group grappling with the mystery of why the cost for a specific surgical procedure varied so drastically from one operation to another, and eventually found the answer through a combination of analytics and human ingenuity.
Using sophisticated analytics, the organisation examined various different variables and seemed to find an answer when it turned out that certain surgeons were typically responsible for the more expensive procedures.
However, further investigation into the surgeons revealed that they all followed the same procedure and so that line of enquiry ended in a dead end.
More analytics were performed, and revealed that the additional costs could be attributed to a particular resin compound used to set bones, with the more expensive procedures using much more than the less expensive ones.
Running analytics again determined that the expensive procedures were all carried out in specific operating theatres, and so some human intervention was called for, and an inspection made of the theatres.
The culprit turned out to be the bigger bowls that were used in some theatres, prompting surgeons to inadvertently use more of the compound than they needed – and so causing additional costs.
This is the interesting thing about business intelligence (BI) and using self-correcting analytics algorithms, Landry says: “You don’t know what you don’t know.
“Using algorithms can help users with the outliers that might otherwise get ignored.”
Typically, when users want to identify anomalies or patterns in data, there’s a good chance they might miss something.
“But algorithms can surface key insights; and help to give users insights that they might not know.”
The human factor is still important, though, Landry says, pointing to the hospital example. “Companies still need to analyse the data and decide on what action to take. The tools can only provide you with a recommendation.
“But the more information you have, the more informed you will be.”
Algorithmic analysis also gives organisations what Landry calls “muscle memory”. Typically, when an organisation makes a particular decision, it’s based on information current or relevant at the time.
But down the line, it’s generally been forgotten what the information was that went into a decision and so it’s not always easy to recognise if the decision is still good.
If the system could remember what the information was back then, and compare it to the information available now, it would be easier to evaluate actions and decide if they are still relevant.
“Currently, we kind of don’t know why decisions were made,” Landry points out.
Collaboration systems are also important for decision-making, she adds. “This helps you to understand not only the data, but who was involved in making the decision. It’s more of that muscle memory.”