Blogger: Craig Roth
Over at the KnowledgeForward blog I wrote that I predict that in 2009 there will be an announcement that information overload or unnecessary interruptions cost U.S. workers $1 trillion. Basex has already released estimates of the cost of information overload being $900 billion and unnecessary interruptions being $650 billion. The figure is soft to me to begin with, but becomes even softer when it becomes a Rorsch test for info-stress. To the New York Times, it's the cost of unnecessary interruptions (well, the recovery time mostly) that are mostly of a mundane nature. To BusinessWeek it's the annual toll on U.S. productivity from interruptions.To Henry Blodget, it's the cost of checking email too often. To Read Write Web it's the cost to the nation of inbox overload and task switching.
Now, I strongly believe organizations can and should take concrete steps to help improve the efficiency of information workers by providing attentional technologies and capabilities that can pull important messages forward and push less important messages back from the worker's focus (see A Manifesto-free Definition of Attention Management). But I don't want to fall into the same trap as Knowledge Management which collapsed in on itself due to overstated estimates of its cost with the implication it could be recovered. I listed 5 reasons I find these estimates dubious (check out the blog entry for more detail on each):
1. Lumping in social interactions and distractions with interruptions
2. By counting all costs and no benefits (quote “total cost” instead of “net cost”)
3. By ignoring closed-loop analysis
4. By playing loose with the definition of “unnecessary”
5. By comparing against perfect short-term productivity instead of long-term sustainable productivity
But if there's a good cause at stake, what's the harm in crafting numbers that help to shock readers awake and spur action on a problem? The Economist ran an interesting story about how AIDS researchers at the World Health Organization and UNAIDS are being accused of using poor statistical methods that "distorted priorities for the treatment and prevention of the disease."
... the agencies spent many years overcounting the number of cases ... Dr Pisani cheerfully admits to being a doctor of the spin variety herself—she refers to the process as “beating up the news”. She absolves UNAIDS's researchers of any blame. They did their best to collect true numbers in difficult circumstances and with little money. But so as to rack the world's conscience, she wrote reports that put the worst possible complexion on those numbers.
As Joel Best, author of Stat-Spotting: A Field Guide to Identifying Dubious Data, said on NPR's Bob Edwards show on 1/10/09, a lot of statistical abuse is introduced by researchers in the social sciences who truly believe in their causes and want to draw attention to them. He questions statistics that support worthy causes, but are nevertheless misleading, overstated, or wrong.
Mr. Best is a sociology professor and takes special notice when statistics involve social causes. Regarding measurements on lost productivity, Mr. Best writes (Stat Spotting, p52-54)
In recent years, it has become very common to hear that this or that social problem costs America so much each year in lost productivity. These estimates ... typically involve billions of dollars.
The basic idea behind these claims is that social problems interfere with people's ability to do productive work. Lost productivity is the value of the work that could have been done if the employees had been able to concentrate on their jobs.
Now, fiddling with any of those numbers [that get multiplied by time and value to obtain the estimates] can lead to wildly different estimates of productivity losses. The particular choices made have everything to do with the resulting statistic.
It is these choices that can distort the information overload and unnecessary interruption cost figures, causing organizations to focus on the wrong problems. Do I choose a definition of "cost" that is net of benefits, or just all the costs alone? Who defines interruption (does it include social interactions and self-imposed interruptions such as distractions)? Who gets to determine whether an interruption is "unnecessary"? As most interruptions only benefit one party involved, do you find the net cost to the organization for the interruption or just total the costs for the person who was put out by the interruption? To make a statement on aggregate costs (e.g., to the U.S. economy, to an entire company), you need to make aggregate judgements about which interruptions count as "unnecessary", not aggregate individual, one-sided judgements.
I encourage any company that suspects info-stress may be occurring or that poor decisions may be made based on information overload to step back and consider their enterprise attention management environment. But be very careful about making a numerical estimate of the actual cost because it is interpreted as a judgement on potential savings and risks distorting the view of what should be done to address it.