I recently ran into this HBR article about why and how to tell compelling stories with data. HBR also had some interesting insights on storytelling and the use and shortcomings of analytics as well (Davenport, Tom. “10 kinds of stories to tell with data”, HBR, May 5, 2014 and Filbin, Bob and Jeff Bladt, “A data scientist’s real job: storytelling”, HBR, March 27, 2013).
It occurred to me quite quickly that I tell stories with data nearly every day. Most importantly, I tell data stories around the virtual campfires of internal customer prep meetings as well as customer workshops. My “not so great big book of everything” is one of them.
Ultimately, doesn’t every good horror story start with sentences like “once upon a time, a boy disappeared in the scary ouse…” (historical use case), “another boy looked for him in the basement as most monsters hide there but once he opened the door – always a bad idea – he never came out as well…” (repeat pattern) and “as it turns out, over the years the house has swallowed teenage boys with red bikes….” (analytical pattern).
Thinking about it now, I would say I am the Steven King of data quality – my trademark – and as such, every activity-based costing expert’s best friend.
It still surprises me every day that the IT-side of the house fully understands the data quality and access problems in the myriad of data sources available; external and internal, front and back office. But there is little understanding about the ramifications of setting up a product attribute incorrectly at design time and the IT infrastructure and productivity cost on the business side it generates once it gets integrated into a made-to-order build, goes into inventory, creates incorrect stocking locations, lost orders, expedited shipments, returns, overestimated re-orders, etc. Once the error gets tracked down via some eager business analyst, stock checks and propagated throughout the enterprise, multiple telephone, webex and physical meetings occurred resulting (as per one of our clients) in sometimes half-a-million dollars lost.
Meanwhile, marketing, sales, service operations, finance, etc. see the end results but either learned to live with it or is afraid of making too many waves without control over all related system aspects. When these get offered to the business via some data governance leadership role, they often back away as it consumes too much of their time and detracts from their primary role: to identify prospects, to educate prospects, to close deals, to ensure clients are satisfied, etc.
Ultimately, data suffers because of a lack of mission, e.g. stop excess customer onsite visits, reduce attrition, and complexity, e.g. too many source and destination applications and dependencies or costly, inflexible integration scenarios.
Cloud-based integration and workflow management sprinkled with data quality assurance are the key ingredients to establish long-term data integrity from cradle to grave.
So I tell the story of a pharmaceutical company grappling with a large portion of its portfolio falling off the patent cliff, paying enormous sums to its CROs for selecting trial sites, where close to half do not perform as intended due to competitive trial activities or subpar results due to recruiting the same patients over and over – most often men, as recently reported in the Atlantic.
Net-net is that data does lie. It lies all the time, not because it wants to but because we let it. The root cause is that we do not tell the story how data becomes a liar and how it disrupts our private and work lives. We only realize it once it gets stolen but by then the outcome is way too personal and way too late.
What are some of the everyday personal and business data horror stories you have heard about? Tell the tale – the bonfire is still warm.
The post Illusion Series – Episode VI: I Tell (Horror) Stories With Data, Do You? appeared first on The Informatica Blog - Perspectives for the Data Ready Enterprise.