Whizzing round the Olympic velodrome last weekend with my family was great fun, if not slightly terrifying as you go up the banking. Not surprisingly data was the last thing on my mind! But at the end we received certificates which were all misnamed. This is pertinent for a couple of reasons. Firstly, it shows there is a manual interface in place. Somebody had copied our names from a screen (our online booking), or printout onto a piece of paper and passed that paper to someone else to create the certificates on another system. That interface was obviously seriously error prone. Secondly my name is John Davis. It is quite common (I blame my parents), and most clerical workers would know that it could be John or Jon, and Davis might be Davies. My son, Guy Davis, became Gille Davies.
So the great experience was tainted by a certificate that was wrong for every single member of the family. No one had thought through the consequences of the data quality issues and had taken a bit more care. We won't be framing it and putting it on our wall, or tweeting a picture of the certificate or sharing on Facebook - losing the Velodrome valuable free marketing. These issues with missing interfaces and poor data quality are the same ones that I came across when I started working in computing and business analysis in the 1980s. Things may have improved for a while, but now that companies have some applications in the cloud we have returned to a situation of many silos, and often some imperfect interfaces between them.
A focus on data quality is an issue that will run on. This is partly an issue with supervision and the reinforcement of the importance of getting data correct before you can rely on it to drive your business (It needs that great MBA stalwart, "Senior Management Commitment"). This will get no easier as the workforce is infiltrated by a younger generation who are accustomed to txting and spreading their focus across a number of apps on a number of devices, with a commensurate loss of attention to detail (am I sounding old now?). Nevertheless we can do better, taking out scribbled manual interfaces by calling APIs, using our systems to do more validation and checking at the point of original data capture, and reconciling data between silos and against other sources. Not only is this more efficient, companies get happy customers as it shows they care about the little things - our names, and who we are!
If you need help with your processes, data, or interfaces, then give me a call. I'm the red and black blur.
John Davis
4 Nov 14.