A few months ago I took the picture below on my iPhone. It was a piece of crumpled paper which had been scribbled on by my children, probably trodden on by the dog and left on my stairs on top of a pile of other rubbish.
I took the picture because it looked like an eerie face or a mask. Look carefully and you’ll see the shape of the chin, a position of two eyes, a nose and even some lips. It is quite weird and spooky, but actually it’s just a load of folds and crumples on a piece of scrap paper.
However if I was a psychic, believed in UFO abductions or was trying to sell you a book on unexplained phenomena I might suggest supernatural forces are at play.
It reminded me that you can usually try and derive some kind of meaning out of anything if you look hard enough.
As digital professionals, I think this is something we always need to bear in mind. Often we are trying to (and are expected to) present output that is objective, scientific and “data-driven”, with a presumption that if it is something which is derived from metrics it will be bias-free.
Examples of this output might be:
- Requirements for new digital capability that has been informed by data and research
- Metrics which infer user behaviour and therefore influence design or content
- KPIs which reflect success
But numbers are open to interpretation and carry our preconceptions, beliefs and biases, often unintentionally, perhaps sometimes intentionally. We leave numbers out, write our own analysis or present data visualizations which emphasize one piece of data over another.
Do any of the following factors ever influence your interpretation of numbers, the presentation of numbers or any output which is presented as being data-driven?
- The need to show some sort of insight, because people want insight
- Covering your own back
- Convenience and tidiness
- Simplification for your audience
- Avoiding saying something difficult
- Avoiding contradiction and untidiness
- Supporting current policy and argument
- Trying to sell an idea because you want to work on that idea
People accept that some are very selective in the data they present, for example in marketing. But metrics and data for internal projects tend to be presented as fact.
If you can avoid these biases then in the long run your metrics and reporting will be more meaningful, have more impact and carry more weight with your colleagues.