The Shudder of Analytical Expectations

I had read an interesting article by Kim Gaddy on 'Utility customer expectations, challenges and solutions' and two sentences, disconnected by a paragraph, had caught my eye.

"There is no denying the fact that customer expectations are on the rise and changing rapidly..... Meeting customer expectations will likely become more difficult over time.".

While this is true and usually leveraged as an argument for the necessity for analytics (Know Thy Customer), it is equally true for the consumers of analytics. Through the continued adoption and success in the marketplace, the expectations of analytics have also been increasing. In fact, it is not uncommon to find individuals spending equal amount of time convincing of the potential and managing the expectations of analytics.

Analytics is a productivity tool, either focusing on increasing revenue (top line) or improving efficiency (bottom line). However analytics is based on data, and it is this data exactly that drives the potential of performance. In other words, the question that should be asked is not only 'what is the accuracy of my analytical models' but 'what is the improvement delta from my current state'. The latter at times is forgotten due to the emphasis on the former.

Chasing a 95% accurate model right from the onset (while may be achievable, however most likely not) simply does not make business acumen. It is more realistic to assess - given the current state, process and data - what is achievable and to identify whether that makes a value proposition. Subsequently to proceed to take incremental steps in operationalising the achievable value.

Analytics is awesome - but it is not a one off exercise with winner takes all. Analytics is a continued process of incremental improvement, and data driving decision making.

"Have no fear of perfection - you'll never reach it." -- Salvador Dalí