The four uses of data (big or small)

Describing. Diagnosing. Predicting. Prescribing. These are the four main uses of data. As a consultant you need to understand the difference.
- Describing is the use of data to set out what has happened – e.g. a website received 1,000 visitors in the first week of May this year
- Diagnosing is the use of data to explain why something happened – e.g. 1,000 visitors came to the website on that week because an email was sent to 10,000 customers
- Predicting is the use of data to define what will happen – e.g. if we send an email to 20,000 customers we will get 2,000 visitors to the website
- Prescribing is the use of data to define what will be done – e.g. we will send a weekly email
Some further thoughts on the uses of data
- Not all data is neatly structured – Data may also by quantitative and not necessarily structured neatly – an example would be interview notes taken during stakeholder meetings
- Correlation and causation are not the same thing – try to be scientific in what you do
- Provide comparison – a number on its own is of little use; provide meaningful and logical data points to compare against
- Simplify your explanations – visualise, round numbers, arrange, tabulate, chart and do whatever it takes to help someone understand the data (Jon Moon’s book Impact is a great way to do this)
- Value is found in prescribing and predicting – this is where actions exist; describing and diagnosing should be hygiene factors, i.e. things that are just correct
- Not just a technology problem – technology is required, but more important is the ability to interpret and model data
- 10% of the budget on tools, 90% on analysis – Avinash Kaushik said it
Let me acknowledge Harvard Business Review’s article on Smart, Connected Products for this article: