Sales Analytics and Metrics that Matter

Avinash Kaushik wrote a great post on "useless web metrics" that discusses how an over abundance of data and easier access to it have contributed to an overwhelming explosion of reports and metrics. However, not all metrics are created equal. The recommendation is to question the validity and impact of all metrics and discard the ones that fail to produce a recommendation for an action. While Avinash focuses on Web Analytics, the same applies to most Business Intelligence or Reporting efforts, particularly "Sales Analytics".

Leading CRM/SFA platforms like salesforce.com enable sales organizations to report on their data easily and reliably. But not because we can create a lot of reports it means that we should. For many, extracting the right data to analyze and deriving true insight might seem like an art difficult to master but in reality it is more of a predictable science. Traditional reporting platforms require a skilled analyst to explore the data, identify trends and patterns, create hypothesis, test them, interpret the results and then create actionable recommendations. This process has 3 key weaknesses:

1) Inefficient. Even with intuitive point and click interfaces, creating and maintaining these reports is time consuming and it gets worse as data volumes increase.

2) Error prone. Without intelligent automation it is easy to accidentally miss certain conditions or critical dimensions from the analysis (we'll discuss Simpson's Paradox in a future post)

3) Unscalable. Effective quantitative analysis is a powerful skill in short demand, this means that the quality of the analysis might vary based on the analyst who prepared it. It also makes it difficult to handle employee turnover.

One way to overcome these challenges is to use better analysis techniques, like predictive analytics and data mining. A modern data mining platform excels at analyzing large amounts of data, automatically without getting tired. You can save hundreds of hours by letting the software identify the right data patterns that are impacting your business and you can eliminate the risk of overlooking critical information. Predictive models study your historical performance and can learn to predict future events. This codified knowledge can be shared within your organization and it will always produce consistent results and recommendations regardless of the person conducting the analysis.

Data mining and predictive analytics can filter the noise, save you time and produce better results, so you can take action and avoid having to ask "so what?".

Do you have any best practices to manage the growth of reports and metrics that you would like to share?