When I get a chance to speak to organizations about their analytics implementations, they’re often overwhelmed by the variety of attributes they can measure about their business. This scenario usually ends in an “anything and everything” approach, leaving businesses with the burden of too much information, and not enough resources to actually do anything about it. To avoid falling into this trap, organizations need to be more purposeful and selective about what they do and do not measure when it comes to analytics.
When implementing a new analytics strategy, organizations need to consider what will be the best strategic use of the data and contribute the most value to the business. In my experience, organizations should focus on measuring the following:
• Data that is relatively easy to collect.
• Metrics that provide direct or indirect insight into the performance of your business.
• Metrics that you have the ability to act on and influence a change.
Given the risk of wasted time and resources, it is also important to consider some things organizations should never measure.
First, organizations should not measure things that are easy to observe yet difficult to measure. A good example of this is determining why a person has decided to leave an organization. Instead of using valuable time and energy to review a wide range of data from surveys or other analytics that could explain why someone chose to leave an organization, you could simply ask them and consider observed behaviors and actions of the employee. By going straight to the source, organizations will save invaluable resources getting this information. Gathered feedback can then be used to make organizational changes and improvements.
Second, avoid the temptation of creating complex metrics and ratios. Many organizations get caught up in building out complex sets of metrics that are challenging to collect, interpret and act upon. Organizations should always consider the cost benefits and decide whether the costs in data preparation, monitoring, and education will actually outweigh the benefits. While BI is a powerful tool, there are some situations that simply don´t need the full force of an analytics solution. Consider a simple example of forecasting sales. While you could develop detailed model based on a wide array of data inputs, it is more efficient and effective to take one key input, your sales pipeline, and use it as a proxy for forecasted sales, with some simple assumptions based on the expected conversion and timing of sales from that the pipeline.
Third, organizations should not use analytics to measure things they have no intention of addressing. Employee and customer satisfaction surveys, for instance, are a common method organizations use to find out what people want from their employer or provider. Yet, organizations should not conduct these types of surveys if they are not willing and ready to put resources into addressing the issues raised. By asking for feedback, organizations are setting an expectation in the minds of their employees and customers that their comments will be addressed and potentially setting themselves up for disappointment. If there is no potential for action, you’re better off not measuring in the first place.
While these points may seem obvious, many organizations continue to struggle with how to approach analytics. It is also important for organizations to realize that there is an inherent cost in measuring. Everything that is reported and tracked creates additional responsibility and a potential burden on organizations to correct and maintain. Determining what to measure from the very beginning will set organizations up to get the most of out their business information.
Another important factor for companies to consider is how they are using the data that they gather. They might measure the most relevant aspects of their business, but without the capabilities to interpret, understand and put the data into the context of the business, even the best data will be useless. Giving data context and answering the “why” of the data, will enable companies to make better business decisions and ensure that good data does not go to waste. However, usually the BI or analytics teams spend the majority of their time preparing data and creating visualizations. While discovering the trends and patterns hiding beneath all that data is essential, by not interpreting the patterns analysts make it difficult for business leaders to do anything with just the raw numbers. Ultimately, they don’t see the true value in BI, as the data is not being presented with the context and understanding needed to truly make it actionable.
Luckily, recent advances in automation are making it possible for organizations to quickly find patterns and outliers minus all the heavy lifting of designing and conducting manual analysis. Instead, teams can refocus their efforts on interpreting and giving context to the data which is essential when it comes to adding value to the business. An experienced analytics team can see a pattern or anomaly and determine what it means, and from there, determine the gaps, possible next steps, and what changes need to be made. A machine can’t do that, yet.
While the challenges are numerous, the most successful data analytic strategy involves finding relevant data to measure and analyze, using innovative technologies to reduce the effort and resources needed to execute, and finally expending time and resources to give the data context and attain maximum value from within the decision-making process.