Helping Businesses Succeed With Key Performance Indicators
Data Analytics is almost becoming a buzz word in that everyone knows data can help companies grow but understanding the best ways to incorporate analytics into your business is very important. Starting off on the right foot by identifying what to measure, why, and using S.M.A.R.T. goals can make or break an analytics team or project. As a CIO, CTO, or analytics team, understanding Key Performance Indicators (KPI) can help ensure your data journey is successful.
What is a Key Performance Indicator?
Key Performance Indicators are metrics regarding the core functions of your business. Utilizing these in analytics creates the ability to use data to make decisions but it doesn’t stop there. Good KPIs are easily measurable to determine if the changes made were successful.
What is a Good KPI?
The most important part of identifying KPIs is following the S.M.A.R.T. guidelines. It must be Specific, Measurable, Achievable, Relevant, and Time-Based. An example of a good KPI could include:
- Increase sales in commedy movies by 3% annually
- Decrease transportation costs by 2% annually
- Increase customer satisfaction by 6% quarterly
- Decrease negative reviews on social media by 3% monthly
Why are these all good KPI examples? Every one of them includes a specific, measurable, achievable, relevant, and time-based goal. They all have a target type of increase or decrease, a specific area to target, the measured amount is realistic, the purpose follows the strategic vision of the company, and includes a time period to measure and correct.
How Do You Measure a KPI?
The most accurate and simplest way to measure a KPI is using data analytics but that isn’t the only way. Before analytics was easily accessible, manual counting was heavily relied on to determine growth and change. Now, we have the ability to automate and calculate changes and growth in a visual, simple-to-use format. No matter what tools you measure, here are some helpful steps to get there
- Identify the methodology for measuring
- Which system(s) do you use?
- How can you obtain that data? Manually or programmatically?
- What data points are available to use for aggregates and filters?
- What is the reliability of the data?
- Identify requirements
- How will you display the data?
- How will users access the data?
- What tools do you have available to use for the project?
- What is the validity tolerance of the report?
- What is your audience?
- Who are the stakeholders?
- Develop a plan
- What is your timeline?
- What are the affects of your report?
- How critical are the decisions being made from the findings?
- How can the report help the consumers?
- Start the project
- Develop a “rough draft” report
- Measure data quality
- Find areas for improvements, inaccuracies, data integrity issues, and workflow complications to simplify the process
- Measure as changes are made
- Confirm requirements are met
- Present findings
This process isn’t easy but it’s critical to accurate and usable reports. When identifying the methodology for measuring, determining what is being measured is far more important than which system. If you have multiple systems, the data point may be included in both so you must identify if one system will limit potential growth of the metric or presents additional data integrity risks.
You work for a marketing firm. Your services include website development, social media management, and graphic design. A key performance indicator could include increase web projects by 4% quarterly. Over each quarter, you should continue to increase by 4% to meet the metric.
You have a number of systems you use including project management, scheduling, and documentation software. Instinctually you may think project management software would be the best option for measuring this KPI. However, if the KPI is “completed” or “identified” projects could be critical to which system you pick. Most likely, the project management software wouldn’t include any scheduled but not performed projects but the scheduling software may not include project tracking and couldn’t measure completed projects.
Once the definition of this KPI is defined concretely, the system can be decided. The next step is identifying how to measure and what the reliability is. Since this example is rather straightforward, there shouldn’t be any data integrity issues and depending on the services, technical resources may be able to gather the information programmatically. The only step in the identification phase is system and data identification. From here, the data is ready to be used.
There should be understanding regarding your data points prior to identifying requirements since this will help create guidelines. It would be hurtful to try to measure something that doesn’t exist or from an untrustworthy data point. Understanding who will be using this report can help determine the level of accuracy required which may prevent certain data points from being used. Knowing how the users will access the data may simplify or complicate the process. It may require custom scripts, software, or connections to be able to meet these requirements. Identifying which tools are available can create a consistent data collection and visualization process between all the KPIs and other reports.
Now that the KPI is clearly defined and the expectations are set, developing a plan can create a very fluid process allowing the analyst or person measuring to understand timelines, benefits, and risks of the projects. This should instill a cautious approach to identify and mitigate all risks. For example, a data point sometimes being populated incorrectly could negatively effect a metric causing unnecessary corrective action. It may be difficult to identify but several tools such as Talend, Power BI, and Tableau are built for this process to identify outliers, incorrect formatting, and data consistency.
The planning part is now complete and the project is kicked off. Starting by building or implementing tools is a controversial approach. However, if the tools or resources are available, this step allows the person measuring the KPI to understand what issues lie ahead of them. For example, if a date field is sometimes not entered correctly or names can be spelled wrong, you may not be able to properly represent the individual associated to the metric. If workflow or data flow correction must be made, the timeline or validity tolerance of the report may need to adjust. As the data is explored, new findings will present themselves which allows for process understanding and improvement. Process improvement can be one of the ways analytics pays for itself so don’t avoid doing so. As data quality issues are addressed, continue to measure the report until the tolerance is met. At this point, the report should be complete and ready for approval or presentation.
Typically, KPIs will be performed in a strategy meeting and will include several modalities of the business so don’t be surprised if this process must be performed multiple times. For presenting, the tools and methods for presenting have already been identified in your requirements step and should be able to be easily displayed for the audience. This may include Excel, Business Intelligence tools, or a PowerPoint depending on the audience and data maturity of the business. The presentation tools are much less important than presenting the findings but can help automate recurring measurements.
Pat yourself on the back for a job well done and don’t forget to re-measure when the KPI period approaches.
Have Questions? Let’s get in touch.
Helping businesses succeed is the reason I started this company. Data can truly help your company grow but it takes a lot to get there. That’s why we’re here to answer questions, help with implementation, data discovery, or report creation. Let’s talk!