Analysis is critical to running a modern enterprise, and decisions should be backed by data. It helps teams understand what happened in the past, make informed predictions about the future, and evaluate alternative options
A well-executed analysis requires developing a sound methodology, processing raw data, and testing multiple hypotheses, all of which are time-consuming. When you don't have the internal bandwidth to run an analysis, we can help
The types of analysis we commonly perform for our clients include:
- Retrospective analysis:
Delving into historical data to understand what happened, and identify the drivers of an issue. This is most commonly triggered by a deterioration in some KPIs - Predictive analytics:
Applying statistical analysis on historical data to determine the likelihood of something happening in the future, for example a customer renewing its contract - Optimization:
Determining the structure and rate of business levers, such as pricing, discount, and commission. Historical analysis is frequently combined with competitor benchmarking to determine the final output - Quantification:
Assessing and isolating the impact of specific business initiatives on the P&L, or on some KPIs. This can be either forward-looking potential impact, or historical-looking realized impact - Modeling:
Creating Excel-based models and tools that allow you to play with different inputs to see the outputs. This is usually for on-going use by your teams in their workflow
Our typical process to run an analysis consists of six steps. The level of effort of each step varies significantly project-to-project
- Understanding the business objectives of the analysis from you and your teams. Every analysis should have a clear purpose that supports a goal
- Developing the methodology and formulating initial hypotheses. We leverage our experience and input from your team for this step
- Collecting the data to be used in the analysis, generally internal but some analysis can also require external data
- Executing the analysis, which usually consists of multiple sub analyses
- Validating the results using both our own experiences as well as inputs from you and your teams
- Review the results with you and your teams, and iterating the analysis based on feedback
For executing analysis we normally use Excel, SQL, R, Crystal Ball, and Azure Machine Learning. For communicating output we use PowerPoint and Excel
Upon project completion, we provide your team with the analysis files and instructions for updating them with new data. If needed, we will also be available to modify or refresh the analysis on a time-and-material basis