Example #1. Modeling
An online retailer wanted to begin using forward-looking analytics (modeling, segmentation) to augment historical dashboarding and reporting.
- Created an analytical data set that could be used across future modeling initiatives
- Built upcoming 12-month gross revenue predictive models with associated gains tables, with separate models for new acquisitions, young customers, and customers with several years of history with the retailer
- Developed a Holiday mailer response model
Example #2. Customer Segmentation
An online retailer wanted to create a customer segmentation to guide channel leads in campaign and communication list selection.
We created a segmentation using demographic attributes, geographic data, and purchase behavior as follows:
- Gathered an analytical dataset specific to the segmentation requirements
- Grouped the various attributes into several dimensions, such as demographic, geographic, purchase type, purchase amount, etc.
- Built a simple model for each dimension so that each customer received a score for each dimension based on how their attributes in that dimension affected their predicted revenue for the upcoming year
- Used the dimensional scores as inputs to a clustering segmentation to create six actionable segments
Example #3. Funnel Reporting
A business-to-business company was seeking to build starter reports around their sales funnel to better understand how they could improve performance at each level of the funnel.
- Created weekly reporting to show the performance for that week at each funnel stage as well as QTD performance
- Developed quarter-end and year-end reporting with summary views to call out key trends across funnel stages
- Built a visual buyer journey report to show volumes and velocities at each funnel level by lead origin
Example #4. Campaign Design & Lift Analysis
A financial services company was seeking to create more structure around upcoming marketing campaigns. They sought to implement tests that would enable them to tailor creative copy and messaging to be more meaningful and to create reporting frameworks that would enable easy interpretation of results.
- Built test grids and helped define cell quantities using the process described below to provide meaningful test results
- Worked with internal teams to reach a compromise between strong campaign results in the present roll-outs and reliable test reads to strengthen future campaigns
- Developed a few key tools to auto-generate the most commonly requested campaign metrics and views without necessitating ad hoc requests for each view
Example #5. Survey Data Aggregation
A volunteer non-profit organization conducted an extensive survey of its volunteer members and was seeking to aggregate, organize, and present findings at an upcoming annual meeting.
The initial data set had many issues that needed to be handled, including missing values and inconsistencies in the scale and type of response across the survey questions.
- Created methodology for standardizing survey results and converting responses to a numeric scale when possible
- Built graphical and chart-based views to evaluate results
- Worked with internal team members to create summarization and commentary to be delivered at the annual meeting