Experimental Design and Analytics Framework to Support Growth Hacking Initiatives
Case Study: Data Science and Analytics
SITUATION & BUSINESS CHALLENGE
A large, multi-chain retailer of consumer electronics, appliances and computers sought better insight into its customer purchasing data to support radical revenue building strategies. The company wanted to leverage data science to experiment with new initiatives in sales and services and created a new division, a “growth office”, to support this effort. The growth office reported to the company’s Chief Strategy Officer and consisted of managers and personnel from across the organization (marketing, finance, strategy, HR, IT, operations and analytics teams). Initiatives designed by the growth offices included in-home advisers to offer electronics recommendations, home-security and health-monitoring systems, and an auto-replenishment program among others. They wanted to implement a “test and learn” approach to quickly assess which initiatives to drop and which to scale.
Strong analytics grounded in data science were needed to make these determinations. Due to the special skillset needed and an overwhelming workload, the company recognized a need for a dedicated leader and full project team. Without a leader to direct the use of insights, initiative leaders were manipulating data to make their programs appear more successful than they were, which ran counter to the objectives of the organization.
Realizing it needed a more robust approach to data science and analytics, the company evaluated three consulting firms, including a major global consulting firm, and chose AIM Consulting to lead the project.
From its Data & Analytics practice, AIM Consulting deployed a team of statisticians, data scientists, and analytics leaders to integrate into the enterprise’s analytics team and growth office.
At the project’s onset, the AIM team developed a process for clearly communicating with the growth office and its many initiative teams how they would be supported in developing and scaling their efforts. AIM then developed an analytics framework to specifically address different needs for metrics and analytics during the phases of the initiatives as they grew. The schema included three branches:
- Short-term success factors: This included setting KPIs to measure levels of success or failure for an initiative and to optimize the successful initiatives. Analysis included customer profiling and comparisons to non-initiative customers (control group), spending analysis to determine specific types of customer revenue, “tail” analysis for determining the existence of sufficient numbers to continue an initiative, and market basket analysis to determine whether initiative customers represented larger revenue than control group customers.
- Long-term success factors: This was designed to understand whether the retailer was having a better long-term relationship with the customer as a result of a initiative. Analyses included review of customer spending patterns, redemption values compared to predicted “customer lifetime value,” and propensity for existing initiative customers to participate in other initiatives.
- Using advanced analytics and machine learning to better predict how to scale: After an initiative proved itself in the short term and was starting to show long-term success, predictive modeling was used to show how the program could scale for the future. This analysis included customer segmenting and targeting for future sales, product recommendation engines, predictions for the next significant purchase, and exhaustive customer journey analyses.
Together, these metrics help the growth office determine whether each initiative is successful against its own criteria for success, how to optimize the succeeding ones, which initiatives achieve the long-term goal of making and cementing better customer relationships, and the potential benefits of scaling programs succeeding in both the short and long term. The AIM team was instrumental in demonstrating success (and failure) in each of the three areas, including suggesting pieces of analysis or models that would help leaders make decisions.
After laying the foundation, the AIM team worked with eight initiative teams in the first stage of the framework to help determine whether each of their programs was successful enough to continue. With this framework in place, the teams are furthering their work on a base of intelligent analytics, moving some initiatives into the long-term and scaling phases.
The benefits of the analytics framework to support the company’s growth hacking and data science ambition are significant and far reaching.
Initial analysis revealed a $5 million spike in incremental revenue from one initiative and $1 million from another, numbers that would not have been so easily discoverable before. In addition, the company saved significantly on operational expenses by not engaging the global consulting firm for the project.
“Test-and-learn” is now a key and respected part of the strategic growth process, with leadership and initiative leads depending on the knowledge and results of data science and analytics to help them make decisions.
AIM’s analytics framework is now tied directly to the company’s revenue and growth, enabling the company to allocate resources in the most productive ways, scale for greater gains, and explore further growth opportunities.