Social Media Data Analytics services (Facebook) helps organizations correlate the change in Facebook fans growth to Sales helping them forecast sales targets, optimize inventory levels, target social media marketing campaigns and plan marketing spends more effectively to achieve the desired results.
Market Equations helped a E-Commerce fashion retailer in the UK institutionalize Sales and Marketing Analytics by building a correlation model linking Facebook "likes" and "fans" growth to Sales. This helped the organization allocate their marketing spends effectively into channels that maximize returns, achieve sales targets and retain the most profitable customers.
- Determine if there is a correlation between Facebook "fans" growth and sales trends
- Build a predictive model that links Facebook fans growth to Sales to determine the increase in fans growth that would achieve sales targets
- Optimizing marketing spends to achieve targets
Sales data, website conversion data and Facebook fans data was provided by the client for a period of 12 months at a daily level. Data audit was conducted on the data to establish data sanity and structure data for further analysis. Data trends were studied further and abnormal spikes in sales caused by seasonal variations were not considered to ensure the results were not skewed by these events.
A correlation study was conducted and the results confirmed a strong correlation between Sales and Facebook fan growth:>
- The sales of a given week were highly correlated to the fan growth in the previous 3 weeks
- Fan growth in a week drives sales for three succeeding weeks after which the effect tapers off gradually
Building the mathematical model
The historical data review clearly revealed the correlation between Sales and Facebook fan growth which was positive but not linear. It also revealed that the relationship was different by Distribution Center (DC)
- As a result of these outcomes, multiple non linear models were developed between sales, fan growth and other factors including website conversion rates at a distribution center level
- Based on model statistics, models where predicted numbers followed the actual numbers closest in the validation datasets were picked and shared with the client
- Separate non linear models were developed for the time periods where behavioral patterns changed drastically around the "occasion" periods
The analytics of social media data was presented in a dashboard format to help the client enter various scenarios to arrive at a number of fans per month that would be required to generate "X" amount of sales. The models helped the client set realistic sales targets and helped them channel and schedule their marketing budgets effectively across platforms. The company was able to scale the business by more than 150% in 8 months and have received a fresh round of (Venture Capital) VC funding.