Churn Model for an e-commerce firm followed by score based pro-active retention campaigns
Problem in Hand
A B2B e-commerce firm was fearing loss of customers and wanted to incorporat a retention program.
Solutioning
We first helped our client define “attition” in their portfolio and then analyzed the data to understand what was the rate of attrition. We also analysed what kind of customers are we losing and when. Post that we developed a churn prediction model to predict the probability of attrition for each customer and a customer attractiveness score based on historic purchase behaviour. Taking into account both scores a monthly retention program was run to retain the customers. A research intiative was also undertaken among high attractiveness score customers to understand the reasons for attrition.
Impact
This initiative resulted in a drop of ~20% in churn rate and also an improvement of customer loyalty scores.
Developed a behavioral segmentation-based dealer strategy for a chemical manufacturer.
Problem in Hand
Our client was a leading manufacturer-distributor of chemicals across industries. With over 1500 distrbutors across the globe, it was difficult to have a visibility on the performance as well as running sales prograsms with them. They wanted to improve their dealer management program.
Solutioning
With an objective of maximizing the lifetime dealer relationship value of the network, we created a segmented dealer management strategy. A 360 degree view of the dealers was created covering their organization details, relation ship detsils, transaction details, payment details etc. It was augmented by derived KPIs like purchase frequency, purchase consistency, credit worthiness, new product trial scores etc.. The mart was then used to do a k-means based behavioural segmentation of the dealers and identify unique characterstics of each segment.
Impact
We were not only able to have a better visibility of the dealers across the globe. This also helped us drive segment specific sales programs such as new product launches were driven by one particular segments or stock clearance through another. This exercise also helped us design a data driven marketing budget allocation for the dealers later. This exercise also observed an incremental business growth of ~15% in the next six months.
Developed an invoice scoring mechanism based on probability of approval from insurance firms for a leading healthcare chain.
Problem in Hand
A leading hospital and pharmacy chain was struggling with cashflows due to their high dependence on insurance approvals and payments of their invoices. They wanted a mechanism to increase the approval rates of their invoices.
Solutioning
As a first step, we developed a regression based scoring model to get the probsbility of approval for the invoices. Based on the scores we prioritized the invoice being sent for approval which took care of the short term cash flow problem. We also analysed the drivers of approval and worked on each factor such as Physician scores etc. to drive the internal strategies for better approval rates.
Impact
The approvl rates increased by 5-7% in the next quarter to ease the cash flows immediately.
Developed a cross sell model-based sales campaign for oral hygiene products for a direct marketing firm.
Problem in Hand
A direct marketing firm was facing an issue with managing their growing sales partner network. They wanted to create a Partner lifecycle management program in an effort to maximize lietime value derived from their parner network.
Solutioning
We developed a 360 degree of the sales partners which included, their demogrphic details, onboarding details, network growth history, order history, network performance hisory, network spread etc. This was supplemented with derived metrices like network influence score, Partner agility score etc. Different data driven programs were developed to increase the quality of partner acquisition, increasing partner contribution, managing partner activity rate. As part of the partner contribution program, we developed a cross model for key product categories and rolled out campaigns for pushing the product category penetration among partner network.
Impact
Cross Sell campaign pushed the average category penetration from ~2.3 to ~3.5 in 6 months.
An earthmoving equipment manufacturer could reduce their procurement cost by 13% through data driven vendor management
Problem in Hand
Client wanted to revisit their procurement function as part of their larger initiative of improving profitability. Organization was procuring 3000+ items through 200+ registered vendors.
Solutioning
An ABC analysis was done to categorize the inventory on the basis of value, time to procure, lead times. Developed prictive maintenance models for for key machine to get predictability in key spareparts requirements. Developed a scorecard based vendor selection model to ensure uniformity in vendor selection process. Developed vendor engagement score, vendor fitment scores and vendor segments to maximize engagement of listed vendors and to ensure sufficieny of bids at any point. Aided the vendor negotiation team with 360 degree visibility of vendor relation in terms of financials, delivery performance and historic engagements to support a four tier data driven negotiation process.
Impact
The initiative put in place a data driven procurement process, higher visibility of vendor relations, segment wise vendor management strategy. This helped in bringing transpartency to the process. Helped improve vendor relations. In addition it improved delivery times, parts rejection rates and cost of procurement. Overall procurement cost were improved by ~13% adding to the bottomline.
A pharmacy chain identified the levers causing business losses despite revenue growth and could prioritize and address them
Problem in Hand
Client had seen a steep growth in number of stores through key acquisitions. Although the revenues grew as expected, client was struggling with mounting losses in the business due to mismanagement at individual store level . The challenge was to identify the issues and provide a sustainable solution to improve profitability
Solutioning
Initial analysis suggested that the issues were not same in every outlet and hence we needed to put in a stronger monitoring mechanism . More than 40 KPIs and their drivers were identified across Inventory Management, Sales, Logistics, recivables etc. Created dashboards to track these KPIs and their drivers for better monitoring. Automated numerous manual calculations to report various derived variables such as Sales Incentives, Average Lead time by vendor and SKUs, Vendor fulfillment rates etc.
Impact
Helped client with identification and definition of 50+ initiatives to create and implement turnaround strategy Led to an increase in operational efficiency, timely escalation of bottlenecks and their resolution Met and exceeded success parameters for turnaround strategy implementation over 18 months
Understanding the demand pattterns across outlets for Quick Service restaurants and implementing a dynamic resource allocation model helped our client rationalizee the strength of their frontline staff. We could reduce the team size by ~ 17% without affecting the peformance and service standards
Problem in Hand
A leading Quick service restaurant was highly overstaffed in anticipation of rush hours across its 100+ outlets in the city. This was having an impact on the profitability inspite of having a steady revenue growth. The existing resource allocation molel was dated and was based on manual inspection of past sales data coupled with the team experience. The existing model was failing with growing business outlets and volume..
Solutioning
We analyzed the sales trends waiting times for dine in customers across outlets and across different days and time intervals to understand the peak loads and waiting time for walk-in customers. A forecasting model was developed to predict footfalls and order volumes. Accordingly a new resource allocator was implemented to adopt a dynamic resourcing model.
Impact
We were able to rotate the reallocate resources as per demand projections during week days and sometime within different time intervals in the same day. This helped us reduce the employee strength by ~17% and still able to manage the variable demand/ footfalls across outlets.
