Measuring the impact of Covid on OPIODS Marketing
Measuring the impact of Covid on OPIODS Marketing https://datazymes.com/wp-content/uploads/2020/12/CaseStudy_Marketing-Mix.jpg 874 600 Vivek g https://secure.gravatar.com/avatar/927c001dd3a13b6e015a946097c9aa5c?s=96&d=mm&r=gMarket Mix Modeling to understand the impact of Covid-19 for a Long Acting Opiod.
Challenge
As the Opioids market is booming worldwide, our client, one of the top pharmaceutical companies, wanted to achieve brand’s strategic objectives of becoming the #1 LAO (Long Acting Opioid) in the market. One of the key prospects for the company was to maintain the consistent promotion presence to various HCP segments, with minimum budget allocation and maximum impact on ROI. An additional challenge was Sales Reps were not able to target HCPs with regular frequency because of COVID. Our clients have adapted to the new realities and wanted us to find ways to be more effective with their non-personal promotion (NPP). NPP, in theory, can be more effective method of reaching and engaging HCPs as it allows companies to connect with HCPs on cost-effective channels, leveraging tactics that seamlessly integrate within a prescriber’s workflow. DataZymes took up the challenging project to explore the impact of COVID-19 through a Marketing Mix Modelling exercise.
Approach
STEP 1
Data Preparation and Insights
A key challenge in most analytical projects is getting accurate and prepared data. DataZymes differentiates itself in that our robust data framework ensures a methodical approach to preparing data. Utilizing the framework helped clearly visualize how the reach and frequency of various channels were affected because of COVID. As early as the data exploration phase, DataZymes was able to share some valuable insights with the client on performance of various channels. It helped client understand the sub-optimal tactics made in the past and helped realign tactics for the future campaigns.
STEP 2
Meaningful Segmentation
The budget for the next fiscal was restricted and it was important to make sure that the right Physicians were being targeted. Employing various segmentation approaches helped identify the most optimal segment. The segment created using K-Means clustering based on Decile, HCP potential, Affinity preference and Affordability was found to be most optimal
STEP 3
Modelling
Timeseries models with different carryover and Ad-stock factors with segments as the class variable were executed. Controlling the impact of COVID and seasonality in the data, was accomplished using various flags. After multiple iterative modeling, one model that was statistically robust and tied well to the business context was finalized
STEP 4
Optimization and results
The impact and ROI at national level were calculated, as a recommendation to the brand team. The response curves for each channel and segment were created to examine the optimal and historical promotion. Additionally, pathways interaction between the Consumer NPP channels and reattributed the sales were documented and presented. Finally, optimization scenarios based on custom budget constraints of promotional channels were created to be presented to Brand Teams
Solution
DataZymes began by consolidating inputs from various data sources. Multiple data sources were used for each tactic and channel. After the data health check was done for completeness, accuracy, and usability of available input, summarizing the data to identify any trend, reach and frequency, customer segmentation, etc. was performed. DataZymes deployed multiple clustering methods to find the most unique segment groups. Determined the appropriate modelling approach (time series regression model) for assessing the promotion response at different granularities. Further, model fit statistics, accuracy and other measures were considered for selecting the best model
