GETTING ON THE SAME PAGE: A GLOSSARY OF COMMERCIAL PHARMACEUTICAL AND BIOTECH ANALYTICS TERMS
In our thought leadership articles, we use a lot of terms specific to our profession and our services. We thought we would take a moment to define some of these foundational terms and concepts in light of the work we do in commercial pharmaceutical and biotech analytics.
To start, Data Science is a field that works with and analyzes large amounts of data using advanced analytic techniques to provide meaningful insights for decision making. For example, we use data science to integrate the promotion, sales, and cost data to feed machine learning algorithms to determine the impact of, and responsiveness to, each promotion. And, most importantly, deliver an optimized recommendation for spend across the promotion portfolio. Another example would be to forecast TRx based on increased or reduced spending on individual promotions. Data science combines deep data manipulation and analytic expertise.
Advanced Analytics is the combination of the approaches, methods, and tools data scientists use to get more insights from the data. Unlike regular “analytics” that looks at history, advanced analytics uses simulation to predict what is likely to happen in the future. Advanced analytics are constantly evolving using new and innovative techniques and technologies.
Analytics as a Service is the practice of delivering advanced analytic results in a way that can easily be consumed by decision makers and where the expertise and platform required to generate those results are all taken care of by the service provider. Sentier’s Promotion Insights service (Marketing Mix Modeling as a Service) delivers comprehensive impact and optimization dashboards; behind the scenes our service includes the people, tools, and cloud based platform to deliver those results. An important aspect of Analytics as a Service is that buyers, based on their level of data science expertise, can also access the data and analytic tools for other purposes.
Artificial Intelligence is the concept that machines can execute tasks and solve problems better than people. Machine Learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed. At Sentier, our models are designed to adapt and improve over time as more data is added and more simulations are run.
Big Data is the idea that businesses are inundated with large volumes of disparate data on a day-to-day basis. It is often described in terms of the 3 V’s: volume, velocity, and variety. How much data you are getting, how fast it is coming in, and how many sources is it coming from? Aside from some of the patient level data and digital data that we receive, the volumes of promotion and sales data that we work with are very manageable. Given that our most advanced clients are running Marketing Mix Modeling four times a year, the velocity is manageable as well. The “Big Data” in our work is the variety. How do you combine non-personal promotion, personal promotion, and HCP vs. Consumer data into a format that machine learning can use. At Sentier, we have spent a lot of time addressing the variety challenge in marketing and sales data for analytics.
Cloud Computing is the delivery of different services, which have historically been “on premise,” through the Internet, like data storage, databases, data science tools, etc. We run our service on a cloud platform. The number one consideration when using cloud services is security. While Azure and AWS certainly have all the tools available, it takes significant expertise to ensure a cloud service is compliant with all the internal and external security requirements out there. We work with our partner Cazena to ensure the complete security of our cloud offering.
Data Integration and Aggregation are the processes by which data are brought together and conveyed in source, discovery, and fit-for-purpose formats. Source is the data as is from the provider. Discovery data is curated, combined, and quality assured data at the lowest level that can be used for a host of purposes. Fit for purpose data is generally aggregated and designed for a specific analytic effort, i.e. one table that contains all the promotion execution data by month and customer segment.
Data Engineering is the function responsible for delivering the three forms of data described above. Data engineers at Sentier take care of the whole data life cycle from source recommendation and acquisition to design and QA. Our Data Engineers are many times Data Scientists in their own right who understand how the data will be used for advanced analytics and work closely with those responsible for the analysis to generate the information that will lead to the best outcome.
The standard definition of Marketing Mix Modeling is that it is the analysis of multivariate marketing and sales data to estimate the impact of various promotional tactics on sales and then forecast the impact of future sets of tactics. But it is so much more. By using machine learning, building the model from the bottom up, and incorporating the full set of customer segments and campaign/promotion combinations, a host of other questions can be answered including:
How do I optimize the execution of this promotion?
What is the best way to interact with various high priority customer segments?
Which promotions have the most effect on each other? and many more…
Optimization is determining the most effective set of inputs to generate a desired outcome. The outcome could be profit or volume driven in the context of brand sales. Sentier has developed a disruptive method for optimization using response curves and corresponding MROI curves to solve for our clients desired outcomes.
Precision Marketing is a technique for breaking the market down into smaller, more specific blocks of customers (Market Segmentation) with unique preferences and ensuring that the marketing/sales interactions have the most meaning for each of them. At Sentier. we generate customer segment preference maps clearly showing the promotions/campaigns that segments are most responsive to. We follow that up with recommendations on sequencing and timing to ensure that promotion delivery is coordinated and set up to create the best possible customer experience.
(Data) Visualization is the usable graphical representation of insights and recommendations gleaned from advanced analytics (see our last post Don’t Chuck Numbers at People. Give Them Answers). Visualization is the bridge between advanced analytics results and executive decisions.
