What Pharma Marketers Need to Know About Artificial Intelligence (AI)

Even though we don’t think a lot about it, everything in our lives is driven by data. We make decisions about how, or even if, we should market a drug based on clinical data. We use historical data to make assumptions in our business and personal lives. We even use data when deciding what to have for dinner. If three-fourths of your family dislikes fish, it probably won’t be on the menu.


Business intelligence and analytics have been analysing data for decades. But with recent advancements in technology, an immense quantity of data is being produced at a vertiginous speed. Now we have big data, and with it comes the potential to transform businesses.Doug Laney from Gartner predicts that:
    •    By 2020, information will be used to reinvent, digitalize or eliminate 80% of business processes and products from a decade earlier


    •    By 2017, more than 30% of enterprise access to broadly based big data will be via intermediary data broker services, serving context to business decisions
    •    By 2017, more than 20% of customer-facing analytic deployments will provide product tracking information leveraging the Internet of Things

It’s no surprise, then, that boards and CEOs of pharma companies are rushing to get big data on their strategic agenda so they can grow, innovate and compete. However, when faced with big data, we found our old ways of analyzing data were insufficient.

Big data does not fit in Excel, and linear approaches —meaning ROI, correlations, multivariate regression analysis, promotional response curves, marketing mix modeling and multivariate statistics—tend to deal with aggregated data. They are simply are not granular enough to really get under the hood of the data. Therefore, they cannot extract the much needed critical information that will transform our results, or even tell us about complex variable interactions. Simply put, they cannot give the depth of information needed in today’s complex pharma environment.

That’s the dilemma facing pharma leadership. There is more data than ever before available to drive intelligent decisions. But collecting, mining, and analyzing that data is more challenging than ever. Pharmaceutical executives are seeking transformational ways to unlock value from their data.

Enter Artificial Intelligence
Artificial intelligence is not new. I was using AI in academia in neural network research on the brain more than 2 decades ago. However, advances in AI, and the various types of AI, such as machine learning, are happening at lightening pace. In fact, the machine learning algorithms Eularis are using today did not even exist 3 years ago.
On top of the speed of advances, AI is no longer locked away in Universities. It is now being applied to business problems. In today’s complex and challenging pharmaceutical world, pharma marketers need to utilize these complex data sets and complex algorithms to create new models that drive real world business outcomes, decrease costs, and improve value from the customers’ perspective.

Artificial Intelligence has the potential to transform companies. Think of it simply as a branch of mathematics, designed for a world of big data, to solve real problems. Pharma executives who want to get the most out of their companies’ data should understand what it is, what it can do, and what to watch out for when using it. Integrating artificial intelligence based analytics with your offering can provide substantial benefits including, among other things:

Understanding of optimal strategic direction
    •    Enhanced value proposition identification
    •    Enhanced competitor differentiation
    •    Optimal resource and budget allocation for maximum market share gain, revenue and profit
    •    Ability to see which segments will offer the most value to your brand as well as what levers to pull to maximize the growth from them
    •    Ability to customize sales and marketing messaging for individuals for what they need for greater customer engagement at that moment in time
    •    Automation of sales and marketing messages and channels by individual

The aim of what we have been doing in AI is to simplify the pharmaceutical executive’s ability to process large volumes of doctor and patient data and derive accurate and consistent findings resulting in improved patient and doctor outcomes and real-world financial results.

The unique machine learning powered platforms available, are able to identify “the wood from the trees” for both marketers and C suite executives. It can determine the best strategies as well as individual tactics for a brand—whether it is just emerging, growing, defending or failing—a brand portfolio, or even a company.
But do you know how best to do it?

Implementing strong AI based analytics does not have to be a daunting task. Eularis have been implementing Artificial Intelligence approaches to solve pharma strategic issues for some time now and we have the process solved.

1. Plan and Discover
Define your strategic needs and understand what challenges you are aiming to solve and what objectives you are aiming to meet. Then examine the data you have with regards to whether it contains what is required to meet your goals. Sometimes you will have all that is needed, but it may need to be restructured, and other times you may need to add additional external data sources. Then plan what approaches are available from the data existing, to meet your needs and provide business value. There are many options in what you can do but these range from relatively inexpensive to seriously extremely expensive depending on the amount of work involved. It is best to start relatively smaller with a discrete number of data sources and a specific objective with measurable outcomes. Then, once you have achieved success with this, slowly add more data sources and more projects to eventually build up to a large scale. The great thing about these projects is that they are all scalable so once data has been put in, the next project adds to that platform.

2. Implementation
Once you have identified your strategic needs and a discrete measurable project, the implementation starts. This involves data scientists cleaning the data. Then big data engineers ingest and standardize each data source, and then add temporal sequencing to the data. Following that, the data scientists code bespoke algorithms to achieve what is required. There are many AI approaches from NLP to evolutionary computation to machine learning and deep learning so it is the data scientist team to evaluate which approaches are optimal for the combination of data and objectives to be created. At the same time the programmers must create a seem-less process for the interfaces for data ingestion, processing and output visualizations. You can use existing visualization software that are compatible with AI languages such as R or Python or Scala, but keep in mind you have a per user license for these and it may be more cost effective to utilize open source code or create your own.

3. Testing and Use
After a rigorous testing process, the systems are released for use. Initial insights are uncovered, and teams are shown how to use it on their day to day work. This involves a change management process so that the systems are understood and utilized effectively.