Boosting results using AI for multi-channel resource allocation

Traditionally pharma sales reps selling to physicians was the largest slice of the sales and marketing pie but this has been shifting to a more integrated multi-channel approach. The challenge for many pharma marketers however, has been the constant addition of more and more channels to the pie, but not an addition of budget. So how should I marketer decide on where and how much to spend their allocations across the channels without wasting any of their budget and spending the right amount where it needs to be?

So much marketing is inefficient and wasteful purely because markets do not know where they are getting the best results and the costs keep increasing, as does the number of channels and competitors. This does not have to be the case.



Big data and AI can change this – however this is not without its’ challenges. The main challenges are in the data – many companies cannot provide strong detail about their marketing expenditures, the data integrity is not optimal due to overlapping channels for marketing expenditures employed simultaneously. There are ways to overcome these challenges with a data normalization approach which allows data gaps to be plugged using government and private sector data (depending on the market and its’ availability) and research data. From this we can determine the optimal mix of marketing expenditures across channels to reduce wasted spend, while increasing the strategic value of marketing insights gained, forecasting effectiveness, and tracking performance. These can be visualized in dashboards for ease of interpretation and use.

Case study

The client
The client was a global team responsible for resource allocation across its’ top priority countries and its top brands. They wanted a way to operationalize a machine learning model that would be an integral part of product planning and marketing execution to ensure optimized resource allocation and boosting of their sales results. The data internally was inconsistent and patchy across the markets.

The solution
Eularis created a model that worked in multiple markets from low data levels to high data levels allowing additional data to be added depending on the market. Some external digital data sources vary by types of marketing expenditure reported (e.g. google, webmd etc) while others were country by country. A vital source of data for every market, required the sales team to add data from the doctor, to be captured in their CRM system and additional market research data. However, in several markets richer data was also available. By collating all of the available data and analyzing against data around new prescriptions written and total prescriptions written, we are able to identify which marketing mix elements work best in which country, region, or target audience segments.

The results
After implementing the system, the sales went up in line with the forecasts with revenue increases at around predicted from the machine learning algorithms. This was not 100% accurate, due to data gaps, but a decent close estimation once we had a few timeline data points and the more data that was added to the system. An additional benefit was that the customers also were showing signs of stronger engagement with the sales and marketing materials and stronger relationships were being forged at the same time. The online dashboards created allowed the marketing teams to log in and track results 24/7 and compare these with competitors. The physicians varied depending on their specialty and geography in terms of which channels work best with them, and the system was able to learn from the data and translate that learning into even more accurate actionable recommendations on optimal allocations as time went on.