PM360 asked experts in everything regarding data and analytics how they are dealing with managing the influx of data and the best ways to turn that information into insights. Specifically, we asked 12 experts, four key questions:
- As many life sciences companies today struggle with data management, what are some possible solutions or tips for how they can better collect, organize, and manage their data so they are able to turn it into actionable insights?
- What new trend, advancement, or application of data analytics do you think holds the most promise for pharma marketers? Why?
- As marketers gain access to more data and the ability to make more particular measurements, what are today’s best practices for establishing which key performance indicators (KPIs) to use to measure a campaign’s success?
- More data has also led to more noise. What are the best ways to help marketers cut through the noise to get the information that will be most beneficial to them?
Why? A single platform cuts down on the number of passwords, access points, and redundancies. To obtain the data needed at your current position, think about this: How many passwords do you need to recall? How many separate locations do you need to access? How many different people do you have to contact? These steps could easily be eliminated.
A single platform allows any department to see their own data, but also to see how their data relates to datasets from other departments. This provides a bigger, more comprehensive picture. The platform should be accessible to all decision-makers and business heads so that single departments cannot derive power by controlling the data. Integrating all data into a single platform also standardizes reporting and helps eliminate reporting errors, allowing fanatic and meticulous diligence on data governance and data quality.
We need to change the way we think and shatter the traditional paradigms of business analytics and data management. Achieving this is not hampered by technology—it is only hampered by our ability to take the critical step to change.
Managing data to achieve actionable insight seems straightforward but without detailed planning on the front-end, then failure on the back-end is a real possibility. It is important to clarify if you are seeking insight to inform next year’s campaign or to optimize while the current campaign is in market. Most of our clients seek both. You must start with clear business questions such as, “Are our targets engaging in content or taking key actions suggesting a behavior shift?” Specific datasets must be captured and analyzed to answer this question. We recommend identifying three to five questions each quarter.
The next step is mocking up key visuals to convey the information and agreeing upon a dashboard prototype. In parallel, you should meet with the individuals that will send the data and clarify the data spec and frequency. Don’t be surprised if you find things missing in the first data exchange—you’ll likely need modifications. When all the correct data is received, data is input for visualizations and additional advanced analytics are completed to answer the business questions. A marketing strategist is also a key team member to provide a marketing lens to convey the most actionable answers.
For maximum leverage across the organization, standardization should be applied to data collection and storage methods to the extent it makes sense to do so. The use of consistent data buckets, such as option lists, age ranges, and geographic regions and levels (e.g., state/CBSA/zip), will decrease time needed to integrate the data and bring rich insights by combining multiple sources. In addition, data should be connected hierarchically by standardized organization names (or ideally, unique identifiers) with organizational structure (parent/child/sister organizations) to extrapolate findings across organizations or groups.
Many third-party data vendors sell cross-reference lists to match their lists to other vendors’ lists; use them when possible instead of recreating the wheel. Share data across your organization and periodically make updated lists of data assets available to other data users within your organization. The more robust access each employee has with a closer view of the big picture, the more they will be able to leverage and integrate various data assets to answer internal and external business questions.
For healthcare, privacy is paramount. To date, this has posed inherent and warranted limitations on data collection and usage. Tokenization will help move pharma forward.
Tokenization technology empowers data-driven advancement while protecting patient privacy. By replacing sensitive patient information with a de-identified, unique value (token), siloed datasets can be linked in new, privacy-secure ways.
Adoption of tokenization technology has allowed companies to combine various disparate datasets—such as clinical, transactional, and social demographic data—within powerful applications where it can be mined and analyzed together.
This improved data access and utility holds great promise and potential. Expect accelerated, more informed medical research; a greater understanding of population health and social determinants; stronger diagnostic accuracy and improved treatment plans; measurable attribution of media’s role in driving positive health outcomes; and more.
- Enterprise reporting is dead. While business users don’t have time to see daily reports in their inbox, IT struggles to maintain and track usage of these reports. This is the Uber and Airbnb generation used to on-demand access to taxi or a room. So why should information access be any different?
- Dashboards are hard, often a dump of irrelevant data for users. Instead of navigating a complex dashboard to find how many prescriptions a doctor wrote in her territory, a sales rep would prefer asking this question to a human-like interface on mobile using voice.
- Learning to use new software is a waste of time for users. Instead a single learning interface that sits like a layer above all enterprise information helps to enable users with contextual insights without the need for training.
- New-age AI platforms automatically adapt to change in an enterprise without the need to re-align data to new organization structures.
All marketing stakeholders must align around a forecast to make key business decisions. Shouldn’t today’s analytical tools make this alignment easier? Pharma companies invariably find advanced analytics for forecasting painful because the data is poorly structured. It is a painstaking process to create, align, and then rework forecasts. The myriad scenarios and versions tend to be highly manual and error-prone. Yet when companies try to implement data management in forecasts, it almost always leads to a trade-off between flexibility and structure. How can pharma achieve the best of both worlds?
Today’s advanced analytics hold the promise of “order-of-magnitude” improvements in forecasting. The most promising applications automate manual tasks that currently take up so much of forecasters’ time, while also enabling them to produce actionable insights and fulfill their critical role in the marketing process. The key to this is the seamless connection of structured data between the tools that each user prefers to work in. The result is seamless data storage and management, reporting and advanced analytics, collaboration, versioning, scenarios, comparison, and other functions. As data analytics advances, forecasters’ positions can genuinely be elevated to serve as the bridge between analytical rigor and actionable insights and alignment.
First, work backward from the ultimate KPI (sales) and list all elements that can drive sales related to the brand (i.e., awareness, equity, site traffic, promotion), the customer (i.e., journey, triggers, media consumption, content drivers), and the marketplace (competition, macro and micro factors).
Second, make sure you have the necessary tools, technology, and talent to bring the data together. The right mix of data scientists and analysts is essential to evolving models and working with marketing teams. Next, set a test-and-learn roadmap that allows you to determine which metrics or combinations are best correlated to sales. The roadmap may vary based on the brand or category life-stage (new vs. mature), marketing/media channels investment (digital vs. offline or email vs. direct mail), and other factors.
Lastly, patience is necessary given the pressure to deliver accurate outputs in “real time” has increased. Models need to be developed and modified over time, especially when new variables come into play. In the end, a shortlist of KPIs will be identified to drive the bottom-line.
When setting campaign KPIs, it’s critical to understand that one size does not fit all. For example, expecting high-decile prescribers to achieve the same growth as low- or medium-decile writers is like expecting your friend who has already eaten two burgers at your barbeque to eat as many hot dogs as someone who just arrived.
In identifying the best and most realistic objectives by segment it is important to let the data guide the way. Analyze the data at both the brand- and segment-level using the following framework:
- Consider the lifecycle stage
- Evaluate prescribing trends
- Are the number of writers and Rx increasing?
- What is the adherence of patients and physicians?
- Identify challenges facing the brand/category
- Quantify opportunities to improve performance
- Map the opportunities by value and time
Using the framework above will help identify and prioritize the most appropriate KPIs. By effectively setting your KPIs at the segment level before developing the campaign, your audience is more likely to consume the message you are serving at your brand’s barbeque.
As pharma marketers increasingly understand that the path to making smart investment decisions hinges on deeply understanding true unmet needs and how people interact with their condition, audience quality—as opposed to the market mix modeling and digital attribution models often used in retail environments—has emerged as the most meaningful KPI.
Audience quality measures how well the reached audience matches with the intended recipients of an advertisement. It is a superior KPI to other measurement approaches because, when an advertisement shows up on a website with organic content that is contextually relevant to a person’s condition, pharma marketers have a better chance of reaching people “in the moment.” As a result, pharma marketers have an opportunity to show up within the context of meaningful engagement, making them highly relevant.
While no measurement model currently tracks audience quality based on both organic and contextually relevant content, pharma teams should work with a measurement company that recognizes the importance of how both contribute to audience quality. Additionally, marketers should find partners that engage and provide content that is contextually relevant to any patient’s journey.
Marketers are eager to collect data, but many don’t act on it. Only 6% of marketers can even view campaign data in anything near real time. How valuable is it to the individual prescriber that a marketing team knows the average email open rate for that prescriber’s specialty over the last six months?
The most successful brands focus their resources on collecting and analyzing data that drives more meaningful customer interactions. If a stakeholder cannot explain how a data point will be used to drive business decisions, there’s not much value in collecting it to begin with.
Of course, it takes great effort to build a digital infrastructure capable of integrating and analyzing data across channels to better understand individual customers. But with the right teams in place, it is possible for marketing teams to gain insight and personalize customer experiences.
Teams that have clearly defined the questions they want their data to answer and found digital experts to collect and manage that data are gaining on their competition quickly. With deeper data visibility and the technology to leverage it, marketers can cut through the noise to evaluate the effectiveness of their efforts, make informed decisions, and strategize for the future.
It is always important to remember that data does not replace common sense. No matter how advanced your analytics capabilities are, there will be gaps in the causal relationship between what the market is telling you and what you should be doing about it. This is often where marketers get lost in the analysis. They react by optimizing their media and messaging mix to improve the performance of individual tactics without the context of the overall marketing objectives. Response date often raises questions about why something happened, leading marketers to confuse “what we did” with “why did that happen.” The original intent gets lost in the details of individual performance measures.
Therefore, when setting up measurement plans, it is extremely important to identify the measures that matter—the measures that have the highest correlation to your primary business goals. Use these as a proxy for measuring your strategic intent and rules for overall marketing success. Set these rules prior to the implementation of any marketing initiative and establish expectations based on an integrated result against established benchmarks. If your rules are clear, your data should enlighten, not obfuscate your success. Shared success metrics act as performance milestones rather than directions.
The abundance of data has created more noise in two ways. First, the amount of data can overwhelm marketers with information and distract them from being able to focus on what is most important. Second, the large amounts of data in themselves create noise when analytics are performed without appropriate techniques and methodologies resulting in outdated or inverted looking spurious relationships that can misdirect attention.
Cutting through the noise requires new methodologies that apply “controlled applications” of machine learning techniques that surface key findings while avoiding spurious findings that “overfitting” ML to datasets can create. New modeling approaches can identify the causal relationships, pinpoint key insights, and drive the most effective course-correcting actions in real time.
Perhaps most important: The translation of heavy-duty analytics into the language of marketers. It is critical to have analytics teams able to communicate the true signal (finding) they have observed and tie the signal back to the business decisions entrusted to marketers to maximize the benefit.