Twenty-first century pharma and biotech organizations are rapidly transforming into data-driven companies. This transformation is critical, future success and discoveries hinge on the ability to quickly and intuitively leverage, analyze, and take action on its data.
In this webinar Lindy Ryan, Research Director at Radiant Advisors, will share her research on how companies successfully manage this transformation by embracing a data unification strategy that’s built on cloud technologies.
Join us and learn how life sciences companies use cloud technology to:
Create a flexible infrastructure with the ability to agilely and quickly unify multiple data sources
TProvide a framework that enables business user agile data access while addressing governance and compliance challenges
Balance the need for data democratization while maintaining proper IT oversight and stewardship
3. 3
Lindy Ryan
Research Director,
Data Discovery &
Visualization
Radiant Advisors
Lisa De Nero
Director,
Life Sciences Solutions
Birst
FEATURED SPEAKERS
21. 21
• Data Governed by a unified semantic layer
• Unification of Data – Multiple data sources
• True Ad-Hoc with logical boundaries
HOW THE BIRST TECHNOLOGY PLATFORM CAN
FULFILL LIFE SCIENCES’ 21ST CENTURY BI
NEEDS
22. 22
WHO IS BIRST
• Enterprise-Caliber BI Platform
– born in the cloud
• 10,000+ organizations rely on
Birst across all verticals
• Founded by Siebel Analytics
veterans, OBIEE
• 80+ Strategic Partners
“ No. 1 in product functionality and
customer (that is, product quality, no
problems with software, support) and
sales experience.”
2014 Business Intelligence and Analytics
Magic Quadrant
23. 23
BUSINESS INTELLIGENCE IS HARD BECAUSE
BUSINESS IS COMPLEX
• The part you see—dashboards, charts,
etc.— is eye candy. The easy part.
• It’s the collection, combination,
integration, de/coding, transcribing and
cleansing of organizational data into a
usable format that is hard.
• Doing so on a repeated basis, over
time is harder.
• Business rules rarely allow you to just
“add-up” data.
24. 24
Business
Agility
Too big, too slow, too old
Data
Governance
Inconsistent siloed results,
lacking security and validation
THE DICHOTOMY…
26. 26
Get Data
(Connect to Source
Applications)
Automated Data Warehouse
Automated Data Model
Intelligent caching /routing
Logical Layer
Arrange Data
(De-normalize Data)
Make data
analytic-ready
(Create Dimensional
Model)
Give data
business
meaning
(Create Business
Model)
Answer
business
questions
(Visualize Analytics)
AUTOMATED MODELING AND DWH
SPEEDS DEPLOYMENT AND DEVELOPMENT CYCLES
27. 27
AUTOMATED MODELING AND DWH
SPEEDS DEPLOYMENT AND DEVELOPMENT CYCLES
Get data
Connect to
Source
Applications
Arrange
data
De-normalize
Data
Answer
business
questions
Visualize analytics
Make data
analytic-ready
Create dimensional
model
Give data
business
meaning
Create business
model
28. 28
AUTOMATED MODELING AND DWH
SPEEDS DEPLOYMENT AND DEVELOPMENT CYCLES
Automated Data Warehousing AND Automated Data Modeling
Intelligent caching /routing
Logical Layer
Get data
Connect to
Source
Applications
Arrange
data
De-normalize
Data
Answer
business
questions
Visualize analytics
Make data
analytic-ready
Create dimensional
model
Give data
business
meaning
Create business
model
29. 29
ANALYTICS NEEDS VARY
Financial
Advisor
Product
Manager
VP of
Supply Chain
Why do we spend so
much time arguing over
who has the “right”
number?
Distribute a report
looking exactly this way
every morning to
thousands of clients
I want to track
performance in a
specially designed
dashboard
Sales
Ops
Manager
Can I ask some ad-hoc
“business” questions – without
touching the dirty data?
30. 30
TOOLS FOR EVERY USE CASE
Pixel Perfect
Enterprise Reporting
Distributed Interactive
Dashboards
Visual Data Discovery
and Exploration
Predictive
Recommendations
Rich Analytic Design Mobile Analytics
32. 32
Business
• Top 15 Global Life Sciences Company
Challenges
• Include digital marketing for better engagement with
HCPs and Patients
• Global BI at business speed for sales & marketing
• Flexibility in sales rep count across product launches
• Disparate data sources lack “conformed dimensions”
• Lack of One Version of the Truth – current solution
has multiple instances and inconsistencies
OPTIMIZED LIFE SCIENCES SALES
WebSources
Results
• Agility to include multiple, new data sources
• Better, targeted messaging to HCPs
• Self-sufficient “specialist”, managers, executives
• Cloud solution < 6 weeks to deploy
• Increased New Patient Starts 10%
Why Birst ?
• Flexibility, agility
• No hardware investment
• Business user friendly – NO training
• Pre-packaged offering lacked flexibility, too heavy
• Full stack in one code base – 1 skill set; 1 solution
• One global solution – unified version of the truth
33. 33
BIRST LIFE SCIENCES ANALYTICS
Sales and Marketing
Analysis
• New Rx and Total Rx
• Call reach and frequency
• Percent to goal
• Top prescribers and top decliners
• Pending patients
• Market share
• % sales target achieved vs. % budget spent
Patient Analysis • Patient profile and demographics
• Target individuals and demographics
Digital Marketing Analysis • How Can I reach my target?
• Once the audience is identified, analyze insights to
broaden scope of campaigns
• What is the optimal mix while maintaining regulatory
compliance?
Product Analysis • Treatment outcome
• Patient historical trend
Supply Chain Analysis • Shelf time
• Inventory in retail and pharmacy
35. 35
LEARN MORE
• Download 2014 WOC Report
– Birst.com/wisdom2014
• Join us for a Live Demo
– Every Tues and Thurs @
11:00 am PT/2:00 pm ET
– birst.com/livedemo
• Contact us
– info@birst.com
– (866) 940-1496 (or +1 415-766-4800)
Today, from pharmaceuticals to global health to the environment, twenty-first century life sciences companies are transforming into data-driven life sciences companies. They are leveraging vast amounts and new forms of data into processes that span from R&D to sales and marketing. And, like in many industries, the data is explosive: already the rate of data generation in the life sciences has exceeded even that of predicted by Moore’s Law itself.
This transformation by life sciences companies into data-driven life sciences companies is challenging, not only because of the sheer volume of data to manage, but because to date there has been a lack of data integration agility, which is a critical success factor in life sciences. Much of the traditional -- and even some of the new -- approaches to data architecture have led to complex data silos that offer an incomplete picture into data, along with slow down the ability to provide access or gain timely insights. Additionally, control of intellectual property and compliance with regulations poses a bevy of operational, regulatory, and information governance challenges.
Of course, the very nature of the life sciences environment is one of non-stop change, growth, and financial investment, too. In fact, projections say that sixty-eight percent of life science companies are expected to increase overall sales and marketing IT spending over the next fiscal year.
And now, adding to this, a strong emphasis on analytics and data discovery for insights is introducing additional challenges in how data is leveraged into the fabric of life sciences organizations.
Future discoveries and successes in life sciences companies hinge on the ability to quickly and intuitively leverage, analyze, and take action on data.
Today’s analytic challenges for life sciences companies can be separated into three distinct categories: the integration challenge, the management challenge, and the discovery challenge, which is the basis for this webinar. We will review these three challenge in depth and then provide three approaches that address these challenges and some supporting case studies from the life sciences industry.
Having highly accessible data not only enables the use of vast volumes of data for analysis, but it also fosters collaboration and cross-disciplinary efforts to enable collective innovation within life sciences companies and among their third party counterparts.
However, while having access to data – all data – is a requirement in life sciences companies, existing data tools and resources lack unification. The integration challenge, then, is ultimately the ability to quickly and agilely unify multiple data sources and provide a full view of information without incurring massive overhead costs. This includes information stored in multiple formats (structured and unstructured), research locations (on-premise, remote premises, or in the cloud), and geographic locations.
After access comes availability -- there must exist the ability to make this data available to support numerous tactical and strategic needs – including providing correct and reliable information to doctors and patients, optimizing multichannel marketing activities, and improving sales force effectiveness through standards-based data access and delivery options that allow IT to flexibly publish data.
Reducing complexity when federating data must also be addressed, and this requires the ability to transform data from native structures to create reusable views for iteration and discovery.
Finally, integration must be agile enough to adapt to rapid changes in the environment, respond to source data volatility, and navigate the addition of newly created data sets.
Traditional data warehouses enabled the management of data context through a centralized approach and the use of metadata, which supported self-service by providing well analyzed business definitions and centralized access rights. However, in highly distributed and fast changing data environments the central data warehouse approach falls short and prioritizes the needs of the few rather than the many. For life sciences companies, this means the proliferation of sharing through replicated and copied data sets without consistent data synchronization or managed access rights.
The management challenge, then, is the guidance and deposition of context and metadata, and the sustainment of a reliable infrastructure that defines and governs access and permissions within the strict context of the life sciences industry.
Management challenges with governance and access permissions are equally procedural and technological. Without a basic framework and the support of an information governance program, technology choices are likely to fail. Likewise, without a technology capable of fully implementing an information governance program, the program itself becomes ineffective.
The third information challenge for life sciences companies could be referred to as a set of “discovery challenges” that ultimately meet the needs of integration, analytics, and discovery while controlling consistency, governing context, and leveraging analytic capabilities.
First is the balancing between fostering the discovery process and environment while still maintaining proper IT oversight and stewardship over data. This is different than what we described in the management challenge in that it affects not only how the data is federated and aggregated, but in how it is leveraged by users to discover new insights.
Then, because discovery is often dependent on user independency, the continued drive for self-service – or, what we refer to as self-sufficiency --, presents further challenges in controlling the proliferation generated by the discovery process as users create and share context. A critical part of the challenge, then, is how to establish a single view of data to enable discovery processes while governing context and business definitions.
Discovery challenges go beyond process and proliferation, to include challenges in providing a scalable solution for enabling even broader sources of information to leverage for discovery, such as data stored and shared in the cloud.
Finally, the evolution of discovery and analysis continues to become increasingly visual, bringing the need for visualization capabilities layered on top of analytics. Identifying and incorporating tools into the technology stack that can meet the needs of integration, analytics, and discovery simultaneously is the crux of the discovery challenge.
Future discoveries and successes in life sciences companies hinge on the ability to quickly and intuitively leverage, analyze, and take action on data.
Today’s analytic challenges for life sciences companies can be separated into three distinct categories: the integration challenge, the management challenge, and the discovery challenge, which is the basis for this webinar. We will review these three challenge in depth and then provide three approaches that address these challenges and some supporting case studies from the life sciences industry.
The first approach to specifically address the integration challenge is choosing abstraction for unification with the support of a semantic layer.
Data abstraction through a semantic layer supports timely, critical decision making as different business groups become synchronized with information across units, reducing operational silos and geographic separation.
The semantic layer itself provides business context to data to establish a scalable, single source of truth that is reusable across the global organization. This is achieved by overcoming data structure incompatibility by transforming data from native structures and syntax into reusable views that are easy for end users to understand and developers to create solutions. It provides flexibility by decoupling the applications -- or consumers -- from data layers, allowing each to work independently in dealing with changes. Together, these capabilities help drive the discovery process by enabling users to access data across silos to analyze a holistic view of data.
Context reuse will inherently drive higher quality in semantic definitions as more people accept – and refine -- the definitions through use and adoption.
One approach to addressing the management challenge is to centralize context in the cloud, which addresses not only the need for integration but for access and storage, too.
Cloud platforms offer a viable solution through scalable and affordable computing capabilities and large data storage. Cloud computing has gone from an idea to a core capability, and many leading life sciences companies are approaching new systems architectures with a “cloud first” mentality. But the cloud also provides the ability to centralize context, collaborate, and be more agile. With the inclusion of a semantic layer for unification and abstraction, data stored on the cloud can be easily and agilely abstracted with centralized context for everyone.
Ultimately, where data resides will have a dramatic effect on the discovery process – and trends support that eventually more and more data will be moved to the cloud.Today, taking the lead to manage context in the cloud is an opportunity to establish governance early on as cloud orientation continues to grow to a core capability over time.
Finally, a third solution relies on embracing visualization for self-service.
Providing users with tools that leverage abstraction techniques keeps data oversight and control with IT, while simultaneously reducing the dependency on IT to provide users with data needed for analysis. Leveraging this self-service or, self-sufficient approach with visual analytic techniques drives discovery one step further by bringing data to a broader user community and enabling users to take advantage of emerging visual analytic techniques to visually explore data and curate analytical views for insights.
Visual discovery makes analytics more approachable, allowing technical and nontechnical users to communicate through meaningful, visual reports that can be published and shared back into the analytical platform to encourage meaningful collaboration. Self-sufficient visual discovery will benefit greatly from users not having to wonder where to go get data -- everyone would simply know to go to the one repository for everything. These tools for visual discovery are highly interactive by nature to enable underlying information to emerge and typically require the support of a robust semantic layer.
And, while traditional BI reporting graphics like standard line, bar, or pie charts provide quick-consumption communications to summarize salient information, exploratory graphics – or, advanced visualizations like geospatial, quartals, decision trees, and trellis charts provide analysts the ability to visualize clusters or aggregate data. And through visual discovery they can also experiment with data through iteration to discover correlations or predictors to create new analytic models.
As we know, life sciences generate an extreme about of information in multiple formats and locations, and each can have a major influence. Integration-of and access-to data enables true democratization of research and information.
In a two year anonymized study, GlaxoSmithKline (GSK) used text analytics software to mine online parenting websites in an effort to understand and analyze concerns – regarding safety, timing, and comfort – that motivate parents to delay vaccinations after a
measles spike in 2011. Capturing candid sentiment data directly from parents allowed GSK to provide doctors with better educational materials and information to supply to parents and patients.
By integrating and analyzing unstructured data against current vaccination data, this research has helped the pharma company reconsider how it helps physicians communicate inoculation information.
Second, through using the cloud as a research enabler, many life sciences companies– including Pfizer, Eli Lilly, and Johnson & Johnson, are demonstrating the viability the cloud for scalability, agility, collaboration, and sharing, which support the claim
that moving larger and larger life science data sets into the cloud is inevitable, and illustrating again the importance of moving abstraction closer to the data to enable global sharing processes and centralize context management.
Eli Lilly launched a 64-machine cluster in the cloud to work on bioinformatics sequence information, then executed the work, and shut down the project within twenty minutes. Lilly’s Senior Systems Analyst for Discovery IT was quoted as saying that while exact cost savings were difficult to calculate, using the cloud helped to circumvent “spiky utilization” and achieve significant time and cost savings.
Finally, today life sciences companies are adopting social media as a new, cost-effective, and rich “source of information” marketing channel to engage directly with customers and patients to measure sentiment and gather real-time market research data to improve existing products and stimulate further innovation.
One case study that has proven the value of visualizing data for reason has been Project: EVO, which is a collaboration between Pfzier and Akili Interative Labs where the two have teamed up to design mobile video game technology to measure cognitive differences in
healthy older adults to identify early warning signs of Alzheimer’s. By comparing levels of amyloid (which is the main component of brain plaques and risk factor for developing Alzheimer’s) and gaming performance characteristics, Pfizer hopes to identify biomarkers that could help identify at-risk populations.
In addition to robust analytic capabilities, this project uses gamification and visualization techniques to discover and communicate insights.
So, in summary, within the life sciences literature, navigating and understanding data has been described as “the greatest challenge to unlocking knowledge and scientific discovery.” Unlocking knowledge and scientific discovery, in this context, requires that analysts and researchers have access to complete, high quality, and actionable information in a way that is agile -- and that leverages available tools and technologies to drive analytics and discovery.
By choosing abstraction for unification, embedding business context into data through the inclusion of a semantic layer, leveraging cloud technologies, and enabling business users with self-service tools that offer robust analytic and visualization capabilities, life sciences companies can continue on their journey to becoming even more data-capable organizations.
Before launching into preso – engage audience –
Ask the question – what tool is used widely in your organization today to get insight into your data?
On what tool do you rely most to make key decisions that drive your business?
The goal is to get them to say “Excel”
Then ask – what is wrong with it? Why not use it for enterprise BI?
Pry as they think about the pains of Excel
The goal is to get them to say
Data anarchy / excel hell – files everywhere with no single version of truth
Very manual / non-repeatable process – that cannot be leveraged continually
Hard to integrate get data from multiple sources
Cannot handle large data sets
Visuals are lacking
But Why is it used so much then?
Offers flexibility to end-users – it can be used to create a “pixel=perfect report” – or to do a pivot table – or to do a fancy chart
What’s the process we follow to make this happen
Connect to Source Applications
Connect securely
Extract data
Full
Incremental
Denormalize Data
Produce “aggregatable” data
Create/flatten hierarchies for roll-ups
Consolidate sources
Cleanse data
Create Dimensional Model
Identify things that are to be aggregated
Identify business entities that
Manage changes and history
Snapshots
Slowly changing attributes
Create Business Model
Semantic layer
Allows business users to create queries without knowing SQL or underlying physical structure
Dsitrubute Insight
Publish heavily pre-digested data (reports)
Adhoc/ visualization
Create interactive analysis (dashboards)
Embed in other applications
What’s the process we follow to make this happen
Connect to Source Applications
Connect securely
Extract data
Full
Incremental
Denormalize Data
Produce “aggregatable” data
Create/flatten hierarchies for roll-ups
Consolidate sources
Cleanse data
Create Dimensional Model
Identify things that are to be aggregated
Identify business entities that
Manage changes and history
Snapshots
Slowly changing attributes
Create Business Model
Semantic layer
Allows business users to create queries without knowing SQL or underlying physical structure
Dsitrubute Insight
Publish heavily pre-digested data (reports)
Adhoc/ visualization
Create interactive analysis (dashboards)
Embed in other applications
What’s the process we follow to make this happen
Connect to Source Applications
Connect securely
Extract data
Full
Incremental
Denormalize Data
Produce “aggregatable” data
Create/flatten hierarchies for roll-ups
Consolidate sources
Cleanse data
Create Dimensional Model
Identify things that are to be aggregated
Identify business entities that
Manage changes and history
Snapshots
Slowly changing attributes
Create Business Model
Semantic layer
Allows business users to create queries without knowing SQL or underlying physical structure
Dsitrubute Insight
Publish heavily pre-digested data (reports)
Adhoc/ visualization
Create interactive analysis (dashboards)
Embed in other applications
Lastly … to compound the problem … your users have varying analytic needs
Tell story of each ‘persona’ –
At end of the day – you have the same problem as before – because you have a different tool for each user’s situation, you get different answers to same question –
And you cannot change your business….
It’s like you need a different tool for each situation
If you are big company:
Pixel perfect reporting is Crystal
Dashbaords are BOBJ/Cognos
Predicitive is SAS
Discovery is Qlik/Tableau
Mobile is… who knows…
And your design studio – is EXCEL!
They all pull the data differently – and give you organization data anarchy…not business intelligence
What Birst is doing is putting all these tools together
On top of a single logical layer – a single business library of all your KPIs
Then we are taking the hardest part, the dirty data management part and making it faster and more accurate then ever before by automating a large piece of that process
We have automated the data warehouse
This gives you a complete set of tools for each individual users – that leverages a single logical later – a single library of your KPIs – to ensure you have business intelligence in a consistent, repeatable, non-error prone way.
Our Key value points:
One single login for entire process, multiple tools for each user
Automation to take care of that Messy Data problem
A logical model – to remove the data anarchy issue and create data synergy
invest in a company-wide BI tool. Previously they had been relying on manual data extracts and a 3rd party data analyst consulting company which would produce monthly static powerpoint and pdf documents summarizing their data. They decided to use Birst as their Company’s first BI tool beginning with a deployment of dashboards to support their product specialists (sales reps) and their regional managers. Since deploying Birst they have seen immediate value. Their data is accessible in a flexible, centralized report environment where specialists can review their performance on-demand and track their progress toward compensation related goals. Important KPIs include the number of patients who have started on their drugs in a specialist’s region and the percent this makes up of their quarterly goal. They also track calls made to prescribers and the number of new doctors who are writing prescriptions. The primary dashboard includes a list of the top prescribers and the top declining prescribers (in reference to the number of patients they have on the drug) which immediately translates into action items for the specialists who can then prioritize their calls to those prescribers.
Looked at pre-packaged offerings – not flexible – too long to customize to their unique needs
This slide is to introduce Birst’s life science analytics solutions at high level. Each life science company may have specific requirements. So most likely this slide will be updated to be relevant by each presenter.
- Sales and marketing analysis: will be particularly introduced in the next slide.
- Patient analysis: Develop more targeted patient profiles that focus not only on products, but also on the ability to pay
for them by analyzing historical health trends in combination with demographics. Identify and target individuals and demographics that could be considered “undiagnosed” with educational campaigns whose goal is to encourage these individuals
to get screened and tested for possible issues. Combine product sales information with patient groups and customer channel information to analyze what tends to lead patients to fill prescriptions at a more consistent rate or what leads physicians to
prescribe certain drugs at a higher rate.
- Operations and financial analysis: Analyze return on marketing events to optimize marketing efforts. Analyze the prescription activity in a geographic region or area to make sales force adjustments according to market size or penetration. Analyze buying trends from the largest customers (managed care providers and governments) to proactively create price points that benefit both the buyer and the company.
- Product analysis: Analyze buying tendencies and treatment outcomes to create more drug and product variations tailored
directly towards different age groups and risk factors. Combine demographics and patient historical trends to target “quality of life” needs of patients (i.e., lifestyle drugs) that improve the day-to-day living standards of patients, especially for non-acute medical conditions.
- Supply chain analysis: Improve production schedules through analysis of which products stay on the shelves the longest and
how well each product is selling. Manage inventories more efficiently based on historical trends and patient behavior to prevent stock-outs at retail and pharmacy locations or other channels.