Here are the key pros and cons of allowing data scientists to publish their discoveries:
Pros:
- It helps attract top talent as data scientists want to publish and advance the field. Publishing is an important part of an academic career.
- It can help the firm gain recognition and credibility in the data science community. This enhances the firm's reputation.
- Published work may provide ideas that other parts of the business can build upon, even if the initial project is not directly useful. This advances the overall data science capabilities.
Cons:
- Important discoveries or techniques could be adopted by competitors before the firm has fully leveraged them internally. This reduces the firm's competitive advantage.
- There may be risks of exposing
2. How do you ensure successful
data-driven decisions?
1. Firm’s management must think data-analytically:
• understand fundamental principles of data management.
• supply the appropriate tools resources to data science teams
• willing to invest in data and experimentation
• steer data team in the right direction
• ask probing questions
2. Management must create a culture where data science, and data scientists will
thrive.
• diverse team
• collaboration
• Once the data science capability has been developed for one application, other application throughout the
business become obvious. “Fortune favours the prepared mind” — Louis Pasteur
3. How to achieve competitive
advantage with data science?
• Examples: Amazon vs Borders, Dell vs. Compaq
• Prerequisite for Competitive advantage:
1. DATA ASSET: the asset must be valuable in the context of our strategy.
2. DATA SCIENCE CAPABILITY: competitors must either not possess the
asset, or must not be able to obtain the same value from it.
• Important Questions:
• Do we have a unique data asset? If not do we have an asset the utilization
is better aligned with our strategy than with the strategy of our competitors?
• Or are we better able to take advantage of the data asset due to our better
data science capability?
4. Strategy for competing
based on data science
• Even if we can achieve competitive advantage, can
we sustain it? HOW? Always keep one step ahead:
• invest in new data asset
• develop new technologies and capabilities
• Note: few companies are able to execute this
effectively
5. Alternative: Achieving Sustainable
Competitive Advantage
• KEY: Achieving sustainable competitive advantage
1. Formidable historical advantage:
• Amazon example (below cost books in the 1990s) : amassed huge data assets
early.
• Switching costs (Amazon recommendation system)
2. Unique Intellectual Property:
• Novel techniques for data mining
• Novel techniques for using the results
• In both cases, these techniques might be patented or simply trade secrets -
competitors will be unable to duplicate these techniques, or will create an
increased expense for them.
6. Alternative: Achieving Sustainable
Competitive Advantage
3. Unique Intangible Collateral Assets:
• Competitors may not be able to put our solution in practice
• It is often not clear to a competitor how algorithm performance is achieved in practice.
• Other intangible asset: company culture that embraces business experimentation
4. Superior Data Scientists:
• Huge variance in the quality and ability of data scientists
• Competition: Netflix, KDD Cup
• Addition catch: Top notch data scientists are in high demand
• Quality data science team leads to another sustained competitive advantage over competitors
• Understanding the learning method of a data scientist will inform your hiring efforts.
• If you can hire one master data scientist, top-notch aspiring data scientists may come to apprentice with
him/her.
• Data Scientists need to have a strong professional network. The better the network, the better the solution.
7. Alternative: Achieving Sustainable
Competitive Advantage
5. Superior Data Science Management:
• Possibly even more critical to success for data science in business is having
good management of the data science team. They must possess the following
set of abilities:
1. Understand and appreciate the needs of the business. And anticipate
the needs to produce new data science project across different
departments
2. Communicate well with and respect “techies” and “suits”
3. Coordinate technically complex activities.
4. Anticipate outcomes of data science projects (similar to R&D process.)
Have and intuitive sense of which projects will pay off.
5. They need to do all this within the culture of a particular firm.
8. Attracting and nurturing data
scientist and their teams
• Firms that have an advantage in hiring are those that create an
environment for nurturing data science and scientists. How?
• Encourage your existing data scientists to become part of the local/
global data science communities
• Publishing advances
• Engaging academic data scientists. Fund research programs, funding
PHD student.
• Take on data scientists as scientific advisors (academics, board
members, etc): this can substantially increase the eventual solution.
• Hire third party to conduct the data science. Caveat: their interests are
not always aligned with their customers’ interests
9. Additional ways to position
oneself for success
• Examine Case Studies: work through many examples of the application of data
science to business problems
• Formulate your own case studies.
• Working through the connection between the business problem and the
possible data science solutions
• Work through many different types of cases. Creates flexibility within a team.
• Accept creative ideas from any source/level:
• Data scientists should be encouraged to interact with employees throughout
the business. This keeps all levels of management open to data-driven
solutions
• Data scientist performance evaluation should be based on how ideas to
improve the business.
10. Additional ways to position
oneself for success
• Be ready to evaluate proposals for Data Science Projects: ideas for
improving decisions through data science can come from any level.
• Managers, investors, and employees should be able to formulate such
ideas clearly, and decision makers should be prepared to evaluate them.
• Important questions:
• Is the business problem well specified? Does the data science solution
solve the problem?
• Is is clear how we would evaluate a solution?
• Would we be able to to see evidence before making a huge investment
in employment?
• Does the firm have the data assets it needs?
11. Data Science Maturity
• To realistically plan data science endeavours a firm should asses, frankly,
and rationally, its own maturity.
• Data Science Maturity Spectrum:
• Low: ad hoc; no formal training; managers have little understanding
of the fundamental principles of data science and data analytical
thinking.
• Medium: employs well-trained data scientists and business
managers who understand the fundamentals principles of data
science.
• High: firms continually work to improve data science processes
(and not just solutions). Challenge the team to instill processes that
will align their solutions better with the business problems.
12. Discussion:
• Should firms allow data scientists to publish their
discoveries? Evaluate the pros and cons of this
approach.