“Focusing is about saying ’no’.”Steve Jobs
This is the best strategic advice I have ever come across. Whether it is your personal or professional life, your ability to say NO is what determines your success. It is the difference between speed and velocity.
For data professionals, knowing when to say ‘no’ is especially important in the COVID19 world. Conflicting demands resulting from immediate, stressful, and emotional events require focus to prioritize.
“Attention, like a flashlight beam, illuminates one subject only to darken another.”Richard Rumelt, Good Strategy Bad Strategy
A good strategy helps you consistently shine the light on the right path. It highlights what is important and what is actionable. A good data strategy lays out which data problems are worth solving, taking into account the capability and capacity of the team. It makes you less-shortsighted than your competition. Staying competitive is not the same as adding the new buzzword to your strategy and our industry is especially prone to those.
If your data strategy contains high sounding buzzwords, without a policy or set of coherent actions, then it is most likely a bad strategy.
3 elements of a good strategy:
- A diagnosis that defines or explains the nature of the challenge. A good diagnosis simplifies the often overwhelming complexity of reality by identifying certain aspects of the situation as critical.
- A guiding policy for dealing with the challenge. This is an overall approach chosen to cope with or overcome the obstacles identified in the diagnosis.
- A set of coherent actions designed to carry out the guiding policy. These are steps that are coordinated with one another to work together in accomplishing the guiding policy.
What is the data diagnosis?
According to Rumelt, before a good strategy can tell you what to do, it must first help you figure out what is going on.
Unlike the field of medicine, where a diagnosis can be any combination of the 10,000 human diseases, the list of diagnoses in the world of data is quite limited and is often a variation of one of these:
- Inability to find the relevant data
- Inability to leverage all of the organization’s data
- Employees having access to data they shouldn’t have
- Too many business intelligence (BI) tools
A common symptom of these problems is a report sprawl. Many reports, in many locations created by many different people using different tools. Report sprawls take up expensive resources: development/maintenance of redundant data artifacts, analysts spend most of their time in redundant data discovery and preparation and, leaders lacking trust in what is presented to them.
Organizations experiencing this vicious cycle can narrow the underlying problem to one or more of these 4 common diagnoses. Once you have the diagnosis, FAIR data principles should make up the guiding policy of your data strategy.
What is FAIR Data?
FAIR principles were launched at a Lorentz workshop in 2014, and in 2016, the ‘FAIR Guiding Principles for scientific data management and stewardship’ were published in Scientific Data.
- Findable: Ensuring that data can be found by both humans and machines. E.g. making sure data artifacts are cataloged with rich contextual metadata.
- Accessible: Ensuring that once data is found, the user needs to know how to access it. This could include going through authorization and/or authentication process.
- Interoperable: Ensuring that data can be integrated with other data and that they can be utilized by other applications or workflows for analysis, storage, and processing.
- Reusable: Ensure that data and related metadata are well-described so that they can be replicated and/or combined in different settings.
Why FAIR data?
Organizations should adopt FAIR principles in the context of enterprise data because they are:
- are agnostic to technology, industry, and type of data artifact
- emphasize machine-actionability to prepare for the future
- make data human actionable to solve the problems of the present
- are stakeholder-friendly
Pick your favorite data industry buzzword (chatbots, augmented analytics, data automation, data fabric), at the core of it all, is the fact that we will rely more and more on computational support to deal with data. Your data strategy needs to account for ever-growing volume, variety, velocity, and veracity. It must be future-oriented. FAIR principles do exactly that. They emphasize machine-actionability to prepare for the future while making data human actionable to solve the problems of the present.
FAIR principles carry knowledge from use case to use case, from stakeholder to stakeholder. It will let business areas to carry the momentum or amplify the impact of actions and decisions. Better interoperability and reusability will give horizontal value to the data—which in turn lowers the barrier for each step on the knowledge ladder
FAIR data is an easier sell to stakeholders. It is jargon-free and without the negative connotations of data governance. Data management concepts and practices like data governance, metadata management, master data management, data quality —all require establishing understanding and use-cases for the stakeholders to understand the value. With “findable, accessible, interoperable and reusable”— the value-add is clear and concise. Embed these words in conversations with the stakeholders at every stage of development of any data solution.
For example, when gathering requirements, questions like the ones below can help clarify expectations and help you decide whether or not a project follows FAIR principles.
Will you re-use any existing data? Which data used in this project will be made available to others in the organization? If certain artifacts cannot be shared, explain why? Are there well-documented processes to request access? What data and metadata definitions/standards will you follow to make this data interoperable? Where will you host/catalog the data artifacts of this project?
FAIR principles give the data teams a concise way to explain why they can or cannot support a solution. Consequently, by following a consistent set of principles to accept or decline projects, your data team can ensure that actions are coherent towards the larger purpose.
If a project doesn’t contribute to making your enterprise data FAIR, say ‘no’.