Yup - there are "just" four steps to creating a data driven organization.
The good news is the first two steps are rather straight-forward. Not hard, but certainly not easy to do well.
The not so good news is that Step Three can be a little challenging.
The bad news is that Step Four is hard. Very hard. And without continuous full commitment from the entire organization - over the course of many months / years - becoming truly data driven will remain an elusive goal.
A recent survey by IBM found that only 38% of firms report that they have created a data-driven organization.
So - how do you become one of the 38%'ers?
Step 1: Defining the end-state for your Data Driven Organization
We certainly have a strong point of view on the end-state. So, feel free to use this as a starting (or end-) point. You are very welcome!
Whatever your end-state vision looks like - it needs to be one that is well communicated and aligned across the entire organization - from data engineering, to the analytics teams, to the stakeholders, to the leadership team.
Step 2: Developing a Robust and Scalable Analytics Strategy
Simply put - without a strong Analytics Strategy, your chances of becoming data driven are pretty slim. So now would be a good time to develop (or revisit) your data and analytics strategy.
And not just any haphazardly developed analytics strategy will suffice. It must support and enable the end-state vision for your Data Driven organization.
It also must be robust and scalable.
Robust:
There is a reason we refer to it as an Analytics Strategy rather than a Data Strategy. It needs to encompass all aspects of the analytics supply chain:
Data architecture alignment with systems and infrastructure architecture
Data Dev and Operations: collecting, storing, augmenting, aggregating data
Analytics dev and delivery: analysis, AI/ML, and visualization
Data governance, quality, security, privacy
Analytics Operations (e.g. comms, access) processes and workflows
Guidelines / "policies": access to data, tool usage, "self service" analytics
Provisions for an analytics "community"
Scalable:
Scalable to support business growth and expansion.
Scalable to support growth and advances in technology and analytics capabilities.
Ultimately this means the analytics strategy is a forward looking framework rather than a statement of implementation details (since the details are sure to change).
As such, the strategy should be:
Driven by business needs rather than specific technology (e.g. "must support large amounts of transactional data" vs "must implement Hadoop")
Tied to capabilities not specific platforms, tools, providers (e.g. "need flexible cloud-based data storage to support rapid but unpredictable business growth" vs "we are implementing Snowflake")
We'll drill down into creating a Robust and Scalable Analytics Strategy in an upcoming article. Stay tuned.
Step 3: Creating an organizational structure that supports the analytics strategy
Things get a little more challenging with this step.
While re-orgs are painful for all involved, if your new / updated analytics strategy marks a major turning point with many changes to the current landscape, then some organizational re-alignment will likely be required.
The affects of a mis-aligned analytics organization structure can cause unnecessary and painful roadblocks to success - everything from confusion over roles & responsibilities, to conflicting priorities, to political battles. So, better to feel a little re-org pain now instead of dealing with major pains down the road....
So while there is no cookie cutter approach to analytics organizational structures, the one universal truth is that it is critical for the organizational structure to be highly aligned with, and supportive of, the analytics strategy.
For example: If the analytics strategy is focused on self-service by the stakeholders, then de-centralizing the analysts from the technology team makes good sense. However, if data is a critical component of the business plan and there is a need for constant, fast-turnaround data engineering, then tightly coupling the technology team and the analysts would likely be a better approach.
Other factors affecting the analytics org structure planning:
Anticipated growth. In data, technologies, data science, analysts, and stakeholders. Look out 2+ years and organize for the future.
Corporate staffing model. What, if any, plans are being considered for remote or near- or off-shore staffing? Data engineering, analytics, data science are all areas where skilled resources exists around the country and around the world - and an off-shore model may mean a different organizational model is necessary.
Of course with any potential change in organizational structure comes the risk of power plays and political battles. Check out our article on "The Politics of Data" for additional considerations.
Step 4: Creating, cultivating, and maintaining a data driven culture.
This is the most challenging step.
It isn't enough to have the right people, processes, and technology. Without a Data Driven culture - a pervasive mindset engrained into the corporate DNA - the organization will fall short in realizing the full potential of the data.
Messaging the importance and power of data
It (almost) goes without saying that the messaging and tone of a data driven culture must come from executive leadership.
However, it isn't enough to make proclamations that the company will become data driven.
Actually, such direct statements would likely hinder the effort. While us "data people" might appreciate the concept, it is important to note that it is not universally embraced. Heck, even many of us "data people" are growing a bit weary of the term "Data Driven". So the messaging will likely resonate more pervasively if positioned along the lines of "we endeavor to support decision-makers by empowering them with the full power of our data and insights".
The messaging also needs to cascade throughout the organization - and not be limited to those functions that are typically swimming in data such as sales and finance. This messaging needs to echo through the corridors and be heard by those in less data intensive areas such as operations, REW, legal.
Backup up the messaging with action
The message can't be empty words. It needs to be backed up with action. Genuine action that is not just a one-and-done internal marketing campaign.
Some ways to "show don't tell":
Avoiding half-hearted attempts that aren't followed up. For example - recommending that everyone read/listen to the latest book on data driven organizations is good. Never doing anything to reinforce it is not so good - as it comes across as being just another fad that will go away in a couple of months.
Fully embracing KPIs across the organization. Each business function should have their own KPIs and these need to cascade down to their departments and be, across the board, actively monitored and actioned.
Ensuring that each employee has a direct connection to (supporting / contributing to) the corporate and/or departmental KPIs.
Project success criteria - that are actually tracked and reviewed after project completion.
Establishing a formal framework that leverages data and insights to support critical strategic decision-making: e.g. data catalogs, business glossaries, easy-to-find and access reports/dashboards, balanced scorecards, processes for "drill-down" analysis, a formal analytics consultancy function.
Championing data and analytics
Companies such as Google and Cisco have "Data Evangelists" or "Data Champions". These are often external-facing roles - enabling their customers/clients to fully leverage the power of their data and the analytics capabilities.
A powerful twist on this approach is an internal-facing Data Champion - or perhaps "Analytics Customer Success Manager (ACSM)" as this better reflects the important task of enabling the stakeholders to successfully leverage the organizations data and analytics capabilities.
This role would be charged with cultivating, fostering, and ultimately maintaining a data driven culture through:
1) Engaging the stakeholders
Understanding the needs and goals of each line of business
Helping align the business goals to analytics needs (feeding the Analytics Strategy)
Helping align the analytics needs to data needs (capabilities, data sets, etc.) and working with the technology team to implement
Helping define and implement business (and system) processes to effectively consume the data/insights
Helping re-define business decision-making processes to more effectively incorporate data and insights
Coaching those who are not "data minded" on the power and potential of data in ways that they can relate to.
2) Being the nexus between stakeholders, governance, data engineering/architecture, operations:
Becoming the embodiment of the analytics strategy across the organization
"Connecting the dots": Fostering cross-functional analytics synergies
Providing input to data engineering on the capability needs of the of the business
Advocating data governance as an enabling function (rather than one that impedes the business)
Removing roadblocks to successful analytics activities. See our article on Analytics Operations which describes the hidden costs of inconsistent communications and informal processes.
3) Leading the charge for well-defined (and properly documented) corporate and departmental KPIs and metrics
4) Tracking data quality and shepherding remediation efforts to drive "trusted data"
Making it stick
Even with a Data Evangelist/Champion function, making a data driven culture truly stick and become engrained in the corporate DNA takes effort across the organization.
Ways to help make this happen:
Data Literacy - promoting the ability for all users to be able to "read", "interpret", and "draw conclusions" from data, reports, dashboards. Some people and teams do this well, but others need help. Adopt the motto of "no team left behind" on the road to a data-driven organization.
If you have not done so, leverage the concept of a Data Ops function to align the data dev, data science, and analysts activities along a business-focused approach.
Foster an analytics "community". Not just the analysts, but encourage participation from non-data people who want to be more data literate.
Hiring / Onboarding. From a candidate perspective, incorporate data awareness into the candidate screening and interview process. During new employee orientation, incorporate information on data capabilities to set the tone early.
One final thought...
Ironically, becoming a Data Driven organization is an art not a science.
Your organization is a living entity. It is fluid, dynamic, and in many ways unpredictable.
Therefore, determining what, when, where, and how to guide the organization towards "Data Driven-ness" can't be done with an algorithm.
Though, of course, that just means a I'll be proven wrong by a data scientist somewhere... and we'll all be OK with that!
Comments