Start DataOps WITHOUT a Massive Investment
My last blog mentioned a Gartner article that said “…use DataOps….without massive investment” but the article didn’t describe how. I said I would explain how to make it a reality, so read on.
There are two main strategies for implementing DataOps; bottom-up evolution or top-down transformation. The quickest way to start is bottom-up. The quickest way to finish and achieve a mature practice that is embedded company-wide is top-down. A common need for both strategies is to use change agents. DataOps will change how the company finds, propagates and uses data so it critical to have one or more change agents who:
- Are voracious learners
- Do not wait for orders to take action on new ideas
- Express excitement freely concerning new ideas and change
- Demonstrate a sense of urgency to capitalize on innovations and opportunities
- Challenge the status quo
- Transcend silos to achieve enterprise results
- Skillfully influence peers and colleagues to promote and sell ideas
- Display personal courage by taking a stand on controversial and challenging changes
Start DataOps quickly by finding a few change agents to begin adopting practices. They may be new staff or long-term established staff with a network of relationships and understanding of how to get things done across the enterprise. You have some change agents in your organization today so find them; maybe you are one of them. Collaborate with them, start applying DataOps methods, and evolve the capability as you have successes.
The top-down Transformation of DataOps leverages same of the same methods as bottom-up, but with more structure and formality. The steps are:
- Identify executive sponsor
- Define vision / charter and inform stakeholders
- Develop Roadmap
- Execute and Advertise/Market
- Periodic assessment and renew plan
- Reinforce the DataOps Culture
You need support from an executive sponsor since you will into resistance from some team challenges, policy roadblocks or funding. It’s a lot easier, if not essential, to have a senior director, VP or C-level officer that you can work with.
Second, formally define and document your vision and charter. One way to start is to simply ask your executive sponsor “What keeps you up at night?” and “If our DataOps turns out to be successful, what would that look like for your perspective? How would you measure or talk about the outcomes?” You should also review your company’s annual report and incorporate priorities of the CEO or Chairman.
Third, launch your DataOps blueprint which consists of three elements:
1. Strategic Roadmap is a “checklist” of milestones or outcomes arranged in multiple tracks and phases. Specific leaders/managers assume responsibility for tracks and phases, but the strategic roadmap does not specify “how” the milestones are to be accomplished, only “what” the outcomes would be.
2. Program Roadmap is for approval of a specific initiate including business justification, costs, change drivers, timelines, current/future state models, risks and constraints. This map adds concrete initiates to the strategic roadmap and plays them out in phases.
3. Project Plan is detailed efforts for a program initiate showing a detailed break-down of activities, resources, dependencies, costs, deliverables and other elements defined by the Project Management Body of Knowledge (PMBOK). Start your DataOps Transformation with at least the first project clearly defined.
The main value of the Strategic Roadmap is to quickly develop the plan; 1 week or a few days may be sufficient. I have created roadmaps as fast as a 4-hour workshop. The components of the roadmap are:
a) Program Owner(s): The person responsible for ensuring that all or parts of the roadmap are completed
b) Program Sponsor: Senior staff supporting the resources
c) Roadmap Milestones ordered by tracks and phases:
o Tracks defining at least three dimensions; People and Organization, Process and Policy, Technology and Infrastructure
o Phases: For a 3-year roadmap phases are Year 1, 2 and 3, For a one-year roadmap phases are Quarter 1, 2, 3 and 4, For a 3-month roadmap phases are Month 1, 2 and 3
The complete blueprint will take a few months to fully define and gain agreement across key stakeholders. It will take longer than the evolution approach to get started, but you will have a foundation to scale the capability and sustain it to become adopted company-wide.
So that it’s. Start with your DataOps with little investment. A massive investment may happen in steps 5 and 6 as it grows and becomes adopted across the enterprise, but when that happens, the payback will be obvious and the investment will be justified. If you want help getting started or help defining your Roadmap, send me note and let’s talk. In any event, have fun!
