What Must Change In The Business To Support Data Science: Transformation Roadmap and Timeline

• Read original article here

The last post was high-level strategy and initial planning. This post is focused on building the business pieces in parallel to these Data Science pieces:

In the last post, I covered the main changes to business strategy, and those need a way to be implemented. The Transformation Roadmap and Timeline create the connection between strategy and implementation. The alignment is critical for a few reasons.

The business makes plans for revenue and margin preservation (cost savings), leveraging Data Science for both. The business’s maturity and the Data Science team’s maturity must move forward together to achieve those goals. If the Data Science team ramps up too quickly, the business cannot fully utilize its capabilities. Projects are set up to fail.

The business moving faster than their Data Science capabilities leads to an execution failure. Projects aren’t delivered on time, and the company will miss projections. Stock prices take a massive hit when that happens.

The Data Science team works with senior leadership to identify new opportunities for revenue growth and margin preservation. These opportunities turn into projects and build out the Product Roadmap. Infrastructure components are required to support projects and capabilities (through automation, deployment, and MLOps).

The Transformation Roadmap is built in parallel. Why? Every opportunity must be evaluated from a Data Science and business perspective. If the business decides to move forward with a project, there are transformations required on both sides.

Senior leadership needs transparency into the bigger picture to decide which projects are feasible. Transformation is expensive. For internal efficiency projects, workflows will be changed, and people are slow to change their workflows. Their tools and processes will adapt first. Training will be required.

For products and features, customer-facing teams have work to do. They must prepare customers for a Machine Learning based experience. Marketing will need time to adapt. Sales teams might too. New pricing models can be necessary.

These are a taste of the most significant changes, but there will be so many more. Building the Transformation Roadmap is the business’s opportunity to see the full scope of change. In immature companies, Data Science projects go over budget because all this is discovered during and after the project.

Data Science is set up to fail because the team can deliver projects that the business is unprepared to use or sell. It can be too expensive to transform in reaction, so the finished projects get shelved. Projects could have generated value but the business as it is now cannot.

Planning upfront brings transformation costs down and reduces resistance to change. Advanced communication is essential because teams can provide feedback before the project is put on the roadmap. The business is engaged during this process. That builds the connection between the Data Science team and the external teams they’ll be working with. Relationships are created from this process.

The Transformation Timeline connects business projections with specific deliverable dates. The Data Science team builds the Data and Analytics Organizational Build Out Plan. This specifies the Data Science Value Stream, and the team’s workflow is implemented based on that value stream.

The build-out is incremental and iterative. It is forward-looking to ensure the work done today supports immediate and long-term business needs. In too many businesses, the Machine Learning capability build-out process stalls repeatedly to fix shortsighted decisions from the past.

Business needs must drive the timeline, but the Data Science team drives the timeline through the Product Roadmap in immature companies. Getting the drivers reversed causes shortsighted decisions. The Data Science team builds capabilities based on a partial picture of business needs. When the two sides finally connect, the Data Science team needs to make changes to support the actual business needs and state of readiness.

The Transformation Roadmap is built from business needs. Those are the Strategic Drivers for Change (opportunities and threats), and they will dictate the rate of business transformation. There is a feasibility component because the business has a maximum transformation rate. Early automation projects are built to increase that rate.

Automation transforms faster than people do. Traditional digital transformation efforts handle most of the early groundwork. The Data Science team introduces intelligent automation to take over a larger share of business processes. Transformation and Data Science are connected because those early projects must happen to accelerate Machine Learning maturity.

The pace of transformation starts slow and increases rapidly. The Data Science team must accelerate their ramp up to meet the rest of the business’s maturity. The Transformation Timeline details:

Again, timing is driven by opportunities and threats. The transformation could happen slower or faster if the business units or Data Science team dictates timing. Either one will break the connection to strategic goals.

The Transformation Roadmap is strategy implementation, and the timeline is execution. Data Science projects make it into production because all the supporting work is in place. Projects produce expected value and stay on budget because the rest of the business has worked parallel towards a common objective.

I explained how senior leadership creates alignment with the Data Science team in the last post. This post has explained how organizations and front-line employees create alignment with the Data Science team. Both are required, or Data Science is set up to fail.