Value Engineering: The Secret Sauce for Data Science Success
If we believe that the Big Data Business Model Maturity Index described in Chapter 1, The CEO Mandate: Become Value-driven, Not Data-driven, is what organizations could do to become more effective at leveraging data and analytics to power their business models, then your next question is "How can I achieve that?"
Let me introduce you to the Data Science Value Engineering Framework (see Figure 2.1).
Figure 2.1: Data Science Value Engineering Framework
The Data Science Value Engineering Framework (process) provides a simple yet effective methodology for exploiting the economic value of your data and analytic assets; a methodology to drive the collaboration between the business subject matter experts (stakeholders) and your data science team to apply data and analytics to improve the operational and business effectiveness of all industries including healthcare, public safety, manufacturing, transportation, energy, education, the environment, sports, entertainment, financial services, retail, and more.
Let's drill into each of the steps of the Data Science Value Engineering Framework—the "How to do it" framework.
Step 1: Identify a Strategic Business Initiative
If we are focused on using Value Engineering to deliver meaningful and relevant value, then the How conversation must start with a focus on a business' Strategic Business Initiatives. A strategic business initiative is characterized as:
- Critical to immediate-term business success.
- Documented (either internally or publicly).
- Cross-functional (involves more than one business function).
- Owned and/or championed by a senior business executive.
- Has an actionable and measurable financial goal (that is, reduce, increase, optimize, rationalize).
- Has a defined delivery timeframe (12 to 18 months).
An organization's key business initiatives can be found in annual reports, analyst briefings, executive conference presentations, press releases, or a chat with your executives.
Moving your data center to the cloud, transitioning from Skype to Zoom, and arming your employees with tablet computers… are not strategic business initiatives. Those are technology initiatives that may or may not have defensible, financially measurable, business or operational impact.
DEAN OF BIG DATA TIP:
- Reduce inventory costs
- Reduce unplanned operational downtime
- Improve customer retention
- Improve yield
- Improve technician "first time fix" effectiveness
- Improve supply chain reliability and quality
Strategic business initiatives focus on business outcomes that have articulated financial value such as optimizing operational efficiency, reducing costs, improving revenues and profits, enhancing customer value creation, mitigating risk, and creating new revenue opportunities.
Step 1A: Identify Metrics against which to Measure Progress and Success
A critical part of understanding your strategic business initiative is to identify the metrics and Key Performance Indicators (KPIs) against which the success or progress of that business initiative will be measured.
For example, if our key business initiative is to increase "Same Store Sales" (where "Same Store Sales" is defined as the difference in revenue generated by an organization's existing outlets or stores over a certain period as compared to the similar previous period), the following metrics or KPIs might be critical in measuring the progress and success of that initiative: Average Revenue per Visit, Volume of Store Traffic, Revenue per Employee, Line Wait Time, % Abandonment, % Mobile Orders, Positive Social Media Mentions, and "Table Turns" (the time it takes to convert or "turn" a table from one customer to the next customer).
DEAN OF BIG DATA TIP:
Make sure you have a robust set of metrics and KPIs to avoid the unintended consequences that can occur due to a too narrow view on how the organization will measure the business initiative progress and success. Brainstorm multiple lead indicators that can provide early readings on the progress and success of the business initiative, and then prioritize those lead indicators (because not all measures are of equal value.
Step 2: Identify Key Business Stakeholders
Once we have identified the targeted business initiative and the metrics against which we are going to measure progress and success, next we want to identify the Business Stakeholders who either impact or are impacted by the targeted business initiative. The stakeholders typically represent 4 to 5 different business functions in order to yield a diverse set of perspectives on how the organization plans to address the targeted business initiative.
Ideally, we want can create a Persona (a Design Thinking tool) for each potential stakeholder to help us better understand their individual as well as the overall organizational challenges, roles, responsibilities, pain points, and key operational decisions for the targeted business initiative (see Figure 2.2).
Figure 2.2: Stakeholder Personas
As you build the Stakeholders Personas, be sure to ask and understand "Why is this business initiative important to them?" and "What is their personal win condition or personal benefit from the successful execution of this business initiative?"
DEAN OF BIG DATA TIP:
It is critical to give all the key stakeholders a "voice" in exploring, ideating, and defining the criteria against which progress and success will be measured. While it is important from a data science perspective to have a diverse set of criteria against which to optimize, organizationally it's even more important that everyone feels aligned and committed to moving forward together. More digital transformation journeys die due to "passive aggressive" behaviors than inadequate technology.
Step 3: Brainstorm and Prioritize Decisions (Use Cases)
The next step in the Value Engineering process is to brainstorm the Decisions that each of the different stakeholders needs to make in support of the targeted business initiative. My findings are that if you identify the right set of stakeholders in Step 2, then the brainstorming and prioritizing of decisions flows very quickly and naturally. Why? Because these stakeholders inherently know the decisions that they have to make in support of the business initiative as they have been trying to make these decisions for year…decades…maybe even generations. Examples of such decisions include:
- Who are my most valuable customers?
- Which students are at risk of attrition?
- What products are likely to break?
- How much inventory am I going to need?
- Which marketing promotion is optimal for the target audience?
- What's the optimal price?
- What's the optimal discount to get the customer to buy?
- What are the right dietary recommendations for this individual?
My observation is that while the decisions have not changed over the years, what has changed—courtesy of massive datasets and advanced analytic algorithms like AI, Machine Learning, and Deep Learning—are the answers. And that's where the Data Scientists who are trained to optimize decisions come into play.
DEAN OF BIG DATA TIP:
A Decision by its very nature is actionable; a conscious pronouncement to take an action. A Question, on the other hand, is useful for validating information or provoking out-of-the-box thinking, but on its own does not imply an action to be taken.
At this stage of the Value Engineering process, we need to aggregate decisions into Use Cases or clusters of decisions around a common subject area that have measurable financial ramifications. To facilitate the aggregation of the decisions into use cases, we use the frame of the targeted business initiative to guide the process.
In Figure 2.3, the targeted business initiative is to "Increase Same Store Sales." On the left side of Figure 2.3 are the brainstormed stakeholder decisions that support the targeted business initiative. Then we use the "Increase Same Store Sales" business initiative to group the individual decisions into clusters of decisions (use cases) around common subject areas such as increase store traffic, increase shopping bag revenue, and increase corporate catering.
Figure 2.3: Pivoting from Decisions to Use Cases
We then label the use cases in Figure 2.3 in an Action Format:
- Identify the appropriate [Verb]. Example Verbs could include Increase, Decrease, Optimize, Reduce, Consolidate, Rationalize, and so on.
- Identify the [Metric] we are looking to impact. Examples Metrics could include Customer Retention, Margins, Visits, Inventory, Unplanned Downtime, Fraud, Waste, Shrinkage, and so on.
- Give the use case a [by X%] goal. Note: you don't need an exact goal at this time in the Value Engineering process. It is sufficient just to use the generic goal of [by X%].
Finally, we use the Prioritization Matrix to drive consensus across the different stakeholders on the top priority use cases based upon the value and implementation feasibility of each use case vis-à-vis each other over the next 9 to 12 months. The Prioritization Matrix process provides a framework for driving organizational alignment around the relative value and implementation feasibility of each of the organization's use cases (see Figure 2.4).
Figure 2.4: Prioritization Matrix
Some key points about the Prioritization Matrix:
- The Prioritization Matrix process weighs the "value" (financial, customer, operational, and environmental) of each use case against the implementation feasibility (data, architecture, technology, skills, timeframe, and management support) of those same use cases over the next 12 to 18 months.
- The Prioritization Matrix process gives everyone an active voice in the identification, discussion, and debate on use case value and implementation feasibility.
The Prioritization Matrix is the most powerful business alignment tool I've ever used. It works every time…if you do the proper preparation work and are willing to put yourself in harm's way as the facilitator.
DEAN OF BIG DATA TIP:
Note: Steps 1 through 3 are covered in excruciating detail in my previous book, The Art of Thinking Like a Data Scientist, including templates and hands-on exercises.
Step 4: Identify Supporting Analytics
Now that we know our top priority use case, we want to identify the predictive and prescriptive analytics that supports the targeted use case. Sometimes it is easier to identify the supporting analytics by asking the stakeholders what Questions they need to answer with respect to the targeted use case.
Then we can walk the stakeholders through the "Thinking Like a Data Scientist" process to convert those questions into predictions and prescriptive actions (see Figure 2.5).
Figure 2.5: Transitioning Questions into Predictions
Figure 2.5 shows some questions and the resulting predictions and the prescriptive actions using an agricultural company example. We start with the question and then convert the question into a predictive statement, such as:
- "What were revenues and profits last year?" (the question) converts into "What will revenues and profits likely be next year?" (the prediction).
- "How much fertilizer did I use last planting season?" (the question) converts into "How much fertilizer will I likely need next planting season?" (the prediction).
Next, we ask the stakeholders if we had those predictions, how would you use those predictions to make operational decisions (which then becomes the focus of the prescriptive actions)?
It's a simple process that builds upon the questions that the stakeholders are already asking today and then guides the stakeholders to the necessary predictive and prescriptive analytics…the key to thinking like a Data Scientist.
Step 5: Identify Potential Data Sources and Instrumentation Strategy
The next step is to brainstorm with the business stakeholders what data you might need to make the predictions identified in Step 4. To facilitate the data sources brainstorming, we simply add the phrase "and what data might you need to make that prediction?" to the prediction statement.
- What will revenues and profits likely be next year…and what data might you need to make that prediction? The data source suggestions might include commodity price history, economic conditions, trade tariffs, fertilizer and pesticide prices, weather conditions, fuel prices, and more.
- How much fertilizer will I likely need next planting season…and what data might you need to make that prediction? The data source suggestions might include pesticide and herbicide usage history, weather conditions, crops to be planted, pest forecasts, soil conditions, and more.
We complete the brainstorming session between the business stakeholders and the data science team by creating a matrix of ranked data sources, using the aggregated judgement and experience of the business stakeholders, that estimates their potential predictive relevance for each Use Case (see Figure 2.6).
Figure 2.6: Data Value Assessment Matrix example
The data science team can then use the relative data source rankings in Figure 2.6 to start their analytic exploration process.
Step 6: Identify Supporting Architecture and Technologies
Finally, we'll need a modern architecture with state-of-the-art technologies (likely with lots of open source options) upon which we can build a solution that delivers the business value. While the architecture and technology choices are never easy, at least you'll understand what technologies you will need AND what technologies you won't need to support your targeted business initiative and the supporting use cases.
If "what" your organization seeks is to exploit the potential of data science to power your business models, then the Data Science Value Engineering Framework provides the "how" your organization can do it.
The Value Engineering Framework starts with the identification of a strategic business initiative that not only determines the sources of value but provides the framework for a laser-focus on delivering business value.
A diverse set of stakeholders is beneficial because they provide different perspectives on the key decisions upon which the data science effort seeks to optimize in support of the targeted business initiative.
The heart of the Data Science Value Engineering Framework is the collaboration with the different stakeholders to identify, validate, value, and prioritize the key decisions (use cases) that they need to make in support of the targeted business initiative.
After gaining a thorough understanding of the top priority use cases, the analytics, data, architecture, and technology decisions now have a value-centric framework within which to make those decisions (by understanding what's important AND what's not important).
- How well do you understand the financial impact of the organization's Strategic Business Initiatives?
- How well do you understand the metrics against which business initiative progress is measured?
- Have you aligned key stakeholders with the Value Engineering Framework
- Have you identified and validated use cases with stakeholders?