Data Strategy: Unleashing the power of enterprise data

In the seventh edition of the newsletter, I shared my views on how companies can make the significant shift to data-driven decision-making, and parallels with the Oil and Gas value chain (because Data is the new Oil!). In the newsletter, we introduced several important concepts in the Data value chain - Data Strategy, Data Engineering, Data Governance, Data Science and Data Visualization. In this edition, we delve deeper into the data value chain, starting with Data Strategy.

?Why Data Strategy Matters

?While Digital investments are front and center in organizations, planning for data is usually an afterthought. Most organizations then kick-off a data collection exercise, and invest significant time and energy into this, to make up for lost time. However, this rarely works as intended. First is a frenzied "accumulate" phase with a variety of data being captured. Then is the "agonize" phase where the sheer volume and variety of data is hard to manage, leave aside analyze. And then comes the "abandon" phase where the organization has tremendous data, but is not usable for any meaningful analysis, necessitating a return to "gut-feel decision making".

"Failing to plan is planning to fail" - Benjamin Franklin

Crafting a data strategy: Start with end goals/ outcomes in mind

Organizations have access to a variety of data sources - both structured and unstructured. Some of it is stored (financial, operational, customer) while several signals (telemetry, sentiment, etc) are consumed but not stored for the future. This is primarily due to the lack of awareness about potential use cases - both current and future. This is why it is important to start with ideas and use cases that enable a data-driven organization, and work backwards to identify the data sources needed. In many cases, the critical step is simply to start storing data. With sufficient data volumes in 18-24 months, this data can then be analyzed and mined for insights.

?

Presenting below a structured framework to develop a data strategy for an organization:?

??

1. The first step is to consider what are the questions that the executive team (management) is seeking to answer. The questions could span the entire gamut of the company's value chain:

Customer - Is my customer acquisition velocity healthy?

Sales funnel - Is my sales conversion in line with past/ market trends?

Market share - Am I losing/ holding/ gaining market share?

Product performance - What is the relative performance of product categories/ SKUs?

Product usage - Are there any trends to be gleaned from product usage (telemetry, etc)?

Production - How is my production efficiency (cost of production, plant utilization) and quality (defects)?

Logistics - Am I able to ensure timely availability of inputs and product shipments (OTIF)?

Profitability - How is my profitability across multiple dimensions (customer/ product/ geography/ channel)?

Customer experience - How is my after sales support and customer experience (sentiment, trouble tickets, complaints)?

These questions should ensure coverage across elements of the business model as well as the operating model. The value driver tree for the industry would be a good place to identify the most relevant and important questions.

?

2. Once the questions have been identified, the next important step is to define the Metrics/ KPIs to track the performance and health of the company. We then typically employ a "Diamond Model" to organize and analyze these KPIs. The Diamond Model organizes KPIs into inputs (tracking actions performed, aligned to strategy) vs outputs (financial metrics that measure how the actions translate to company performance). Similar to the value driver tree, the model also establishes the hierarchy of metrics breaking them down from higher-level/ composite metrics (such as ROIC) into underlying KPIs (such as Gross Margin, and further into Cost of Production/ COGS and Cost to Serve).

?In platform/ digital business models, a Customer Journey based model is more apt, given the exponential growth potential and relevance of metrics such as Customer Acquisition Cost (CAC), Customer Lifetime Value (CLTV).

?

3. After defining the metrics, the next step is to define "Dimensions". Here we define the levels of detail by which the KPIs are analyzed. In most organizations, performance is reviewed at a:

Customer level - Based on segment (named accounts, large enterprises, mid-market), sector (auto, energy, pharma, etc), account classification (platinum/ gold/ silver/ etc)

Geography level - Theatre (Americas, EMEA, Asia Pacific), country, region/ state

Product level - Category, Brand, Product Family, Product, SKU

Channel level - In industries with a strong channel mix, this is an additional consideration

These dimensions are required as a "drill-down" capability for management and key stakeholders to review. This requires the source data to be available in the Systems of Record (appropriate Chart of Accounts) and a robust cost allocation framework. These dimensions need to be decided based on practical considerations - essentially aligned to the organization structure. This ensures that each KPI when drilled down has an owner who is accountable for performance.

?

4. The journey of data to insights needs to culminate in a Dashboard. It is important to think through the decisions that will be enabled by the dashboards to enable appropriate presentation (visualization), analysis (forecasting/ predictions), and "Calls to Action" (triggers/ alerts).

?Best practices in dashboard design also include creating "Role-Based" views (which are tailored to the individual/ role) as well as "Self-Serve" capabilities (allowing additional exploratory analysis).

?

5. Finally, dashboards in organizations are also complemented by reports. It is not feasible to create dashboards spanning the breadth of people and roles within an organization. This problem is addressed by creating a catalog of reports which enable cascaded views across business units and functions. Reports need to be created at the unit level which can then be aggregated as appropriate up the ranks and into dashboards.

?However, even the reports catalog needs to be planned and designed carefully. Most organizations report having a proliferation of reports, which can be counterproductive as it results data overload and potential omissions.

Getting started with Data Strategy - The fuel for your AI Strategy

?Every company is currently navigating the Artificial Intelligence (AI) wave. There is significant pressure on organizations to quickly understand and integrate AI across their value chain. This is leading to the belated realization that an AI strategy can only be built on a solid data foundation, which algorithms can extract insights from. It's never too late to build a data strategy, but it is especially important when considering large or significant digital investments such as ERP or CRM systems.

"Well begun is half done" - Aristotle

#genai #artificialintelligence #technology #digital #TechTonicThursday