Start from decisions, not data: Meet Val Matisons

Zennify Team

‍“What sets organizations apart isn't their technology stack. It's the people they have and the decisions those people make, and whether those decisions are visible to everyone who needs them.”

Val Matisons frequently starts discussions with leadership teams by asking: “What are the first 20 questions you want your data to answer?”

The conversation that follows often reveals an important gap: while organizations invest heavily in data, they are not always aligned on the specific decisions that data should enable.

As Zennify’s newly appointed Head of Data Strategy, Val brings nearly four decades of experience helping organizations translate data into better decisions. Over his career, he has built and led data practices at global consulting firms including Avanade, Accenture, PwC, and Tata, advising clients across financial services, healthcare, telecommunications, and other highly regulated industries. His financial services work includes extensive engagements with Canada’s major banks and numerous insurance providers.

His approach has remained consistent: start with the business outcome, then work backwards to the data foundation that supports it.

"I help clients make better decisions. That sounds simple, but consider: the decisions you and your coworkers make determine whether your business maintains or loses its position in the marketplace. Making the correct decisions quickly and executing them consistently requires maximizing the utility of data, the most fundamental component in decision making."

Here's the approach Val uses to make that happen.

Start from decisions, not data

Many data initiatives begin with discussions about data quality, data management, or data governance. While these are essential capabilities, they often address the mechanics of data rather than the outcomes the business is trying to achieve. The real value comes from designing information that supports specific business decisions.

Val’s approach reverses the traditional sequence. Instead of starting with the data, he begins with the decisions that matter most to the organization.

“The goal is to enable answers that move the needle for the business. That means working backwards from high-value decisions and designing the data products and workflows required to support those decisions consistently.”

This decision-first approach starts by identifying three to five recurring decisions within each line of business that materially affect revenue, risk, or operating cost.

In financial services, examples include:

Which customers should we proactively retain or cross-sell this month?
Which loan segments or regions are showing early signs of increased default risk?
Which fraud patterns require new detection rules or models?

Each decision is then linked to a specific performance metric, such as non-performing loans, fee income, or customer satisfaction, and aligned to a business cadence (daily, weekly, or monthly).

At that point, the conversation shifts. The focus is no longer data in the abstract, it becomes about delivering actionable insights that support concrete business outcomes on a defined timeline. This is where Tableau becomes essential, not as a reporting tool, but as the shared decision layer that makes those insights visible across the organization.

Design a targeted portfolio of data products

Once an organization has identified the decisions that matter most, it can begin designing the data products required to support those decisions.

For each high-value decision, Val defines a data product that delivers the core insights needed to answer key business questions reliably and at the right operational cadence.

Within this model, Salesforce acts as the operational container for these data products, bringing together curated customer data and recommended actions. Tableau then serves as the decision layer, where executives track portfolio health, product leaders monitor customer segments, and frontline managers see performance against targets. Agentforce handles AI-assisted workflows, while Tableau ensures every stakeholder can see the same truth at the pace their decisions require.

Examples within financial services include:

Customer Growth & Attrition Hub
A consolidated view of customer lifetime value, churn risk, and next-best-product recommendations at the household level. Updated daily and integrated into Salesforce to guide relationship management and targeted campaigns.
Credit Risk Early Warning Hub
Portfolio monitoring capabilities that track roll-rates, vintage performance, and regional stress indicators, with drill-down visibility to the individual account level for underwriting and credit teams.
Fraud & Anomaly Watchtower
Near-real-time detection of suspicious behavior, generating alerts and investigation queues to help fraud teams respond quickly to emerging threats.

Each data product is defined with clear ownership, service-level expectations, level of detail, and delivery mechanisms whether through CRM systems, BI platforms, APIs, or automated alerts. Importantly, these are not simply reporting tools. They are operational data products designed to provide consistent answers to recurring business questions.

A data product is a self-contained, reusable asset that packages data with metadata, pipelines, and interfaces to deliver actionable value for specific business needs. It combines raw or processed data (e.g., datasets, ML models, dashboards) with governance, discoverability features, and quality standards to serve users like analysts or AI systems without exposing underlying complexities. Data products boost reusability across teams, reducing duplication and costs while ensuring trust through built-in quality, versioning, and interoperability.

But delivering value requires more than building the product. The insights must be accessible and embedded into the workflows where business decisions actually occur.

Make insights accessible and actionable

The greatest value from data emerges when everyone who needs to make decisions can see the same metrics, updated in real-time, without waiting for IT or analysts. Val approaches this challenge by designing data products with consumption and usability as core requirements.

Shared decision visibility through Tableau: Executives, product leaders, and risk managers see the same core metrics (customer health scores, portfolio risk indicators, operational efficiency) but filtered to their role and refreshed at the cadence their decisions require. No more version control issues or "which report is current?"

Action embedded where work happens: Key indicators flow into Salesforce where business teams work daily, with automated alerts when thresholds are crossed. Tableau handles the "why is this happening" analysis; Salesforce handles the "what do I do about it."

Close the loop with experimentation and AI

Effective data strategies create a continuous learning cycle that improves those decisions over time.

As Val notes: “For every decision, organizations need a mechanism for testing and learning. The learning component is often the most overlooked part of data programs. The goal is simple: increase the activities that deliver results and reduce those that don’t. Markets evolve, and both the data foundation and the models that support decisions need to adapt accordingly.”

This is what separates data initiatives that deliver sustained business value from those that gradually lose relevance. In practice, this approach includes:

AI-assisted decisioning

Agentforce can leverage these data products to recommend actions, guide frontline employees, and automate routine responses. At the same time, Tableau provides performance monitoring so leaders can evaluate the impact of experiments and model-driven decisions.

Through this feedback loop, organizations move beyond static reporting and toward adaptive decision systems that continuously improve performance. Each iteration produces new insights. Initial models provide direction, but repeated testing and refinement allow organizations to optimize decisions for their specific markets and customers.

Over time, the data products evolve from reporting tools into living operational capabilities—embedded within Salesforce, informed by AI through Agentforce, and monitored through Tableau—continuously improving the way the organization makes decisions.

The biggest obstacle: Breaking the data-first habit

The hardest part of implementing this approach? Getting organizations to break a deeply ingrained habit.

"I've seen companies spend months perfecting their data governance framework before asking a single business question. By the time they're ready to deliver insights, the business priorities have shifted. The executives who championed the initiative have moved on. And the ROI case has evaporated."

The cultural shift from "let's get our data house in order first" to "what decision are we trying to make" requires executive sponsorship and discipline. It means saying no to comprehensive data cleansing projects in favor of targeted fixes that support specific decisions. It means accepting that 80% data quality is often enough if it's the right 80%.

Organizations that succeed start small, pick one decision, build one data product, show value, then expand. The companies that struggle are trying to boil the ocean before they've proven they can boil a cup of water.

What this means for your organization

This approach fundamentally shifts how organizations think about their data investments. Instead of asking "How do we improve our data quality?" the question becomes "What decisions would better data enable?"

AI can handle the grunt work: the pattern recognition, the data cleaning, the routine analysis. But that elevates the importance of human judgment. The relationships you build, the strategic questions you ask, the context you bring. That's where competitive advantage lives.

"Data strategy is at its best when it becomes invisible. When decision-makers don't think about 'using data' because insights are simply part of how they work. Information woven so seamlessly into the business that better decisions happen naturally."

That's the goal. Not more dashboards. Not more governance committees. Better decisions, made faster, with confidence.

What questions do you have of your data (and Val)

A question for you: Can your leadership team name the top five decisions that drive your business outcomes? Can they tell you what data supports those decisions today?

If there's hesitation in either answer, you're not alone, and you're leaving value on the table.

Ready to start the conversation? Book time with Val to explore how a decision-first approach could transform your data strategy.