Beyond the dashboards & AI hype: Real data leadership insights with Jacob Bedard

‍

When Jacob Bedard builds data systems, he doesn't start with algorithms or analytics. He starts with recording what actually happens in the business.

On this episode of Crunching Data, Jacob, a distinguished data and analytics leader with over 16 years of experience, shares how the real work of data happens before any analysis begins. He opens up about AI's dangerous authoritative tone, why most companies don't need streaming data, and how to measure data team success beyond vanity metrics.

If you're leading a data team or rethinking your data strategy, this conversation offers essential insights from someone who's built data organizations at OpenBorder, Apollo Insurance, Dialpad, and Hootsuite.

Here are some key areas we dive into:

The Foundation Problem: It's Not About the Data

Jacob's most fundamental insight challenges how most organizations think about data work. As he puts it:

"If you don't record it, we can't report it. So if you simply have not captured a piece of information that you can't infer or deduce downstream, like it's gone, your ability to report on that, it's come and gone."

The real challenge isn't in building sophisticated models or creating beautiful dashboards. It's in ensuring that your systems actually capture the data necessary for meaningful analytics. Jacob's experience shows that the most successful data initiatives start with aligning business processes and systems. "You're doing a lot of work on changing their CRMs and other connected systems and then changing processes," he notes.

The key is making these processes trackable without creating unnecessary overhead for business teams.

‍

The AI Reality Check: Authority Without Accuracy

When discussing AI's role in data work, Jacob offers a refreshingly pragmatic perspective that cuts through the current hype. His concern isn't about AI's capabilities, but about its convincing authoritative tone.

"I think it's the same thing that is a problem when you have very authoritative tone being used by a person, right? It's the fact that they sound so correct... these models that people are using, they come up with very convincing answers."

The problem becomes more acute in data work, where there's "zero place for mistakes." Unlike engineering teams that can review AI-generated code through pull requests, data teams often work with more abstract problems where errors are harder to catch but can be devastating to organizational trust.

Jacob advocates for a more structured approach, similar to WolframAlpha's model, which combines natural language processing with deterministic functions rather than relying purely on statistical predictions.

Measuring Data Team Success: Beyond Vanity Metrics

One of the most insightful parts of our conversation focused on how to measure data team effectiveness. Jacob suggests that the percentage of time spent on recurring reporting serves as a crucial maturity indicator. Successful data teams automate routine work to free themselves for high-value, complex questions that emerge monthly.

This insight challenges the common practice of building hundreds of dashboards, most of which go unused. Instead, Jacob emphasizes the importance of starting with stakeholder interviews to understand real pain points before proposing solutions:

"You have to start with the stakeholder interview, understand the pain points and then make sure that what you're proposing is really about solving it, not hey, I really want this technology and it's gonna solve your problem... If you just want to start with the tool, you're toast. You have to start with the business."

‍

The Underrated Role: Business-Data Bridge Builders

Looking toward the future, Jacob identifies a critical gap in the data ecosystem: professionals who can bridge business processes with data collection. While data science has matured significantly and data engineering has seen major advances with tools like dbt, the intersection of source systems and business processes remains underserved.

For newcomers to the field, Jacob recommends developing skills that combine business knowledge, system architecture, and data expertise—essentially becoming translators between business needs and technical implementation.

The Streaming Data Myth

Our conversation concluded with Jacob debunking a persistent myth in data engineering: the universal need for streaming data:

"I really got to say streaming data. 999 times out of a thousand, you don't need to stream data from your source system to your destination. Usually it's just going to be a batch."

This reflects a broader pattern where tool selection drives decision-making rather than business requirements—a trap that even experienced teams can fall into.

Key Takeaways for Data Leaders

Jacob's insights offer several actionable principles for data leaders:

1. Focus upstream: The biggest wins come from improving data collection, not just analysis
2. Build trust first: Accuracy and consistency matter more than speed or sophistication
3. Start with stakeholders: Understand the problem before proposing solutions
4. Automate the routine: Free your team for high-value strategic work
5. Question the tools: Don't let technology choices drive business decisions

As data continues to play an increasingly critical role in business success, Jacob's experience reminds us that the fundamentals matter most. Building trustworthy, accurate, and useful data systems requires as much focus on people and processes as it does on technology.

The full conversation offers even more insights into data leadership challenges and solutions. It's essential listening for anyone looking to build more effective data organizations.

Watch the Podcast above or here

Watch Other Episodes of Crunching Data

• From Governance to Ecosystems: Rethinking Data Strategy with Dylan Anderson
• The Leadership Metric That Changed His Data Team’s Reputation with Jeff Baird
• Context, Curiosity, and Cutting Through the Noise with Qun Wei (Lemonade)

‍