About This Episode
In episode 6 of Data & AI Heads, host Ian Allison sits down with Wyatt Larsen, a data and analytics leader with over 15 years of experience across energy, education tech, healthcare tech, and HR tech. Together they tackle one of the most persistent challenges in the field: how analytics teams earn the organizational trust needed to work on initiatives that actually move the business forward, rather than getting sidelined as a cost center.
Wyatt makes a compelling case for a slightly decentralized analytics model, where analysts embedded in business units gain deeper context and domain knowledge. He also introduces practical mechanisms for cross-functional learning, including peer code reviews and paper reading groups, that build trust across analyst communities without requiring a heavy top-down reporting structure.
The conversation goes deep on how analytics teams demonstrate and communicate value: framing insights so stakeholders feel ownership of conclusions, documenting downstream business impact, and using counter metrics to keep A-B test results and KPI improvements honest. Wyatt shares a candid personal story about a 30% revenue projection that delivered 10%, and why that gap underscored the importance of counter metrics before any result goes public.
Key Takeaways
- Analytics teams hold soft influence across the entire organization but have no formal budget authority, so building trust through relationships and discovery is the only real lever they have.
- A slightly decentralized model, with analysts or subject matter experts embedded in business units, gives those analysts the day-to-day context that a purely centralized team cannot replicate.
- Cross-functional peer feedback sessions and paper reading groups, kept to roughly seven or eight participants, create safe spaces for shared learning and build trust across the analyst community over time.
- Coming into a stakeholder conversation with a discovery mindset, rather than leading with 'your process is broken,' is critical because processes often exist for valid reasons and leading with criticism destroys trust immediately.
- When presenting an analysis, letting the stakeholder reach the intended conclusion on their own terms increases their buy-in and makes them more likely to credit and advocate for the analytics team in the future.
- Counter metrics are essential guardrails: optimizing one KPI in isolation, such as win rate or website views, can mask deterioration in a downstream metric like pipeline volume or checkout conversion.
- Projecting a 30% revenue lift from an A-B test and delivering 10%, even though 10% was historically strong, created organizational disappointment because expectations were set too high, illustrating why under-promising and over-delivering matters.
What We Cover
Frequently Asked Questions
How can a data analytics team build trust with business stakeholders?
According to Wyatt Larsen, the most effective approach starts with a discovery mindset rather than immediately telling stakeholders their processes are broken. Analytics teams build trust by embedding themselves in business units, understanding context, and presenting insights in a way that allows stakeholders to reach the right conclusions on their own terms. When a leader feels the insight was their own idea, they are far more likely to champion the analytics team's work going forward.
Should analytics teams be centralized or decentralized?
Wyatt Larsen favors a slightly decentralized model where analysts are embedded in business units, because proximity to daily operations gives them domain context a centralized team cannot easily replicate. He notes the right structure depends on context: a marketing attribution analyst requires a specialized skill set tied closely to one problem, while a finance department analyst may be a generalist. The risk of being too centralized is that the data team becomes an entity unto itself, optimizing its own processes rather than serving the business.
What are counter metrics and why do they matter for analytics teams?
A counter metric is a secondary measure that can reveal whether an improvement in a primary KPI is genuinely positive or is masking a problem elsewhere. Wyatt Larsen gives the example of a sales team whose win rate rose in one quarter, which looked positive until the data revealed the cause was a drop in client meetings, signaling a weaker pipeline for the following quarter. Tracking counter metrics prevents analytics teams from over-claiming value from an improvement that will hurt the business downstream.
How should data leaders communicate the value analytics teams deliver?
Wyatt Larsen recommends two complementary approaches. First, document the downstream business impact of each analysis, even when the credit goes to the decision-maker who acted on it. Second, when presenting insights, frame the conversation so the stakeholder draws the intended conclusion themselves. When stakeholders feel ownership of a decision, they naturally acknowledge the analytics team's role and seek more of their support on future decisions.
What is the risk of over-projecting results from an A-B test?
Wyatt Larsen shared a personal experience where an A-B test projected a 30% revenue increase, but the actual result was approximately 10%. Although 10% was historically strong, the gap created disappointment across the organization because expectations had been set too high. He advises analytics teams to account for counter metrics and variables that may not hold constant, and to frame projections with appropriate conditions rather than presenting a single optimistic headline number.
How do peer feedback sessions and reading groups help analytics teams grow?
Wyatt Larsen advocates for two specific practices: cross-functional code reviews or analysis feedback sessions, where an analyst presents work in progress and invites critique, and monthly paper reading groups where team members share interesting findings from research. He recommends keeping these groups between roughly four and eight people to maintain focus and avoid tangents. He notes that the session leader should model vulnerability by sharing their own work first, which encourages others to participate openly.
How should analytics teams handle business requests for access to PII or private data?
Wyatt Larsen notes that business units sometimes want access to private data because it would genuinely improve decision-making, even when privacy regulations restrict it. His recommended approach is to bring legal into the conversation directly, so the business can understand what use cases are permissible and under what conditions, rather than having the analytics team simply say no. Helping the business navigate that complexity is itself a trust-building activity that demonstrates the analytics team as a strategic partner rather than a gatekeeper.
About the Guest
Wyatt Larsen
Wyatt Larsen is a data and analytics leader with over 15 years of industry experience spanning energy, education technology, healthcare technology, and HR technology. He has worked across both embedded analyst and centralized team structures, giving him a nuanced perspective on how analytics organizations should be designed to maximize business impact. Wyatt is a strong advocate for cross-functional learning environments, counter-metric discipline, and influence-based approaches to stakeholder engagement. He is candid about hard-won lessons, including the organizational consequences of projecting results that outpace what an A-B test ultimately delivers.
Connect on LinkedIn →More Episodes