Three Things Every Data Leader Should Kill Before Building Another Dashboard

Introduction

Every data leader has been there: another stakeholder wants another dashboard, promising that this one will finally bring clarity. But the more you build, the less anyone trusts them.

According to former Worldpay and FIS data leader Phil Thirlwell, the root problem isn’t the dashboards themselves, it’s the culture, the habits, and the incentives behind them.

“At one point there were 600 Power BI dashboards. You know for a fact they were not all getting updated, probably some of the people that had built them were no longer with the business, but they’re still out there.”

In this episode of Data Matas, Phil joins host Aaron Phethean to explore what data teams should stop doing before they can truly make an impact.

You’ll learn:

• Why dashboard sprawl erodes confidence instead of building it
• How to escape the “service desk” trap and focus on outcomes
• Why trimming KPIs improves alignment and decision quality
• How co-developing metrics with engineers builds trust
• What “conversational BI” really means for the future of analytics
• This isn’t about tools or trends, it’s about rebuilding trust, one decision at a time

Meet Phil Thirlwell

Phil Thirlwell spent nearly a decade leading analytics and data strategy across Worldpay and FIS, two of the biggest names in financial technology. Now operating as a fractional data and analytics leader, he helps mid-sized firms modernise their data practices without the overhead of a full-time Chief Data Officer.

“Most companies don’t need a full-time CDAO on the payroll. They need specialist skills on demand, someone who can meet them where they are, solve the problems, and then step back.”

That perspective, pragmatic, hands-on, and outcome-led, runs through everything Phil does. He’s not interested in shiny platforms or long strategy decks. His focus is simple: get the right data to the right people at the right time.

Why This Challenge Matters Now

Dashboards were supposed to make data more democratic. Instead, they’ve multiplied to the point of chaos. In most organisations, the number of dashboards now outpaces the number of people who meaningfully use them.

The pressure to “move fast” and “serve the business” has turned many data teams into reporting factories, busy, but not impactful. As Phil explains, the result is an illusion of productivity: lots of activity, very little clarity.

“It feels good to be answering those questions because it feels like you’re giving immediate value, right? But you’re not really producing scalable solutions that are maybe getting to the heart of the problem.”

Common pitfalls include:

• Building dashboards on demand instead of by design
• Treating ad-hoc requests as success metrics
• Measuring everything that moves, instead of what matters

It’s not a tooling problem, it’s a trust problem. Leaders don’t trust the data because they don’t trust how it’s produced, shared, or understood.

How to Rebuild Trust Before Building Anything Else

1. Kill Dashboard Sprawl

When every team can spin up dashboards in minutes, the result isn’t insight, it’s overload.

“At one point there were 600 Power BI dashboards. You know for a fact they were not all getting updated, probably some of the people that had built them were no longer with the business, but they’re still out there.”

Phil’s advice is to treat dashboards like products: each one needs an owner, a purpose, and an end-of-life date.

Implementation guidance:

• What to do first: Audit what exists. Archive anything with zero usage in 90 days.
• Tools or structures: Track usage analytics (Power BI, Looker, Tableau) to see what’s actually opened.
• Watch-outs: Don’t delete dashboards blindly – communicate with the user base before retiring them.
• Expected benefit: A smaller, cleaner set of dashboards people actually trust and use.

2. Kill the Service-Desk Mentality

Most data teams pride themselves on being helpful. The problem? That helpfulness often comes at the cost of focus.

“It feels good to be answering those questions because it feels like you’re giving immediate value, right? But you’re not really producing scalable solutions that are maybe getting to the heart of the problem.”

Phil advocates running the data function more like a product team. Instead of reactive ticket queues, teams should operate in sprints with clearly prioritised outcomes.

Implementation guidance:

• What to do first: Replace ad-hoc requests with structured intake forms that link each request to a business outcome.
• Tools or structures: Use agile boards (Jira, Monday.com) to plan deliverables per sprint.
• Watch-outs: Saying “no” to low-value requests requires political backing from leadership.
• Expected benefit: Fewer distractions, faster delivery, and stronger alignment with company goals.

3. Kill KPI Overload

“It’s definitely more or less when it comes to dashboards and even KPIs, they cease to be key performance indicators if you’ve got too many of them, right?”

Phil’s point is brutally simple: too many metrics dilute focus. Dashboards packed with KPIs might look impressive, but they leave decision-makers wondering what actually matters.

Implementation guidance:

• What to do first: Limit executive dashboards to fewer than 10 KPIs.
• Tools or structures: Define each metric’s owner and source in a shared data catalogue.
• Watch-outs: Don’t aim for “perfect” definitions; aim for consistent ones.
• Expected benefit: Decision-makers aligned around the same numbers and language.

4. Co-Develop Metrics to Build Trust

True data literacy doesn’t come from dashboards; it comes from collaboration.

While not a direct quote, Phil describes how involving development leaders and business users in defining KPIs transformed engagement. When people understand how a metric is built, they take ownership of it, and that ownership strengthens governance naturally.

Implementation guidance:

• What to do first: Run KPI definition workshops across departments.
• Tools or structures: Use shared documentation tools (Notion, Confluence, dbt Docs) for transparency.
• Watch-outs: Avoid over-engineering – focus on the 20% of metrics that drive 80% of decisions.
• Expected benefit: Higher engagement, cleaner data, and improved cross-team trust.

5. Simplify Before You Automate

Automation is seductive, but Phil warns that it can easily multiply chaos.

“AI and automation amplify whatever’s already there. If your pipelines are messy, you’ll just get to chaos faster.”

The smartest teams start small: simplify their data model, document transformations, and only then introduce automation layers.

Implementation guidance:

• What to do first: Run a “simplify audit” – remove unused tables, pipelines, or scripts before optimising.
• Tools or structures: Lightweight lineage tools (OpenLineage, dbt) help visualise dependencies.
• Watch-outs: Don’t assume automation equals improvement – measure quality, not just speed.
• Expected benefit: More reliable pipelines, lower maintenance effort, and better cost control.

Putting It All Together

Rebuilding trust in data isn’t about rolling out new tools, it’s about stopping what doesn’t serve the business.

Start by auditing dashboards and KPIs, then co-design what’s left with the people who use it. Shift from “how fast can we deliver” to “how clearly can we decide.”

Metrics of success:

• 50% reduction in redundant dashboards
• Consistent KPI definitions across departments
• Fewer ad-hoc requests hitting the backlog

The result is a data function that feels smaller but performs bigger.

Real-World Impact

At FIS, Phil’s team didn’t just cut dashboards, they cut confusion. By consolidating hundreds of reports into a handful of trusted products, engagement jumped, and decision-making became faster and more transparent.

The lesson: clarity builds momentum.

Your Next Move

Audit what you’ve built. Retire what isn’t used. Reconnect with your stakeholders and ask what decisions they actually need to make. Then start building from there.

Every dashboard, metric, and model should earn its place by improving how people decide, not how much data they can see.

Listen to the full conversation with Phil Thirlwell on Data Matas for the complete breakdown of how to get there.

Dig Deeper

🎧 Listen: Three Things to Kill Before You Build Another Dashboard
📺 Watch: YouTube.com/@Matatika
👤 Connect with Phil Thirlwell on LinkedIn
🌐 More episodes: Matatika.com/podcasts

#Blog #Data Engineering #Data Infrastructure #DataStrategy #ETL #ETL Tools #KPI overload