That quote was dropped in a recent LinkedIn debate I was part of. And you know what?

I agree.

But instead of being an argument against data teams, to me it's one of the most compelling arguments for self-service.

Because self-service isn’t just a buzzword or a dream. It’s the only operational model that scales data work in a way that actually matches how companies move, learn, and decide.

Full stop.

Let’s break down why that is, and why I’ll happily die on this hill.

What Even Is Self-Service?

Let’s get our terms straight.

When I say "self-service," I don’t mean stakeholders writing raw SQL against a production replica at 2am while sipping wine and hoping for the best.

Self-service means:

People can answer and iterate on their own questions (safely), without waiting on the data team, and can reliably take action based on what they learn.

That’s it. And that’s everything.

It doesn’t mean “no data team.” It means “the data team stops being the bottleneck for every single insight.”

Why It Matters

Here’s a hard truth I’ve learned working with 50+ organizations - from VC-backed scaleups to traditional SMEs to large, multinational enterprises:

If you want your data team to have any strategic leverage, some level of self-service is non-negotiable.

Yes, even in orgs where “nobody knows Excel.”

Yes, even when the company is just starting out.

No, not everything needs to be decentralized on day one. But from day one, you should be laying the groundwork for stakeholders to learn, explore, and decide without waiting in a Jira queue.

The 5 Ways Data Teams Create Value and How Self-Service Enables Each One

Let’s walk through the five business outcomes a great data team drives , and how self-service supports each one.

1. Increase in a company's revenue, growth, or market share

Self-service can drive top-line growth in two ways:

• When self-service is the product. Think: external users paying for access to embedded dashboards, APIs, or insights (common in B2B SaaS).
• When self-service enables customers to realize more value from your product. For example, retention platforms where customers can analyze their own funnel data - leading to stickier usage, upgrades, and expansion without increasing support burden.

(Anything like “marketing makes more money because they used self-service to better understand customer churn” is real, but that’s enabling revenue, so it belongs in Outcome #5.)

2. Decrease in a company's cost

Self-service is a cost-killer in the best way:

• Lower cost per answered question: stakeholders self-serve routine insights without analyst time.
• Less repetitive “reporting work” for the data team: fewer manual refreshes, one-off pulls, and recurring ad hoc requests.
• Faster decision-making reduces operational drag: shorter cycles mean less wasted time coordinating, waiting, and re-aligning.
• Reduced duplicated work across teams: fewer parallel spreadsheets / shadow metrics when people use shared self-serve sources.

This isn’t (just) about cutting headcount. It’s about freeing up time for higher-leverage work - and eliminating the overhead of coordination theater.

3. Enhancement of a company's reputation or brand

Harder to quantify - but not absent.

In cases where customers or partners use your analytics (via embedded dashboards or portals), self-service becomes part of the user experience. And fast, reliable answers - without escalation - translate into professionalism and trust.

However, most brand and reputation improvements (e.g. data consistency, auditability, GDPR compliance, data security) don’t come from self-service itself. They’re driven by governance and infrastructure - the foundations that enable safe self-service.

So while it’s not the primary lever here, self-service still plays a supporting role especially in externally-facing data products.

4. Strengthening of relationships between critical organizational stakeholders

This one is big.

Self-service reshapes the dynamics between the data team and the business:

• Stakeholders feel unblocked. They don’t have to wait in line to do their job.
• Data teams get to partner, not just serve. Less “please pull this,” more “help us think this through.”
• Ownership shifts. When domain teams can self-serve, they take more accountability for their own outcomes. In other words: The people who make decisions based on data have "skin in the game"
• Friction drops. Fewer angry pings, fewer missed handoffs, fewer dropped tickets.

Done right, self-service transforms the data team from a ticket queue into a multiplier.

5. Increase in organizational knowledge and capability of achieving organizational goals

This is the heart of it.

Self-service supercharges how an org learns, iterates, and aligns:

• Data in the hands of domain experts: the people closest to the problem can answer their own questions.
• Higher-quality decisions through proximity + context: interpretation happens with full business context, reducing wrong conclusions.
• Better questions and deeper exploration: self-service shifts behavior from requesting reports → investigating drivers and segments.
• Faster learning loops / experimentation behavior: quicker validation of hypotheses encourages test-and-learn.
• Transparency and trust in numbers: people trust and use what they can inspect and understand.
• Scalability of decision-making: more teams can operate analytically without waiting for centralized support.
• Resilience / reduced single points of failure: knowledge moves from individuals into shared assets (datasets, definitions, documented logic).

This is how real data cultures get built: not by sending dashboards, but by empowering decisions.

⚠️ But What About…?

You’re probably thinking: “What about early-stage companies? What about low data literacy? What if nobody wants to self-serve?”

I get it.

Even in very old-school small companies where people barely use Excel, I’ve implemented some form of self-service. Often it starts small: letting a team explore a BigQuery-connected Google Sheet. And it grows from there.

Yes, early on, things will be more centralized. That’s fine. But even then, I usually start planting seeds for decentralization:

• Simple semantic layers
• Light training for Super Users
• Early exposure to tools they’ll one day own

Because I don’t want to build a cathedral of dependencies I’ll just have to dismantle later.

A Final Note on Nuance

Is self-service black and white? No.

There are shades. Levels. Maturities.

You don’t need Looker + dbt + a metric layer + lineage + observability just to get started.

But you do need to treat self-service as the goal, not an optional feature. Because the longer you delay, the more brittle, expensive, and frustrating your operating model becomes.

What To Take Away

Let’s wrap with some actionables:

• Reframe your goal. My self-service success metric: 80% of requests can be self-served by business decision makers independently and immediately
• Audit the queue. If you’re spending 80% of your time on requests that stakeholders could answer with access + training, that’s your roadmap.
• Read next week's newsletter. Next week, we’ll get tactical.

👉 This week was about why self-service is non-negotiable.

👉 Next week: how to actually make it work (without throwing your team into chaos or handing SQL to the marketing intern).

See you then.

Cheers,
Sebastian