Data-Driven Decisions: From Gut Feeling to Evidence

Data-Driven Decisions: From Gut Feeling to Evidence

"We decide based on data" – a claim many companies make. But what does it really mean? And why do so many data strategies fail? This article shows the path from intention to reality.

Table of Contents

1. What Does "Data-Driven" Really Mean?
2. Data Maturity Levels: Where Does Your Company Stand?
3. Dashboards Without Impact: Why Data Strategies Fail
4. The 5 Most Important KPIs for Every Industry
5. From Data Collection to Actionable Insight
6. The Difference Between Data Consulting and Data Implementation
7. Tools: BI Systems Compared
8. Your Path to a Data-Driven Company

---

What Does "Data-Driven" Really Mean?

Being data-driven doesn't mean having data. It means making decisions based on data – consistently and systematically.

Gut Feeling vs. Data: A Comparison

Gut-feeling decision:

• "I think our customers want feature X"
• "That worked last year too"
• "My gut tells me Q4 will be strong"

Data-driven decision:

• "73% of support tickets concern feature area X"
• "Conversion rate is 23% higher with variant A than B"
• "Based on trends from the last 24 months, we forecast +12% for Q4"

The Three Levels of Data-Driven Decisions

1. Descriptive: What happened?

• Revenue last week: €127,000
• New customer rate: 18%
• Average processing time: 4.2 days

2. Diagnostic: Why did it happen?

• Revenue decline correlates with ad campaign ending
• High new customer rate from referral program
• Long processing time due to holiday season

3. Predictive: What will happen?

• If we extend campaign: +15% revenue expected
• Churn probability for segment A: 24%
• Inventory lasts 18 more days at current sales rate

The Maturity Check

Most companies are stuck at "Descriptive." Real competitive advantage only comes at higher levels.

---

Data Maturity Levels: Where Does Your Company Stand?

Level 1: Data Chaos

Characteristics:

• Data scattered across Excel, emails, folders
• Every department has its own "truth"
• No unified definitions
• Reports compiled manually

Typical statements:

• "The numbers don't match"
• "Where do I find the current version?"
• "That takes time, I need to gather the data first"

Risk: Wrong decisions due to wrong or outdated data

Level 2: Data Silos

Characteristics:

• Departments have their own systems
• Data is structured but isolated
• No cross-departmental reporting
• Manual data transfers between systems

Typical statements:

• "You need to ask the other department"
• "Our CRM shows something different than accounting"
• "Linking the data is a project"

Risk: Inefficiency and incomplete picture

Level 3: Data Warehouse

Characteristics:

• Central data storage
• Unified definitions (Single Source of Truth)
• Automated reports
• BI tool in use

Typical statements:

• "Look at the dashboard"
• "Data is updated overnight"
• "We have standardized KPIs"

Risk: Data available but not used for decisions

Level 4: Data-Driven

Characteristics:

• Decisions are systematically data-based
• Self-service analytics for all departments
• Real-time data where needed
• Data culture anchored in the company

Typical statements:

• "What data supports this hypothesis?"
• "Let's A/B test that"
• "The metrics show we need to adjust"

Level 5: Data-Intelligent

Characteristics:

• Predictive and prescriptive analytics
• AI-supported decision support
• Automated anomaly detection
• Continuous experimentation

Typical statements:

• "The model predicts..."
• "The AI detected an anomaly"
• "Based on predictions, we're adjusting"

Self-Test: Where Do You Stand?

Question	Yes	No
Is there a central data source for key KPIs?	Level 3+	Level 1-2
Are decisions regularly justified with data?	Level 4+	Level 1-3
Can business departments create their own analyses?	Level 4+	Level 1-3
Do you use predictive models?	Level 5	Level 1-4
Is "What do the data say?" standard in meetings?	Level 4+	Level 1-3

---

Dashboards Without Impact: Why Data Strategies Fail

The uncomfortable truth: Most data strategies fail. Not because of technology – but because of implementation.

Failure Pattern 1: Dashboard Graveyards

The pattern:

1. Company buys BI tool
2. IT builds 20 dashboards
3. Everyone looks at them initially
4. After 3 months: 80% of dashboards no longer used

Why it happens:

• Dashboards show data nobody needs
• No connection to decision processes
• Too complex to provide quick answers
• Data isn't trustworthy

Failure Pattern 2: Analysis Paralysis

The pattern:

1. Team analyzes data intensively
2. More and more questions arise
3. "We need more data"
4. Decision gets postponed

Why it happens:

• No clear question at the start
• Perfectionism over pragmatism
• Fear of wrong decisions
• Data becomes an excuse

Failure Pattern 3: Vanity Metrics

The pattern:

1. Team tracks impressive numbers
2. "We have 100,000 website visitors!"
3. But: No correlation to business success
4. Wrong optimization on irrelevant metrics

Why it happens:

• KPIs not aligned with business goals
• Easy-to-measure things get measured
• Complex, important metrics are ignored

Failure Pattern 4: Technology Without Culture

The pattern:

1. Company invests in data warehouse, BI tool, data scientists
2. Technology works
3. But: Decisions still made by gut feeling
4. "The numbers are nice, but I know better"

Why it happens:

• No buy-in from leadership
• No incentives for data-based decisions
• Data culture not built
• Lack of data literacy

The Way Out

1. Question Before Data Not: "What can we do with our data?" But: "What questions do we need to answer?"

2. Fewer, Better Metrics 5 relevant KPIs beat 50 irrelevant dashboards.

3. Link Data to Decisions Every important decision needs a data basis – and gets documented.

4. Leadership as Role Model If management decides "by gut feeling," so does everyone else.

---

The 5 Most Important KPIs for Every Industry

Not every metric is a KPI. A KPI (Key Performance Indicator) is a metric directly linked to business success.

Universal KPIs for All Companies

1. Customer Acquisition Cost (CAC) What does it cost to acquire a new customer?

CAC = Marketing and Sales Costs / Number of New Customers

Benchmark: Should be <1/3 of Customer Lifetime Value.

2. Customer Lifetime Value (CLV) How much value does a customer generate over the entire relationship?

CLV = Average Order Value × Purchase Frequency × Customer Lifespan

Benchmark: CLV:CAC ratio should be >3:1.

3. Net Promoter Score (NPS) How loyal are your customers?

NPS = % Promoters (9-10) - % Detractors (0-6)

Benchmark: >0 is good, >50 is excellent.

4. Employee Satisfaction (eNPS) How engaged are your employees?

Benchmark: >20 is good, >40 is excellent.

5. Cash Flow / Runway How long can you continue at current burn rate?

Runway = Cash / Monthly Burn Rate

Benchmark: At least 6 months, better 12+.

Industry-Specific KPIs

E-Commerce

KPI	Formula	Benchmark
Conversion Rate	Orders / Visitors	2-4%
Average Order Value	Revenue / Orders	Industry-dependent
Cart Abandonment Rate	Abandoned / Started Carts	<70%
Return Rate	Returns / Orders	<20%

SaaS / Subscription

KPI	Formula	Benchmark
Monthly Recurring Revenue (MRR)	Sum of all monthly subscriptions	Growth >10% MoM
Churn Rate	Lost Customers / Total Customers	<5% monthly
Net Revenue Retention	MRR with Upsells & Churn	>100%
Time to Value	Time until customer realizes value	<7 days

Manufacturing / Production

KPI	Formula	Benchmark
OEE (Overall Equipment Effectiveness)	Availability × Performance × Quality	>85%
First Pass Yield	Good parts on first try / Total	>95%
On-Time Delivery	On-time deliveries / Total	>95%
Inventory Turnover	Cost of Goods / Avg Inventory	Industry-dependent

Services

KPI	Formula	Benchmark
Utilization Rate	Billable Hours / Available Hours	70-80%
Project Margin	(Revenue - Cost) / Revenue	>25%
Client Retention Rate	Returning Clients / Total	>80%
Revenue per Employee	Revenue / Employees	Industry-dependent

---

From Data Collection to Actionable Insight

Data alone is worthless. Value only emerges when data becomes insights – and insights become actions.

The Insight-to-Action Funnel

Level 1: Raw Data

• 1,247,832 records in the database
• No interpretation

Level 2: Information

• "Revenue in March was 15% lower than February"
• Description, but no explanation

Level 3: Insight

• "The revenue decline correlates with the end of the discount campaign and mainly affects new customers"
• Explanation and pattern recognized

Level 4: Recommendation

• "We should launch a targeted reactivation campaign for March buyers"
• Action recommendation

Level 5: Action

• Campaign is launched, results are measured
• Cycle begins again

Practical Techniques for Insights

1. So What? Test After every analysis ask: "So what? What does this mean for our actions?"

If you don't have an answer, the analysis isn't complete.

2. Segmentation Averages hide patterns. Segment by:

• Customer groups
• Product categories
• Time periods
• Channels

3. Comparisons Numbers without context are meaningless. Compare with:

• Previous periods
• Targets
• Benchmarks
• Segments

4. Correlation Analysis Which factors are connected?

• Customer satisfaction and repurchase rate
• Response time and churn
• Training and productivity

5. Root Cause Analysis Ask "Why?" 5 times to get to the actual cause.

Example:

• Revenue is declining. Why?
• → Fewer orders. Why?
• → Fewer website visitors. Why?
• → Less Google traffic. Why?
• → Algorithm update hurt ranking. Why?
• → Outdated SEO strategy. → Actionable!

---

The Difference Between Data Consulting and Data Implementation

Another uncomfortable chapter.

What Traditional Data Consulting Delivers

Typical Deliverables:

• Data Strategy Document (80 pages)
• Data Governance Framework
• Tool recommendations
• 3-year roadmap
• Data Maturity Assessment

Typical Cost: €50,000-200,000

Typical Result After 12 Months:

• Documents exist
• No data warehouse implemented
• No new dashboards in use
• No changed decision processes

What Data Implementation Delivers

Typical Deliverables:

• Working data warehouse
• 5-10 productive dashboards
• Automated data pipelines
• Trained users
• Documented processes

Typical Cost: €30,000-100,000

Typical Result After 12 Months:

• Decisions are made data-based
• Teams use dashboards daily
• Time savings from automated reports
• First predictive analyses

The Fundamental Difference

Data consulting tells you what to do. Data implementation does it.

Like all consulting topics: A mediocre strategy that gets implemented beats a perfect strategy in a drawer.

---

Tools: BI Systems Compared

For Beginners and Small Teams

Metabase

• Strengths: Open source, easy installation, intuitive interface
• Weaknesses: Limited enterprise features
• Price: Free (self-hosted), from $85/month (cloud)
• Suitable for: Startups, small teams, first BI steps

Google Looker Studio (formerly Data Studio)

• Strengths: Free, good Google integration, easy to share
• Weaknesses: Limited data modeling, performance with large data
• Price: Free
• Suitable for: Marketing teams, Google ecosystem users

For Growing Companies

Tableau

• Strengths: Powerful visualizations, large community, many connectors
• Weaknesses: Learning curve, expensive for larger teams
• Price: From €70/user/month
• Suitable for: Analysts, data-intensive companies

Power BI

• Strengths: Microsoft integration, affordable, extensive features
• Weaknesses: Complex, best experience only in Microsoft ecosystem
• Price: From €9.40/user/month
• Suitable for: Microsoft environments, cost-conscious companies

Superset

• Strengths: Open source, very flexible, SQL-first
• Weaknesses: Technical setup required, less polished
• Price: Free (self-hosted)
• Suitable for: Technical teams, companies with DevOps capacity

For Enterprise

Looker

• Strengths: Strong data modeling (LookML), good governance
• Weaknesses: Expensive, complex setup
• Price: On request (typically €50,000+/year)
• Suitable for: Large companies with data teams

Qlik Sense

• Strengths: Associative engine, strong self-service capabilities
• Weaknesses: Older architecture, UI no longer modern
• Price: On request
• Suitable for: Established companies with complex requirements

Our Recommendation by Company Size

Company Size	Recommendation	Why
Startup (<20 employees)	Metabase	Quick start, free, grows with you
SMB (20-200 employees)	Power BI or Metabase	Value for money, good features
Mid-market (200-1000 employees)	Tableau or Power BI	Scalable, enterprise features
Enterprise (>1000 employees)	Looker or Tableau	Governance, scaling

---

Your Path to a Data-Driven Company

The 90-Day Plan

Days 1-30: Foundation

• Conduct data maturity assessment
• Define top 5 business questions
• Identify and document data sources
• Select and set up BI tool
• Build first dashboard for one core question

Days 31-60: Expansion

• Connect additional data sources
• Create 3-5 more dashboards
• Set up automated reports
• Identify and train power users in departments
• Make first decision explicitly data-based

Days 61-90: Anchoring

• Enable self-service for business departments
• Set up data quality monitoring
• Establish weekly "Data Review" meetings
• Document and communicate quick wins
• Plan next phase

The Biggest Levers

1. Start Small One dashboard that's used is worth more than ten gathering dust.

2. Answer Real Questions Not: "What can we track?" But: "What do we need to know?"

3. Involve Leadership If the CEO looks at a dashboard every Monday, so does everyone else.

4. Celebrate Successes Every data-based decision that proves right should be communicated.

5. Be Patient Data culture doesn't happen overnight. Expect 12-18 months for real change.

---

Conclusion

Becoming data-driven isn't a technology project – it's a culture change. The tools matter, but are secondary. What counts:

1. Ask the right questions – don't collect data for data's sake

2. Implementation over perfection – a simple dashboard that's used beats a perfect data platform nobody uses

3. Leadership as role model – data culture starts at the top

4. Continuous improvement – not a one-time project, but a journey

The companies that will lead in 5 years aren't those with the most data – but those best at turning data into decisions.

---

Want to establish data-driven decisions in your company? Talk to us – we don't just build dashboards, we create data culture.