Why Leaders Can't Trust Their Own Data (And What to Do About It)

It's Wednesday morning. Your COO asks for last month's revenue numbers. Finance pulls a report from the ERP: CHF 2.3M. Sales pulls from Salesforce: CHF 2.5M. Operations has a third number from their own spreadsheet. Nobody knows which one is right. Your leadership meeting turns into a data reconciliation session instead of a strategy discussion.

This isn't a technology problem. This is a trust problem. And I see it in nearly every Swiss B2B company I work with — from 50-person manufacturers to 400-person distributors. The systems work individually. But nobody trusts the numbers they produce together.

The Real Cost of Unreliable Data

When leaders can't trust their data, the effects go far beyond inaccurate reports. Here's what actually happens:

• Decisions slow down. Every number needs to be verified before anyone acts on it. A decision that should take 10 minutes takes a week because three departments need to confirm their version is correct.
• Teams build shadow systems. Your best people start maintaining their own spreadsheets "just to be sure." Now you have five versions of the truth instead of one.
• Revenue leaks. Quotes that should have become orders sit in a queue because the data between Salesforce and ERP doesn't match. Nobody notices until the customer calls.
• AI investments fail before they start. You can't train an AI model on data you don't trust. Companies that skip the data foundation end up with expensive tools that produce unreliable outputs.

A Swiss distributor I recently worked with estimated they were losing 8-12 hours per week across their leadership team just reconciling data from different systems. That's an entire working day — every week — spent verifying numbers instead of running the business.

Why This Happens (Even to Good Companies)

The root cause is rarely incompetence or bad software. It's usually one of three patterns:

Pattern 1: Systems Grew Organically

Your company started with an ERP. Then you added Salesforce for sales. Then a separate tool for inventory. Then one for project management. Each system was the right choice at the time. But nobody planned how they'd talk to each other.

The result: each system has its own version of customer data, product data, and transaction data. When they disagree — and they always do — there's no clear rule for which one wins.

Pattern 2: The "Quick Fix" Integration

At some point, someone connected Salesforce to the ERP. Maybe a point-to-point integration. Maybe a CSV export every night. It worked — until business rules changed, a field was renamed, or the volume of data grew beyond what the original design could handle.

Now you have an integration that sometimes works, sometimes doesn't, and nobody fully understands what it does or doesn't sync.

Pattern 3: No Single Source of Truth

The most common pattern. Ask your team: "For customer addresses, which system is correct — Salesforce or the ERP?" If nobody can answer confidently, you don't have a source of truth. You have competing opinions stored in databases.

---

How to Restore Data Reliability

Fixing this isn't about buying a new tool or starting a 12-month data governance project. It starts with three practical steps:

Step 1: Define Who Owns What

For every data type in your business, answer one question: which system is the source of truth?

• Customer master data → probably ERP
• Opportunities and pipeline → probably Salesforce
• Product pricing → ERP or a dedicated pricing system
• Order status → ERP

This sounds simple. It's not. It requires leadership alignment, not just IT decisions. When the COO says "Salesforce is the truth for customers" and the CFO says "No, the ERP is" — that's not a technical problem. That's a governance problem that needs executive alignment.

Step 2: Make Data Flow in One Direction

Once you know who owns what, data should flow from the source of truth to the consuming systems. Not both ways. Not sometimes one way, sometimes the other.

Customer created in ERP → syncs to Salesforce. Not the other way around. Price updated in pricing system → pushes to both ERP and Salesforce. Never edited directly in either.

This eliminates the "which number is right?" question. The answer is always: the one in the source system.

Step 3: Make Failures Visible

The worst data problems are the ones nobody knows about. A sync fails silently. A record gets duplicated. A price doesn't update.

Build visibility into your data flow:

• Dashboard showing sync status: what synced, what failed, what needs attention
• Automated alerts when error rates exceed a threshold
• Weekly data health report for leadership — not IT, leadership

When data problems become visible, they get fixed. When they're invisible, they compound.

A Diagnostic You Can Run This Week

Before investing in any tool or project, answer these five questions with your leadership team:

1. For each major data type (customers, products, orders, prices), which system is the source of truth? If people give different answers, start here.
2. When was the last time you compared numbers across systems? If the answer is "months ago" or "never," you likely have drift you don't know about.
3. How do you know when a sync fails? If the answer is "when someone complains," your failures are invisible.
4. How many people maintain their own spreadsheets "just to be sure"? Shadow systems are a symptom of distrust. Count them.
5. Could you give an investor accurate revenue numbers in under an hour? If not, your data reliability has a direct business cost.

If more than two of these questions make you uncomfortable, your data foundation needs attention — before any automation or AI project can succeed.