From Gut Feel to Good Data: Building the Foundation Your Planning Process Runs On 

Key Takeaways

• Your planning cadence is only as good as the data feeding it. Structure plus bad data equals organized guessing. 
• You already have most of what you need—it’s just scattered, stale, or stuck in someone’s head. 
• Don’t chase perfect data. McKinsey found that no company with a deployed planning system reported perfect data—yet many still saw strong results by applying the 80/20 rule. 
• Start by measuring the Confidence Gap—the distance between what your systems say and what your team actually believes. Close it on five or six core data elements and the whole cadence transforms. 
• The data habits you build now are the foundation for AI-powered planning later. Low data quality the top barrier to scaling AI in supply chains. 

From Gut Feel to Good Data: Building the Foundation Your Planning Process Runs On

You took the advice from Part 1 in this series. You’ve set up the monthly planning meeting with the right people in the room. You’ve laid out the four steps. The first demand review starts strong. 

Then someone asks a simple question: “How much of Product X did we actually ship last month?” 

Three people give three different answers. One pulls from the ERP. One has a spreadsheet they’ve been maintaining on the side. One remembers “roughly” from a conversation with the warehouse manager. 

The meeting doesn’t collapse. It does something worse—it drifts. The next 30 minutes are spent debating whose numbers are right instead of deciding what to do about the gaps. By the time the team aligns on the data, they’ve burned through the meeting time and the team’s energy. 

The planning cadence didn’t fail because the structure was wrong. It stalled because the data feeding it couldn’t be trusted. 

When the data is unreliable, planning meetings degrade into one of two failure modes. In the first, the team debates every number and defers every decision until someone can “go check.” In the second, the team ignores the data entirely and reverts to gut feel—which is exactly what the planning process was designed to replace. 

Both modes share the same root cause, what we call the Confidence Gap—the difference between what your systems say and what your team actually believes. 

The ERP says you have 400 units on hand. The warehouse manager says “it’s closer to 300 once you account for the quality holds and the orders picked but not yet shipped.” A lead time is listed as three weeks in the system, but everyone knows that supplier runs five. That’s the Confidence Gap in action. 

Every business has one. It’s not a sign of failure—it’s the natural consequence of systems that were set up years ago and haven’t kept pace with changes in how businesses operate. The good news is, it’s measurable, and it’s fixable without a six-month data project. 

Gartner estimates that poor data quality costs the average organization $12.9 million per year. For an individual mid-market business, their loss may look manageable, but the proportional impact is often greater—because you have fewer people and less margin to absorb the cost of decisions made on bad information. 

The Data You Already Have (And Where It’s Hiding) 

The monthly planning cycle from Part 1 runs on six core data elements. You almost certainly have all of them already. So the problem isn’t missing data—it’s that the data is scattered across systems, spreadsheets, and institutional knowledge that never makes it into the planning meeting in a form the team can trust. 

These core data elements include: 

• Demand history—what actually sold or shipped over the past 12 to 24 months at the product family or category level. For a manufacturer, that’s shipment history by customer. For a distributor, order and fill data by account. For a retailer, POS data by location or channel. This almost always exists in your ERP or order management system. But it’s buried in transaction records nobody’s aggregated into a planning-friendly view, so the team pulls it ad hoc each month and gets slightly different numbers depending on who runs the report. 

• Forward demand signal—what’s coming. Open customer orders, committed purchase orders, promotional plans, seasonal expectations, and project pipelines. This data exists but often lives in sales reps’ heads, email threads, or scattered CRM notes rather than somewhere the planning team can access and act on. 

• Inventory position—what you have on hand, where it sits, and what’s available versus committed. For multi-location distributors or retailers, the challenge multiplies—the product might exist in your network, just not in a single consolidated view that gets updated frequently enough to trust. 

• Open supply—what’s in the pipeline and when it’s expected to land. Purchase orders, production orders, and inbound transfers. The gap here is almost always lead time accuracy—the system shows a PO due date of June 15, but the buyer knows that supplier runs two weeks late, and nobody updates the system until the goods actually arrive. 

• Lead times and supplier reliability—the realistic view of how long it takes to get materials or finished goods, including known variability. These were often entered during system setup years ago and never revisited. Reality has changed since the numbers were established. 

• Capacity constraints—what can hold you back. Production lines, warehouse throughput, and shelf space. This almost always lives in people’s heads rather than anywhere the planning team can reference when making allocation or commitment decisions. 

None of this requires a new system. It requires someone to name the gap, close it, and agree on a single source. 

Six Data Traps That Keep Your Planning Process Running Blind 

If the five hidden costs of unstructured planning from Part 1 tell us why you need a cadence, these six data traps are can be what keeps that cadence from delivering on its promise. 

1. Multiple versions of truth. Sales tracks revenue in dollars. Operations tracks units shipped. Finance tracks invoiced amounts on a different calendar. So when the planning team sits down, they’re comparing apples to oranges without realizing it. Every disagreement feels like a judgment call, but it’s actually a unit-of-measure problem hiding in plain sight. The fix is to agree on one unit of measure and one source system for each core data element. It doesn’t have to be sophisticated, but it does have to be singular. 

2. Stale master data. Lead times, safety stock levels, and reorder points were configured when the system was implemented and haven’t been touched since. The business has changed—new suppliers, new products, different customer mix—but the planning parameters reflect past conditions, not the present. Better to build a quarterly review of key planning parameters into the cadence: fifteen minutes, once a quarter, assigned to one person. This is the single highest-leverage data maintenance activity a mid-market company can do. 

3. The shadow spreadsheet. Someone on the team doesn’t trust the system, so they maintain their own spreadsheet with “the real numbers.” Over time, it becomes the de facto planning system—but nobody else can access it, update it, or validate it. This is Part 1’s hero dependency problem expressed through data. To remedy, bring the shadow spreadsheet into the light. If it’s more accurate than the system, figure out why and fix the system. If it’s not, retire it. 

4. Confusing precision with accuracy. This is the counterintuitive one. A company invests time building a SKU-level forecast with decimal-point precision, but the underlying demand history is incomplete and the assumptions are stale. The result is a forecast that looks authoritative—polished spreadsheet, three decimal places—but is no more reliable than a gut estimate. Precision without accuracy is worse than a rough estimate everyone acknowledges is rough, because it creates false confidence. The better approach is to prioritize accuracy at the category level over precision at the SKU level, because a view that the whole team trusts will produce better decisions than a granular forecast nobody believes. 

5. Treating data quality as an IT project. When leadership decides “we need better data,” the instinct is to hand it to IT or launch a data cleansing initiative. Six months later, the project is stalled in requirements gathering, and the planning team is still using the same spreadsheets. Improving data quality for planning is an operations initiative, not a technology project. The people who know where the data is wrong are the planners, buyers, and warehouse managers who work with it every day. So keep ownership with the planning team—assign a data steward from operations while IT supports, but doesn’t own the effort. 

6. Waiting for perfect data before starting. The most dangerous trap, because it sounds responsible. “We need to clean up our data first” becomes the reason the cadence never gains traction. McKinsey’s 2024 Global Supply Chain Leader Survey found that data issues are the biggest bottleneck in supply chain digitization—but also found that none of the companies with deployed planning systems reported having perfect data, and many were satisfied with their results. The researchers recommend applying the 80/20 rule: Press forward once most data are available, and fix the gaps as the process exposes them. The planning cadence itself is the best data quality tool you have. 

The Data Trust Audit: Close Your Confidence Gap in Two Weeks 

This isn’t a data cleanup project. It’s a focused, two-week exercise designed to close the Confidence Gap on the data elements that matter most to your planning cycle. Think of the Data Trust Audit as the data equivalent of the planning cadence itself—a structured rhythm that replaces ad hoc scrambling with deliberate action. 

In week one, quantify it. Gather your planning team—the same people who attend the monthly cadence meeting—and evaluate each of the six core data elements across three dimensions.  

1.  Source clarity—is there one agreed-upon source, or does everyone pull from a different place? 

2. Freshness—how current is this data when it enters the planning meeting? 

3. Team trust—would the team act on this number without debating it? Score each dimension 1 to 5. The average is that element’s Confidence Score. Anything below a 3 is your priority. 

Picture what this actually looks like: Your demand planner, your buyer, your operations lead, and your sales counterpart sitting together for an hour, going element by element. “Do we trust our inventory position?” Differences come to light. That conversation—the one where the gaps between perceptions surface—is where the real value lives. You’re not just scoring data. You’re building alignment around what’s trustworthy and what isn’t. 

In Week Two, close it. For each priority element, take three actions.  

1. Designate the single source—one system, one report, one spreadsheet. If there are competing versions, pick the one closest to the transaction and retire the others. 

2. Update the stale inputs—if lead times or safety stocks haven’t been reviewed in six months, spend thirty minutes with the buyer or supplier manager who knows the current reality. 

3. Assign the steward—one person, by name, responsible for keeping this data element current between planning cycles. Not a full-time role. A standing commitment. 

The output is a one-page scorecard the team can reference and update quarterly. Tape it to the wall in the planning meeting room if you have to. Over time, Confidence Scores should trend upward—and that trend becomes a leading indicator that the planning process is maturing, just as a declining Chaos Index from Part 1 signals that the cadence is taking hold. 

This exercise also lays the groundwork for what comes next.  To take advantage of artificial intelligence (AI) in supply chains, your data foundation must be connected, contextualized, and continuous. The Data Trust Audit is the first step on that path. The data discipline you build now isn’t just about making the monthly meeting work—it’s the foundation that makes AI-powered planning possible. 

The Discipline Compounds 

The planning cadence from Part 1 gives you the rhythm. The data work in this post gives that rhythm something real to run on. Together, they represent the shift from “we have a planning meeting” to “we have a planning process”—one where decisions are made on shared facts rather than competing instincts. 

Neither installment required new software. Neither required a dedicated data team or a six-figure consulting engagement. Both required discipline—and the willingness to stop debating whose numbers are right and start agreeing on a single version of the truth. 

That discipline is about to pay off. In Part 3, we’ll cover what happens when the cadence is running and the data is flowing: How to recognize when you’ve outgrown spreadsheets, what to look for in planning technology that fits a lean team, and how to bring in tools that automate the manual work—without the complexity that derailed past attempts. 

The debates don’t disappear. They just stop being about the data—and start being about the decisions.

Frequently Asked Questions 

We have data in multiple systems that don’t talk to each other. Do we need to integrate them first? 

Not yet. Start by designating one source per data element and pulling the planning view together manually, even into a spreadsheet. Integration is a longer-term improvement, not a prerequisite. The Data Trust Audit will show you which elements to integrate first when you’re ready. 

Our ERP has data, but nobody trusts it. How do we fix that without a system overhaul? 

Trust doesn’t break across the board. It breaks on specific data elements for specific reasons. Run the Confidence Gap assessment. You’ll likely find two or three elements dragging down overall trust while others are fine. Fix the stale lead times, reconcile the inventory discrepancies, and/or update the safety stock parameters that haven’t been touched in years. Trust is rebuilt one element at a time. 

Who should own data quality if we don’t have a data team? 

Assign a data steward. This is the person who already knows the numbers best and is probably maintaining the shadow spreadsheet. Make the role official, give them time, and keep ownership within the planning team. IT supports with system access; operations owns the accuracy of what goes in. 

How do we know our data is “good enough” to plan on? 

If most core data elements score a 3 or above on the Confidence Gap scale—“we trust it directionally”—you’re ready. The cadence itself will surface where the remaining gaps hurt most. That tells you exactly where to focus next. Waiting for all 5s is the last trap mentioned in our list above. 

You mentioned AI-ready data. Do we need to worry about that now? 

This comes up often from companies at this stage, and it’s the right question to ask early. Not directly, no, but every step in this post moves you in the right direction. The need for data that’s connected, contextualized, and continuous is really describing what a mature version of the Data Trust Audit produces: data from agreed-upon sources, enriched with context from the planning cycle, and updated on a reliable cadence. You’re not building an AI pipeline. You’re building the habits that make one possible later.