Manufacturing Data

Your manufacturing data isn't perfect.
Neither is anyone else's.

Every manufacturer we've worked with came in with data problems. Some were obvious. Most weren't. This is what we've learned across hundreds of ERP environments, and why making progress from where you are matters far more than waiting until the data is clean.

8 min read Data quality ERP & operations

We have connected to hundreds of ERP environments across manufacturing portfolio companies. The data is always messy, not because the teams are careless, but because ERP systems were built to process transactions, not to anticipate ten years of acquisitions, floor workarounds, and configuration drift. Here are the fifteen data challenges we see most often, and what they actually cost you analytically.

Entity & master data
01
Same customer in five ERPs, no consensus on who they are

When a PE portfolio company acquires another manufacturer, it inherits a second customer master. Three acquisitions later, revenue by customer is a reconciliation exercise. Deduplication is manual, taxonomy is improvised, and no record is authoritative. Consolidated cross-site revenue visibility requires resolving records that were never designed to align.

Data Mastering · Sales IQ
02
The customer name field, used as the only status tool anyone has

"Acme Inc (Do Not Use)." "Meridian Attractions (CLOSED)." "Global Parts (Use 4782 Instead)." The ERP doesn't have a status field that flows cleanly into reporting, so teams write the status into the name. Your AR team may be chasing a company that shut down two years ago. Nobody did anything wrong. The ERP didn't give them a better option.

Data Mastering
03
Same part, different item number, different description, different language at every site

Cross-site inventory visibility requires knowing that Part A100 at Plant 1 and Part 100A at Plant 2 are the same component. Without a unified item master, consolidated purchasing spend, vendor consolidation negotiations, and cross-site demand planning all run on guesswork. Translations add another layer: the German plant's description for the same SKU may share nothing with the US record.

Data Mastering · Inventory IQ
Costing & labor
04
Standard costs and routing standards set at go-live and never touched since

The variance between what the ERP thinks a product costs and what it actually costs has been growing quietly since implementation. Without an accurate standard to compare against, cost analysis is an estimate dressed up as a calculation. Planned costs are often missing entirely, which means there is no baseline for variance analysis at all. You cannot measure improvement you cannot baseline.

Operations IQ · Pricing IQ
05
Time not logged to production jobs, so efficiency cannot be measured

Without hours posted to production orders, you know total payroll but cannot measure shop floor efficiency. The floor could be running at 70% or 95% of routing standard. The ERP has no way to tell you which, and neither does your operations report. Labor efficiency becomes a gut feel rather than a managed metric, and the gap between standard and actual stays invisible.

Operations IQ
06
The transaction that posted wrong and cannot be reversed

A receiving clerk processes a shipment of 1,000 units. The purchase order was written in cases. The ERP unit of measure is each. The clerk receives 1,000 (meaning 1,000 cases) against a UOM of 1 (each). The ERP accepts the transaction. Of course it does. Inventory is now overstated. Cost is exploded by the case-pack multiple. The posted transaction cannot be cleanly reversed, and every weighted average cost calculation downstream reflects it permanently.

Inventory IQ · Operations IQ
Pricing complexity
07
Pricing is a four-layer system nobody fully understands

Global list price, customer-specific price curves, long-term agreements, and pricing trade agreements interact in ways that make effective price visibility nearly impossible to maintain. What the customer actually paid versus what your pricing policy said they should pay often lives in a spreadsheet on someone's desktop, not in the ERP. The ERP records the transaction. It doesn't explain the logic behind it.

Pricing IQ
08
Every customer gets a discount, so there is no real list price

When every customer has a negotiated deviation from list, the list price loses meaning as an analytical benchmark. Price-volume-mix analysis needs a defensible starting point. Without one, the analysis reflects whatever proxy assumptions the analyst constructed. You end up measuring the deviation from a fiction, which doesn't tell you much about pricing power.

Pricing IQ
09
Replacement parts look like mix in PVM, but it is actually price

When a customer moves from a standard component to a higher-spec replacement or service part, PVM analysis classifies the margin change as a mix shift because the item number changed. The real story is a price change. The distinction matters when deciding whether to raise prices or rationalize the portfolio. Replacement and service parts need to be tagged and handled separately or the analysis gives you the wrong answer.

Pricing IQ
Sales & commercial
10
Open orders and quotes that age in the system indefinitely

A quote that should have expired in Q1 is still sitting in open backlog in Q4. Conversion rate analysis, pipeline forecasting, and backlog accuracy all depend on timely status updates that sales teams have no structural incentive to provide. The ERP backlog is technically accurate. Practically, it is not, and measuring sales velocity against it produces results nobody trusts.

Sales IQ
11
CRM is not connected to ERP, so quote-to-cash is a gap

Quotes live in Salesforce. Orders live in the ERP. The space between them is where your quote-to-cash analysis disappears. Win rates, average deal size, sales cycle length, and customer lifetime value all require bridging two systems that were designed with no intention of sharing records. Most teams maintain both manually and hope they stay synchronized.

Sales IQ
Planning & operations
12
MRP settings last reviewed at implementation, planners running on spreadsheets

Lead times are stale. Safety stock levels were set when the business was half its current size. Reorder points reflect a supplier relationship that no longer exists on the same terms. Planners compensate by maintaining Excel files and overriding the ERP daily. The system is technically in control. Operationally, it is not, and the spreadsheets become the system of record nobody can audit or inherit.

Inventory IQ · Operations IQ
Multi-entity & currency
13
No centralized exchange rate service, everyone uses their own rate

Multi-site manufacturers operating in multiple currencies often have someone manually updating a rate table, sometimes quarterly, sometimes only when the numbers look suspicious. Exchange rate inconsistencies across plants and portfolio companies create phantom margin shifts that are hard to explain and harder to trust. Two sites reporting in nominally the same currency but using different rates will never reconcile at the portfolio level.

Finance IQ · Pricing IQ
14
Daily Sales Average uses calendar days because sales days are never loaded in the ERP

DSA (Daily Sales Average) divides revenue by the number of days in the period, but the right denominator is sales days: the days on which orders are actually placed or shipped. ERPs carry a production calendar for work center scheduling. They do not carry a sales calendar. A plant may run five days a week, but orders only come in four. A distributor channel goes quiet the week of a regional holiday the ERP doesn't know exists. Using calendar days or production days as the denominator inflates or deflates DSA by a predictable but uncorrected amount, and the result is trend lines that move for structural reasons, not commercial ones. Comparing DSA across regions with different sales day patterns is comparing different denominators dressed up as the same metric.

Sales IQ · Finance IQ
Bills of materials
15
One BOM error at one level, dozens of corrupted cost rollups

A misconfigured sub-assembly in the bill of materials propagates upward through every finished good that uses it. A single BOM error can corrupt cost rollups across dozens of items before anyone notices. The first symptom is usually a margin exception report with a number nobody can explain. The cause, once found, is often a configuration decision made years earlier by someone who no longer works there.

Operations IQ

The data will never be perfect. That's okay.

Every challenge above is one we have seen, worked through, and built for. Not one of them is a reason to delay getting better information. The better move is a platform that works with manufacturing data as it actually exists, improves it as a byproduct of use, and gets your team to better decisions from where you are today.

Data quality isn't a project with a completion date. It's a continuous process. Every flagged anomaly, every corrected master record, every question the platform can't quite answer yet becomes a signal about where to focus next. Progress compounds, even when perfection isn't the destination.

Marquis IQ · Built for imperfect data
We've seen your data before. Here's how we work with it.

Marquis IQ connects to manufacturing ERP environments as they actually exist. We handle customer and supplier mastering across multiple ERPs, fragmented item records, stale standard costs, complex pricing structures, and missing planned costs as part of the integration process, not as blockers to getting started.

Customer and supplier mastering that resolves duplicated records across every ERP in the portfolio
Pricing analytics across list price, price curves, long-term agreements, and trade agreements
Centralized exchange rates and DSO calculations that account for regional business calendars
Analytics that surface where data quality gaps exist so you know where to focus first

FAQ

Questions from manufacturing teams

The questions we hear most often when companies are starting to work through their data situation honestly.

Do we need to clean our data before we can get value from Marquis IQ?

No, and if you're waiting for that, you'll be waiting a long time. Marquis IQ connects to your ERP data as it exists and improves it as part of the process. Customer deduplication, UOM anomaly detection, fragmented item records. That's handled during integration, not before it. You learn what your actual data problems are by using the platform, rather than having to guess at them in advance.

We have three different pricing structures across our ERPs. Can Marquis IQ work across all of them?

Yes, and this is one of the most common situations we see in PE-owned portfolios. Each acquired company tends to have its own pricing logic: different list price structures, different discount frameworks, different LTA formats. Marquis IQ normalizes these into a common analytical layer so you can compare pricing performance across entities without requiring the ERPs themselves to align. You don't need to standardize the source systems to get standardized visibility.

How long does it take to get reliable data out of a messy ERP environment?

Most customers are running live dashboards within weeks of their first call, not months. The data won't be perfect at that point, but it will be meaningfully better, and it keeps improving as the platform runs. The more useful question is usually: what specific decision do you need the data to support, and how clean does it need to be to support it? Some questions can be answered with what you have today. Others need targeted cleanup first. We help you figure out which is which.

Tell us what your data looks like. We've probably seen it.

Bring your messiest ERP environment and your most frustrating data problem. We'll tell you whether we've seen it before, what we did about it, and what it looks like on the other side.