Point of View
How the numbers get made. And how we fix them.
Written from the work Marquis does with manufacturers, distributors, and the firms that own them. Each one ends with how we deliver it on Marquis MCP, Agents, Apps, and IQ.
Start here
The four, explained.
- Marquis MCPWhat is an MCP server?The door between Chat, Cowork, and your company’s data.
- Marquis AgentsWhat is an agent?One job, a guard, and a person who signs off.
- Marquis AppsWhat is a vibe app?A prototype you can click. Not yet a system.
- Marquis IQWhat is Marquis IQ?One warehouse for every ERP. Token free.
AI and the data underneath
The model is the easy part. The data under it decides whether the answer is right.
- ERP MCP for Claude, ChatGPT, Cursor, and Excel AI: The Architecture That Makes Manufacturing AI Actually WorkEveryone wants to ask AI about their ERP data. Almost no one has the data foundation to make the answer trustworthy. Marquis IQ MCP Server gives AI tools secure access to clean, normalized, multi-ERP manufacturing data.→
- The Stylus Moment: AI, plain English, and the return to human-driven instructionSteve Jobs killed the stylus by giving us back our fingers. Twenty years later, AI is doing the same thing to business data. The interface is reverting to plain English, and that changes everything about who can ask questions and what they can know.→
- Self-Service Analytics for ManufacturersI was a Tableau certified trainer in the thick of the self-service analytics movement. I believed in it completely. Here's what I actually learned from training people across manufacturing, healthcare, and aerospace, and why AI finally changes the picture.→
- One Power BI Developer Is Not a Data Strategy (And Neither Is Your New AI Tool)Businesses are doing the same thing with data that shoppers do in the Costco TV aisle. They walk in knowing what they need, get completely overwhelmed, and either buy nothing or buy the wrong thing. Here's why that keeps happening and how to fix the order.→
Margin, price, and procurement
Every point of margin has a cause. Name it, size it, and give it an owner.
- What Is Price-Volume-Mix Analysis?PVM analysis names every force that moved your gross margin from one period to the next. Done right, it separates revenue effects from cost effects and gives each driver a number. This guide covers all six: Price, Volume, Mix, Materials, Labor, and Burden.→
- Why Did My Gross Margin Change?When gross margin moves, the answer is almost never a single cause. Price, volume, mix, materials, labor, and burden are all moving at the same time. This framework separates every driver and sizes each one, so the conversation at the board table starts with evidence, not instinct.→
- 5 Data Problems That Corrupt Your PVM BridgeThe price-volume-mix bridge is only as accurate as the data it runs on. Before you trust its output, five data problems routinely corrupt the inputs - and each one silently misattributes margin movement to the wrong driver.→
- Gross Margin Analysis: Your ERP Tracks Both Sides. Most Tools Only Show You One.Real gross margin analysis decomposes unit cost (labor, material, burden) and realized price (list, price list, volume slope) independently. Most margin analysis tools skip that work and report the net. Here is what you miss when they do.→
- ERP Procurement Analytics: The Eight Signals Hidden in Your Purchase Order DataEvery PO your team places, every goods receipt that comes in, and every invoice that clears AP generates a procurement signal. Most manufacturing teams see those signals once a month, in a spreadsheet, after the fact. Here is what real-time procurement analytics looks like built directly from your ERP data.→
Inventory, cash, and operations
Working capital hides at the SKU, the plant, and the invoice.
- Inventory Days on Hand by SKU: The Metric Every Manufacturing Ops Team Is MissingYour ERP reports 45 days of inventory on hand. That number is accurate and useless. Somewhere in your warehouse right now there are bearings with 700 days of supply three aisles from a control relay four days from a stockout. Here is how SKU-level DOH changes every inventory decision you make.→
- How to Calculate Inventory Turns: Why the Common Formula Fails ManufacturersThe inventory turnover formula in your ERP uses two data points: inventory on the first day of the period and inventory on the last. For seasonal operations, that two-point average swings 40 percent between quarters without a single item moving faster or slower. Here is why the daily average method produces reliable turns at any level of aggregation, and why turns-by-quantity and turns-by-value diverge the moment you roll up above the SKU level.→
- Inventory Aging Analysis: Identifying Dead Stock Before It Costs YouAn inventory aging report categorizes every stocked item by how long it has been sitting without demand. At 30 days, an item warrants monitoring. At 90 days, a purchasing pause. At 180 days, an active disposition decision. At 365 days, a write-down conversation with finance. Here is how aging analysis works for manufacturers, what drives each category, and how to turn the report into a working capital action list.→
- 5 Data Problems That Corrupt Your Inventory AnalyticsDays on hand, inventory turns, and aging buckets are only as reliable as the data they run on. Five problems - unit of measure mismatches, negative balances, ghost inventory, location gaps, and stale lead times - silently corrupt every metric your team uses to manage working capital.→
- My Favorite Workstream: The Inventory Data Model That Has Answered Every Question for 15 YearsPeople ask me what my favorite workstream is. After 15 years building manufacturing analytics, the answer is always the same. Not because it is the easiest to explain. Because when the data model is right, a buyer, a planner, and an ops leader can all get answers from the same two tables. That almost never happens in this industry.→
- Inventory Analytics for Manufacturers: Why Plant-Level Reporting Is Not EnoughEach plant has its ERP and its inventory reports. Plant A knows its own on-hand, its own days on hand, its own slow movers. What Plant A does not know is that Plant B is sitting on 140 days of the same component under a different item code. That gap is not a failure of either plant's ERP. It is a structural blind spot that no single-plant system can close. Here is how multi-site inventory analytics closes it.→
- Accounts Receivable Aging Analysis for Manufacturers: Beyond the ERP Bucket ReportYour ERP produces an aging report that groups overdue invoices into 30-day buckets and sums the totals. The totals are accurate. The report does not tell you which single customer accounts for 40 percent of your overdue balance, whether DSO has trended up for three consecutive quarters, or which invoices the collections team should call on today. Here is the AR analytics framework that actually shortens collection cycles for manufacturers.→
- The 7 ERP Reports Every Manufacturing CFO Needs (and Why Your ERP Doesn't Build Them Automatically)Your ERP captures the raw data for all seven of these reports. It does not structure it into them. Here is what each report contains, why the gap exists between operational ERP data and management-ready analytics, and what decisions each one enables.→
- Marquis Playlists is built. Put your ERP data on the TV.A fully configurable rotation of the dashboards you already have, with a timer on every screen, controlled by the people who own that place. It is in Marquis IQ now. A customer can turn a wall into a live view of the plant this week.→
Forecasting and demand
A forecast is only useful if someone can explain it, and it runs again next month.
- Why Manufacturing Demand Planning Fails: And the data problem at the root of itMost manufacturing demand plans fail before a single model runs. The algorithm is rarely the problem. Three structural data issues - fragmented transaction history, unreliable feature data, and a grain definition that doesn't match the decisions it's meant to support - corrupt every forecast built on top of them.→
- From ERP to MRP: Building a Forecasting Competency That Runs ItselfThe difference between a forecasting project and a forecasting competency is what happens after the model finishes running. A competency runs again automatically next month, writes its primary output back into the ERP as a demand plan, and gets more accurate with every period that passes. Here is the architecture that makes that possible.→
- Demystifying AI Forecasting: From Black Box to Boardroom AnswerProducing a demand forecast is not the hard part. Explaining it is. When an executive asks "how did it come up with $1.5M for next month?" the answer requires opening the black box - feature matrices, model competition, SHAP attribution, and forecastability scoring. This article walks through every layer.→
Acquisitions and private equity
A different ERP at every company. One view for the board.
- ERP Data Consolidation After an Acquisition: The 100-Day Playbook for PE-Backed ManufacturersAt close, a PE-backed platform company inherits whatever ERP stack came with the deal. Three plants, three ERPs, no common data layer, no shared supplier master, and a board expecting operational visibility inside 100 days. Here is how to get there without touching a single ERP.→
- The Five Analytics Questions Every PE Operating Partner Should Ask at Portfolio CompaniesA PE firm closes on a manufacturer. The board wants visibility. The portfolio company's ERP holds all the data. Here is the framework for extracting the five answers your investment thesis depends on, and the warning signs that the data infrastructure isn't ready to provide them.→
- How PE Firms Are Flipping the Script with Marquis IQ: Master Data as a Strategic Advantage for OperatorsThe PE firms pulling ahead are not just running better analytics. They are treating master data as a strategic foundation - establishing the vocabulary, resolving the duplicates, and extending the data model in ways that create a compounding advantage across every portfolio company and every reporting cycle.→
- Beyond the 12-Month Forecast: How PE Firms Use Demand Models to Understand Revenue QualityEvery month you report $15M in revenue. But how much of that would have arrived without the sales team lifting a finger - and how much depended entirely on this month's commercial effort? Time series decomposition answers that question, and for PE operating partners, it is more valuable than the forward forecast it also happens to produce.→
The data foundation
Nobody’s data is clean. Progress beats waiting for it to be.
- 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.→
- Your ERP Runs the Business. It Was Never Designed to Understand It.For almost 15 years we have been helping manufacturers grapple with disconnected, duplicated, dirty, and distrusted data. With the explosion of AI, the need for clean, context-aware data has never been more urgent. Here is what we have learned about why the old tools failed, and what actually works.→
Bring the question.
Margin, inventory, the forecast, or the next acquisition. Tell us where it hurts.
