# Your ERP Says You Have 200 Units. You Have 147. Canonical: https://granular.to/blog/erp-says-200-units-you-have-147 Published: 2026-05-19 Updated: 2026-05-19 Author: Trey Category: Field notes Tags: distribution, erp, inventory, operations > A field dispatch on why ERP inventory counts break down at mid-market distributors, covering receiving errors, adjustment abuse, returns processing gaps, and the shadow systems every warehouse runs alongside the official system. > **TL;DR.** The average distribution warehouse runs at somewhere between 63% and 85% inventory accuracy, depending on who you ask and how they measure it. That means anywhere from one in six to one in three line items in your ERP is wrong. For a $50M distributor, that gap translates to hundreds of thousands in lost sales, excess carrying costs, and emergency freight. The fix is not a bigger ERP or another bolt-on module. It is closing the gap between what your system says and what your warehouse team actually sees, one process failure at a time. Your ERP says you have 200 units of that high-margin connector fitting. Your biggest customer calls to place a rush order for 180. Your sales rep confirms availability, promises Thursday delivery. Wednesday morning, the warehouse picks the order and finds 147 units on the shelf. Now you are scrambling: calling suppliers for expedited stock, eating overnight freight costs, and explaining to a customer why the system lied. This is not a technology failure. Your ERP is doing exactly what it was told. The problem is that nobody told it the truth. ## The Accuracy Gap Nobody Talks About The [Inventory Management Review](https://www.inventorymanagementreview.org/) puts average warehouse inventory accuracy at roughly 63% for companies without robust cycle counting programs. The [CAPS Research](https://www.capsresearch.org/) benchmarks are slightly more generous, landing around 80% to 85% for distribution operations with regular audits. Either way, the gap is massive. For a $50M distributor carrying $8M in inventory, even a 15% error rate means $1.2M worth of stock records are wrong at any given time. Some of those errors cancel out (overages here, shortages there), but the net effect hits in three places: lost sales when you cannot fill orders you thought you could, excess inventory on items you overbought because the system showed zero, and emergency procurement at premium prices when reality catches up. The [Institute for Supply Management](https://www.ismworld.org/) has tracked this pattern for years: mid-market companies invest in ERP implementations, see initial accuracy improvements, and then watch those numbers degrade over the following 18 to 24 months. The system did not get worse. The processes around it did. ## Where the Counts Go Wrong We have seen the same five failure points across every distribution operation we have worked with. They are not exotic. They are mundane, which is exactly why they persist. ![Barcode scanner resting on partially opened pallet at warehouse receiving dock with shipping manifest](/images/blog/erp-says-200-units-you-have-147-receiving-dock-scanner.jpg) ### Receiving errors This is where most accuracy problems start. A PO says 200 units. The pallet arrives, your receiving clerk scans the packing slip barcode, confirms 200 in the system, and moves on. Nobody actually counted the boxes. Maybe the supplier shorted you by 12. Maybe they sent 208. The ERP now has a number that was wrong from the moment it entered the system. [Modern Distribution Management](https://www.mdm.com/) has reported that receiving discrepancies account for roughly 30% to 40% of total inventory variance in mid-market warehouses. The fix is not complicated: require physical count verification at receiving, flag discrepancies for supplier follow-up, and track supplier accuracy rates over time. But it takes discipline, and it takes time your receiving crew does not think they have. ### Adjustment abuse Every ERP has an inventory adjustment function. It exists for legitimate reasons: damaged goods, quality holds, unit-of-measure corrections. But in most mid-market warehouses, it becomes a pressure valve. Picker cannot find the item? Adjust it out. Surprise overage during a spot check? Adjust it in. Month-end physical count does not match? Bulk-adjust everything to match what is on the shelf. Each adjustment is technically correct in the moment. The problem is that nobody investigates why the discrepancy existed. Adjustments mask root causes instead of exposing them. We have seen warehouses averaging 400 to 600 inventory adjustments per month with zero root-cause analysis on any of them. ### Returns processing Returns are where accuracy goes to die. A customer sends back 15 units. The return hits receiving. Maybe it gets inspected immediately, maybe it sits on a pallet in the staging area for three days. The credit gets processed in the ERP before the physical inspection happens. Now you have 15 units on the books that may or may not be saleable, sitting in a location the system does not track, waiting for someone to decide what to do with them. For distributors with return rates above 5% (common in electrical, plumbing, and HVAC distribution), this single process gap can account for 20% or more of total inventory variance. ### Cycle count theater Most mid-market distributors have a cycle counting program. On paper, it looks good: count a percentage of SKUs every week, investigate discrepancies, adjust as needed. In practice, the counts happen on paper or handheld scanners, the "investigation" is a 30-second look around the shelf, and the adjustment goes in without any process change. [Research from Auburn University's RFID Lab](https://rfid.auburn.edu/) has shown that cycle counting programs without root-cause protocols actually increase error rates over time. The counts create a false sense of accuracy while the underlying process failures continue unchecked. You are not improving; you are just resetting the clock. ### The shadow system Here is the one nobody puts in a vendor pitch deck. Every mid-market warehouse has a shadow system. It is the spiral notebook the warehouse manager keeps with "real" stock levels. It is the whiteboard near the shipping dock with notes like "Bin C-14 actually has ~40, not 72." It is the experienced picker who knows that whenever the system says they have something in Row 7, they should check Row 12 first because stuff migrates. ![Warehouse desk with printed ERP inventory report next to handwritten notebook with different numbers circled in red pen](/images/blog/erp-says-200-units-you-have-147-shadow-system-notebook.jpg) These shadow systems exist because they work. Your best warehouse people have learned not to trust the ERP, so they built their own tracking. The problem is that shadow systems are not scalable, not transferable, and not visible to anyone making purchasing or sales decisions. When the warehouse manager who keeps that notebook [retires or leaves](/blog/capture-tribal-knowledge-before-key-people-leave), the shadow system goes with them. ## What Actually Fixes This The instinct is to buy something: a WMS module, RFID tags, automated counting drones. And some of those tools are genuinely useful, eventually. But bolting technology onto broken processes just gives you faster, more expensive wrong answers. The distributors we have seen close the accuracy gap do three things, in this order. **First, they measure accuracy at the transaction level, not the SKU level.** Instead of asking "is our inventory accurate?" they ask "which specific transactions create discrepancies?" That means tracking accuracy at receiving (did the PO quantity match the physical count?), at picking (did the picker find what the system said was there?), and at adjustment (what triggered this adjustment?). Transaction-level tracking exposes the root causes that SKU-level audits hide. **Second, they fix receiving before anything else.** Every unit that enters the warehouse with an incorrect count creates a cascade of downstream errors. A 3% receiving error rate on 500 daily receiving lines means 15 wrong records per day, 75 per week, 300 per month. Each one compounds through picks, adjustments, and cycle counts. Fixing receiving accuracy from 95% to 99% has more impact than any other single change. **Third, they kill the shadow system by making it official.** Instead of fighting the spiral notebook, they formalize it. The warehouse manager's annotations become systematic exception flags. The picker's knowledge about where things actually live becomes a location audit trail. The goal is not to eliminate tribal knowledge; it is to capture it in a system that survives personnel changes. ## The Real Cost of "Close Enough" Most distributors live with inventory inaccuracy because the pain is distributed and hard to quantify. A lost sale here. An expedited freight charge there. A customer who quietly shifts volume to a competitor because your fill rates dropped. [Altavant Consulting](https://www.altavant.com/) estimates that inventory inaccuracy costs mid-market distributors between 3% and 5% of annual revenue in combined effects: lost sales, excess carrying cost, emergency procurement premiums, and labor spent working around bad data. For a $50M distributor, that is $1.5M to $2.5M per year. Not in one dramatic failure, but in a thousand small ones that never make it onto a P&L line item. Your [ERP is not the problem](/blog/no-erp-does-everything-what-works-instead). Your ERP is a mirror. If you do not like what it is showing you, the answer is not a better mirror. ## FAQ **How often should a mid-market distributor run cycle counts?** Most operations benefit from counting high-velocity A-items weekly and slower-moving B and C items monthly. The frequency matters less than what you do with the discrepancies. A cycle count without root-cause investigation is just an expensive adjustment. **Can AI or automation fix inventory accuracy problems?** Tools like automated receiving verification, image-based counting, and predictive discrepancy flagging can help, but only after the underlying process gaps are addressed. Automating a broken process just produces wrong answers faster. **What is a realistic accuracy target for a mid-market distributor?** Best-in-class distribution operations hit 97% to 99% accuracy at the SKU level. Most mid-market distributors should target 95% as a first milestone, which typically requires fixing receiving processes and formalizing cycle count root-cause protocols. **How long does it take to improve inventory accuracy meaningfully?** With focused receiving process fixes, most distributors see measurable improvement within 60 to 90 days. Getting from 85% to 95% accuracy typically takes six to nine months of sustained process discipline. The technology investments (WMS, RFID, barcode verification) layer on top of that foundation. If the gap between your ERP and your warehouse floor sounds familiar, [Granular builds focused tools](/) that close it. Not a new ERP, not a six-month consulting engagement. A working solution in four weeks, fixed price. [Book a discovery call](/) and bring your worst inventory report. --- ## Keep Reading - **[No ERP Does Everything. Here Is What Works Instead.](/blog/no-erp-does-everything-what-works-instead)** Where ERPs fall short for mid-market operations and what to build around them. - **[How to Capture Tribal Knowledge Before Key People Leave](/blog/capture-tribal-knowledge-before-key-people-leave)** The playbook for documenting what your best people know before they walk out the door.