The 5-Day ERP Integration: The Fourth Shift Story Behind EDGEBIC's Approach

Fast ERP scheduling integration is measured in days, not months, when the interface is a file instead of a connector: the documented proof is Plastilite Corporation, where a complete Fourth Shift ERP integration (data in, optimized schedule out, dates synchronized back) ran from Monday to Friday. That 5-day story belongs to the User Solutions lineage that built it, and it is the origin of the architecture EDGEBIC ships today: reusable import masks that turn a week of expert integration craft into a saved configuration any planner re-runs in two clicks.

This post retells the case as documented, extracts the principles that made one week possible, and shows how EDGEBIC by User Solutions turned each principle into product machinery. If you want the step-by-step instructions for your own Fourth Shift install, the companion EDGEBIC + Fourth Shift integration guide has them.

The setup: an ERP vendor recommends someone else's scheduler

Plastilite Corporation of Omaha, Nebraska makes insulated and protective packaging: their products ship everything from Omaha Steaks to vaccines. To support growth they selected a new ERP, and Fourth Shift (from Softbrands, Inc.) was the frontrunner for accounting and inventory.

Scheduling was the sticking point. Plastilite's floor runs many injection molding machines, each with different tool configurations holding different combinations of molds for different products. Picking the right machine means picking the right machine with the right chest and the right molds, against finite capacity, across the whole order book. That was, in the words of the case record, far too complex for the ERP's basic infinite capacity scheduling.

What happened next is the unusual part. Fourth Shift itself recommended Resource Manager DB (RMDB) from User Solutions as the scheduling add-on for its customer. An ERP vendor sending its customer to a third-party scheduler is the strongest kind of endorsement in this category, because the vendor has every commercial reason not to. Jon Ehly, Plastilite's Chief Financial Officer, confirmed the recommendation came from Fourth Shift directly. The full account is in the Plastilite case study.

Monday to Friday: what actually happened

User Solutions arrived on Monday. By Friday, Plastilite had a complete, optimized, finite capacity schedule with full ERP integration. One week contained:

No API project. No middleware. No months of specification. The ERP was never modified: it exported orders and received dates, and stayed the untouched system of record throughout.

Why one week was enough: four principles

Strip the story to its mechanics and four principles fall out. They are the reason the week worked, and they are testable against any integration proposal you are evaluating.

1. The interface is a file the ERP already produces. Every dependency you add (API access, sandboxes, vendor coordination, certification) adds a queue. An export the ERP already generates has no queue. This is why the same lineage later scheduled Cummins across 33 locations from AS400 data: the file interface predates and outlives every integration fashion.

2. The ERP is never touched. Risk approvals kill timelines. An integration that only reads exports and hands back Excel needs no change control on the ERP side, so nobody has to say yes except the planner.

3. Schedule on day 3, not month 6. A real schedule from real data, early, converts the project from speculation to correction. Plastilite's team spent the back half of the week tuning a working schedule, not specifying a future one.

4. Data before perfection. The routings entered that week were good enough to schedule, then improved against results. Waiting for perfect data is how integrations die; scheduling imperfect data is how it gets perfect.

How EDGEBIC industrialized the method

In the RMDB generation, that week rested on expert hands. EDGEBIC's contribution is turning each principle into machinery a planner operates alone. The EDGEBIC ERP integration architecture is the 5-day method productized:

The import mask is the consultant's mapping, saved. A mask stores which entity you are importing (products, work centers, routings, orders, actuals), the file format, whether the first row is headings, and how each source column maps to an EDGEBIC field. You build it once by dragging your export's headings onto target fields. Every later run is two clicks, and the mask remembers the file path and worksheet.

Unit conversion is a stored factor, not a spreadsheet step. ERPs disagree with schedulers about units constantly. A conversion factor of 0.016667 on a setup-time column turns minutes into hours during every import, forever, with no one remembering to do it.

The routing import wires itself. Routings are the hardest data to move because operations reference each other. EDGEBIC imports them in two passes: pass one validates rows and auto-creates any work centers or products the file names that do not yet exist; pass two sorts each product's steps by sequence number, saves them, and wires step 10 to 20, 20 to 30, and the final step to the finished product. Alternate and parallel work centers (the exact shape of Plastilite's machine-tool-mold problem) import as routing attributes.

Every row is accounted for. Each imported row lands as exactly one of Created, Updated, Reused, or Failed. The result dialog shows the counts; a per-run log file records every row and every failure reason. The Friday-afternoon question "did everything make it in?" has a written answer.

Re-running is safe by design. Matching records come back Reused (untouched) unless you enable updates, and on updates a blank cell preserves the existing value instead of wiping it. Routing re-imports wipe and recreate a product's steps in one run, so repeated imports never accumulate duplicates. And scheduled jobs run from a frozen snapshot of their routing, so data maintenance never rewires work in progress: completed operations are never moved at all.

Imports never schedule. Imported orders wait as unscheduled demand until you run the scheduler. Data movement and planning stay separate, so an import can never silently rearrange the floor. It is the same discipline that let Plastilite's ERP stay untouched, applied one layer up.

What did not make the week fast

The story invites a wrong conclusion: that speed came from heroics or from a simple shop. Neither is true, and knowing what did not matter is as useful as knowing what did.

It was not simple data. Injection molding with variable tool configurations and mold combinations is among the harder scheduling shapes there is: the machine, chest, and mold have to be chosen together, against finite capacity, for every order. Complexity was the reason the ERP module failed, and the week absorbed it anyway.

It was not pre-cleaned data. The scheduling data was broken down during the week, not before it. Day one started from what the ERP and the planners actually had.

It was not a small order book. The open orders imported from Fourth Shift were the live book, not a pilot subset. The schedule generated on that data was the schedule the shop ran.

What made it fast was the absence of queues: no API to negotiate, no middleware to build, no change request against the ERP, no steering committee between an export and a schedule. Every activity in the week produced something the next activity consumed the same day. That property is architectural, which is why it transfers to your shop even though your machines, your ERP version, and your product mix do not match Plastilite's.

The lineage, and where it points

The 5-day week was not a one-off stunt; it was one visible data point in a 35+ year record. User Solutions has integrated scheduling beside ERPs since 1991: Fourth Shift by the vendor's own recommendation, Macola, Cummins across 33 locations from AS400 data, and results like GE Railcar lifting on-time shipping from 30 percent to 90 percent. The general implementation playbook that grew out of this history is covered in the 5-day scheduling software implementation process.

EDGEBIC is where the lineage points now: the same file-first integration doctrine feeding a current-generation engine (work center groups, sequence-dependent setup optimization with a proven optimality gap, operator skills, lot streaming, and reschedules that never move completed work). The EDGEBIC product overview covers the engine; the JobBOSS and Epicor guides show the same method applied to other ERPs; and the ERP scheduling add-on page frames the category. If you are new to the discipline itself, start with what production scheduling is.

Your Monday

The story's practical lesson fits in one sentence: bring an export, not a project plan. Pull this week's open orders and one product's routing from your ERP (Fourth Shift or any other) and bring the files to a demo. Mapping them takes minutes. A schedule from your own data exists before the meeting ends, and then you are living the same week Plastilite did: tuning a working schedule instead of specifying a future one.