Most manufacturers I meet are not looking for software. At the end of the day, they're looking for an answer to a question that their current system cannot answer: what did that job actually cost, and can we promise the customer a date we will hit? Business Central answers both, but only if the manufacturing side of the product is set up deliberately rather than switched on and hoped for the best.
The manufacturing granule set is broad, and a first-time implementation can drown in it. So here are the five areas that carry the weight. Get these right and the rest of the module tends to fall into place; get them wrong and no amount of reporting will rescue the numbers underneath.
Everything downstream is built on these two tables. The production BOM says what goes into the item; the routing says where it goes, in what order, and how long it takes. Business Central lets you version both, with certification status and starting dates, so an engineering change becomes a new certified version rather than an edit that quietly rewrites history on jobs already closed.
Multi-level BOMs are where the accounting interest sits. A subassembly can be its own produced item with its own routing and its own cost roll-up, or it can be flattened into the parent. That choice decides whether you can see margin at the subassembly level later, and it's very hard to change once a year of history has accumulated. Decide it during design, with the controller in the room.
Routings carry setup time, run time, wait time, and move time against work centers and machine centers. Those four fields drive scheduling, capacity load, and the capacity portion of your standard cost. If the shop floor tells you the times are approximate, that's fine to start with, but write down that the standard cost is approximate too.
A production order moves through simulated, planned, firm planned, and released status, but each status means something different to planning. Firm planned orders are visible to the planning engine and can be rescheduled by it; released orders are committed and are what the floor works against. Teams that release everything immediately lose the flexibility the status ladder was built to give them.
Execution happens through the output journal, the consumption journal, and the production journal. Flushing method is the quiet setting that decides how much data entry the floor has to do:
Pick-plus-backward is the right answer in a warehouse-managed plant and the wrong answer everywhere else.
Subcontracting deserves its own mention because it's handled well and used rarely. Flag a routing operation with a subcontractor work center, run the subcontracting worksheet, and Business Central generates the purchase order for the outside service and posts the operation when the receipt happens. The outside cost lands in the job rather than in a general overhead bucket.
If I could only implement one thing for a manufacturer, this would be it. The planning worksheet takes demand from sales orders, forecasts, and component requirements, nets it against supply on hand and on order, and returns a list of suggested actions: create this production order, expedite that purchase order by six days, cancel the transfer that is no longer needed. It's the single biggest step up from a spreadsheet-driven purchasing process.
The quality of those suggestions depends entirely on item-level planning parameters. Reorder policy, safety stock, lead time calculation, reorder cycle, and the ordering multiples are not master data housekeeping; they are the model. Four reorder policies are available, and most catalogs need a mix rather than one applied across the board.
Reorder Policy |
How It Plans |
Typical Fit |
|
Lot-for-Lot |
Orders exactly the net requirement, period by period |
Made-to-order items and expensive components |
|
Fixed Reorder Quantity |
Orders a set quantity whenever stock drops below the reorder point |
Steady-usage parts with a natural batch size |
|
Maximum Quantity |
Tops stock back to a defined ceiling |
Bulk or shelf-stock items with storage limits |
|
Order |
Creates supply tied one-to-one to a specific demand |
Custom or serialized configure-to-order work |
Run planning as a regenerative calculation while you're learning the data and switch to net change once the parameters are trusted. Resist the temptation to accept every suggestion the first month. Read the action messages, argue with them, and fix the parameter that produced the one you disagree with. That argument is the implementation.
This is the section accountants care about and the one that gets shortchanged most often. Business Central supports FIFO, LIFO where permitted, average, specific, and standard costing, and manufacturers overwhelmingly want standard. Standard costing gives you an expected cost to plan and quote against, and then measures reality against it.
The measurement is where the value sits. When a production order is finished and costs are adjusted, the difference between expected and actual is broken into named variances: material, capacity, capacity overhead, subcontracted, and manufacturing overhead. That breakdown answers the question a plant manager and a controller usually argue about in the dark, which is whether the job lost money on the metal, on the hours, or on the outside process.
Two operational habits make or break this.
Run the adjust cost routine on a schedule rather than when someone remembers, because until it runs your inventory valuation and your general ledger are describing different months.
Roll standard costs on a deliberate cadence with a documented revaluation, so the variance you are reading is production performance rather than a price list that went stale in the spring.
Remember: standard cost provides the ability to review variances.
If standard cost variances are not reviewed, another costing method might serve you more appropriately. Production orders provide expected costs (based on the BOMs and routings) and actual costs (based on consumption, capacity and output.) Standard costing for accounting value of inventory may not be the answer, just the production order variances.
Lot and serial tracking runs through the whole chain, from purchase receipt through production consumption and output to shipment, so a recall question becomes a trace rather than a project. For food, pharmaceutical, aerospace, and medical device work, don't consider this a nice-to-have. Also, the expiration and warranty date handling on lots should be reviewed before anyone writes a customization for it.
Warehousing scales in tiers, which people often miss. You can run simple location-level inventory, add bins, add directed put-away and pick, and only take on the full warehouse management overhead where the plant actually needs it. Turning on the top tier everywhere is the most common way a manufacturing go-live gets slower instead of faster.
Capacity planning ties back to those routing times. Work centers and machine centers carry calendars, efficiency, and capacity, and you can plan against finite capacity where a real constraint exists and infinite capacity elsewhere. Modeling every machine as a constraint produces a schedule nobody believes; modeling the two that actually gate throughput produces one people follow.
Assembly management is the lighter sibling of full manufacturing, and for kitting or light configure-to-order work it's often the better answer. It gives you an assembly BOM and an assembly order without routings, capacity, or production order overhead, and you can assemble it directly to a sales order line.
Dimensions turn production data into analysis. Tag output by product line, plant, cell, or customer program and profitability questions stop requiring a data project. Set the dimension structure before the first production order posts, because retrofitting dimensions onto history is a job nobody enjoys twice.
Analysis mode and financial reporting cover more shop floor reporting than you might expect. Before commissioninga Power BI model, try answering the question inside Business Central with analysis mode on the item ledger and capacity ledger entries. A surprising share of the standard manufacturing questions are already there.
If you are scoping a manufacturing implementation, sequence it in the order above. Data first: clean BOMs and honest routings. Execution second, so the floor is posting real output and consumption before anyone judges a number. Planning third, once the parameters have something true to work from. Costing and variance analysis last, because the variances only mean something when the three layers beneath them are trustworthy.
Manufacturers who take that order tend to reach a believable standard cost within a couple of closes. The ones who start with the costing reports usually spend a year explaining variances that were really data problems all along.
Need someone to dig deeper into your system specifically? Centre Technologies helps you connect and automate your business operations with dedicated, ongoing support. Contact one of our ERP experts and they'll help you get everything set up so you're happy with the results.