Independent IFS Cloud practice · Business process optimisation
Reorder points in IFS Cloud: classify the parts first, then let the job calculate
A reorder point is only as good as the group of parts it was calculated for. Inventory Planning and Replenishment (IPR) in IFS Cloud classifies parts by value, movement and lifecycle stage, and calculates a lot size, a safety stock and an order point for them. In my view the classification is the part that decides the result: if every part gets the same parameters, a fast mover and a part nobody has issued for a year are treated alike, and the second one ends up in the aging report.
Key takeaways
- IPR is an inventory control solution based on reorder points. IFS positions it for independent demand, mainly distribution and spare parts, and it can run alongside Kanban and MRP.
- Parts are classified three ways: ABC by volume value, frequency by issue transactions per month, and lifecycle (Introduction, Mature, Decline, Expired). The planning parameters are then set per combination.
- The job calculates four things: lot size, safety stock, order point and next order date. Models are chosen for demand, safety stock, lot size and order point, from Manual to statistical ones.
- Values are inherited from Company down to Supplier, and the lower level overrides the higher one, so the most specific rule wins.
- The classification can be overridden on the part until a locked-until date. After that date the history takes over again.
1.What problem does Inventory Planning and Replenishment solve?
Parts with independent demand are not driven by a production plan, so there is nothing to explode and nothing for MRP to net. Someone, or something, has to decide when stock is low enough to reorder and how much to order. With no MRP behind these parts, the decision falls to whoever orders them, in practice a planner with a spreadsheet, which is the pattern described in the shadow spreadsheet running a factory.
IPR moves the decision into the system. A part is planned this way when its Planning Method on the inventory part is set to B for reorder-point planning. The documentation lists two prerequisites besides that: accurate and complete basic data, and, for a switchover, the "Import Inventory Part Daily Issues" assistant, which can load historical issue data so the first classification has something to work with.
2.How are parts classified, and why does it come first?
The step is called Perform ABC, Frequency and Lifecycle Classification, and it runs on issue history. Three views of the same part are produced.
| Classification | Based on | Classes |
|---|---|---|
| ABC | Volume value: the part's inventory value multiplied by its issued quantity | A, B, C. By default 80%, 15% and 5% of turnover within the asset class or site |
| Frequency | Issue transactions per month, compared with limits set per site or asset class | Fast, Medium, Slow and Very Slow Movers |
| Lifecycle | Stage boundaries, set as offsets per site or asset class | Introduction, Mature, Decline, Expired. Parts move between stages automatically |
A classification matrix is created for each lifecycle stage and each site and asset class combination. The asset class is the grouping you choose, for example spare parts against raw materials. It is also where seasonality is set, and where the Classification Periods field decides how many months of history the classification reads.
Four details are easy to miss. The ABC ranges are global, so the same percentages apply to every part that IPR plans. And on the ABC Classes page, a system administrator can change the percentages, but the accumulated total of all classes must be 100% or the page cannot be saved. According to the 25R2 page, a new part record gets class C until an ABC analysis has been run, and the calculation is not carried out for configured parts.
The classification can be overridden on the Inventory Part page until a locked-until date. After that date it reverts to history. The documentation gives new parts with a forecast but no sales history as an example: you set the class yourself for a defined period.
3.What does the job actually calculate?
Four outputs: the lot size (the quantity proposed when replenishment is needed), the safety stock (the buffer for demand variation), the order point (the level at which replenishment is due) and the next order date, which assumes consumption follows the forecast.
| What is modelled | Models in the documentation |
|---|---|
| Demand | Forecast (from IFS Demand Planning), Yearly Prediction (field Pred Year Cons Qty), or History |
| Safety stock | Manual, Time Coverage, Historical Uncertainty (uses the service rate), Mean Absolute Error (uses the Demand Planning error) |
| Lot size | Manual, Time Coverage, Economic Order Quantity. Max Order Cover Time, Durability and Min, Max and Multiple lot sizes also apply |
| Order point | Manual or Lead Time Driven. For slow movers: Slow Movers by Lifecycle or Lead Time, and Croston by Lifecycle or Lead Time |
The inputs behind them are lead times, ordering cost, the inventory interest rate, the service rate, available quantities and demand. Lead-time variation is taken from historical order occasions, so it reflects how replenishment actually arrived.
The documentation does not describe the sequence of the IPR job, the default service rate values, or exactly how the results are written back to Inventory Part Planning Data. Treat those as things to confirm in your own system before you rely on them.
4.Which rule wins when several levels define a value?
Parameters can be set at several levels, from the top: Company, Site, the ABC, frequency and lifecycle combination, Asset Class, Commodity Group, and Supplier. Lower-level values override higher-level ones.
My suggestion for a start: a few rules at the top that cover everything, then exceptions only where a part group behaves differently. A value set at Supplier level beats one set at Site level without anyone needing to touch the parts.
In my experience the same inheritance is also how a setting gets forgotten: a value placed low in the hierarchy keeps winning after the reason for it has gone.
5.Where does reorder point planning go wrong in practice?
These are conclusions from the documented behaviour and from reviews of how companies run planning, not features IFS Cloud raises on its own.
| What you see | What to check first |
|---|---|
| Parts that should be planned this way show no proposals | The Planning Method on the part is not B, so IPR never looks at it. |
| Many new parts sit in class C | No ABC analysis has run since they were created. New parts get class C until it does. |
| Seasonal parts get a low order point in the season | Classification Periods on the asset class covers too few months, or seasonality is not set. |
| A part keeps an old class long after its demand changed | A manual override with a locked-until date far in the future. |
| Order points rise and stock ages | Safety stock and lot size models are too generous for slow movers, and the extra stock shows up later in the stock aging report. |
| Counts keep finding differences on the same parts | Planning is working from quantities that are not true. See cycle count variance reconciliation. |
The pattern behind four of the six rows is that a parameter was right on the day it was set. After go-live nobody owns the question of whether it is still right, which is the subject of process drift after go-live.
6.How do you introduce it without disturbing the warehouse?
These steps are my recommendation, not an IFS procedure.
- Start with one asset class. Pick a group of parts with real issue history and a planner who knows them.
- Run the classification and read it. Check that the A parts are the ones the business already calls critical. If they are not, look at what drives the volume value and at the Classification Periods setting.
- Compare proposals with current practice for one full replenishment cycle, or one season for seasonal parts. Do not order from the proposals yet. Note where they disagree with what the planner did and why.
- Choose the simplest model that holds. Manual and Time Coverage are easier to explain to a buyer than Croston. Move to a statistical model for a class only where simple rules keep missing.
- Give the overrides an owner and an expiry. Every locked-until date and every low-level rule needs a name next to it and a review date.
Steps two and five are a checking job, not a configuration job: which parts have a class that no longer matches the demand, and which overrides have outlived their reason. A read-only monitoring pattern can produce that list without writing anything back to IFS Cloud.
The primary sources are the IFS Cloud documentation pages Inventory Planning and Replenishment (26R1) and ABC Classes (25R2). They are written for different releases, so check the pages for the release you run.
7.Frequently asked questions
Does IPR replace MRP in IFS Cloud?
No. The documentation describes IPR as a solution for independent demand, mainly distribution and spare parts, and says it can run alongside Kanban and MRP. A part is planned by IPR when its Planning Method is set to B for reorder-point planning.
What are the default ABC classes in IFS Cloud?
The system ships with three classes: A at 80%, B at 15% and C at 5% of turnover value within the asset class or site. A system administrator can change them on the ABC Classes page, as long as the cumulative total stays at 100%. Parts are ranked by volume value, which is the part's inventory value multiplied by its issued quantity.
Can I override the classification of a single part?
Yes. Classification values on the Inventory Part page can be set manually until a locked-until date. After that date the classification returns to what the history says. The documentation names new parts with a forecast but no sales history as the case it is meant for.
How much history does the classification use?
The Classification Periods field on the asset class sets the number of months of issue history to read. Seasonality is also set on the asset class, so in my view seasonal parts are best grouped in their own asset class.
8.About the author
Dariusz Myśliwiec brings 18 years of hands-on work with IFS (Apps 7.5–10 and IFS Cloud). IFS Certified Associate Consultant. PRINCE2® 7. Based in Kraków, delivering remotely across Europe and globally as an independent practice, so you talk to the consultant who builds it.
Selected clients: Fugro · LGC · BVI Medical · Betafence (PRÆSIDIAD) · Barlinek · NGK Ceramics · Newag · Oleofarm.
IFS is a registered trademark of IFS AB; this practice is not affiliated with IFS AB.
Check your reorder points against how the parts really move
Tell me which part groups cause the most stock-outs or the most aged stock. On a 30-minute fit call we can look at how they are classified today, which parameters were set at which level, and whether IPR is the right tool for them.