The short version: For decades, the supply chain industry has sold expensive planning software as the answer to better planning. But the real constraint was never the planning engine — it was data availability, accessibility, and granularity. Once your data lives in a properly structured foundation like Microsoft Fabric, the planning layer becomes something you can build with AI and statistical tools, not something you license for millions.
Ask most manufacturers or distributors how they will improve planning, and the answer usually involves a platform: Blue Yonder, Kinaxis, o9, or SAP IBP. Each is capable, and for the right organization each is worth it. But each also carries the same quiet assumption — that better planning requires buying a better planning engine. That assumption is worth questioning, because it gets the bottleneck backwards.
The assumption worth questioning
Walk into most planning teams and the engine is not what slows them down. What slows them down is that the data needed to plan is scattered, stale, or trapped. Delivery times live in the ERP. Inventory positions live in the ERP too, but in a slightly different corner of it. Shipment histories are somewhere else. The forecasts retailers send arrive as spreadsheets. The lead-time realities from factories arrive as different spreadsheets. Someone spends the first few days of every planning cycle piecing it together by hand — and by the time it is assembled, it is already a little out of date.
The planning engine was never the problem. The data was.
What "pieced together from the ERP" actually looks like
Consider a national distributor running a large, capable ERP. On paper they have everything: inventory, shipments, receipts, order history, delivery performance. In practice, answering a real planning question means pulling from several ERP modules, exporting to Excel, reconciling against a couple of external spreadsheets, and stitching the picture together manually.
It works. Teams do it every week. But it carries three costs that compound:
- It is slow. The assembly work eats the time that should go to analysis and decisions.
- It is fragile. The picture depends on the person who knows how to build it, and breaks when they are out.
- It is shallow. The ERP holds what the ERP holds. The moment you want to augment it — a factory's actual delivery-time history, a retailer's forward forecast, a web source, a freight or commodity signal — there is nowhere natural to put it. So it gets bolted onto a spreadsheet, or it gets left out.
That third cost is the important one. The ERP is a system of record for transactions, not a foundation for analysis. It was never designed to be the place where you blend internal truth with external signal at the granularity that planning actually needs.
Microsoft Fabric as the foundation
This is the gap Microsoft Fabric fills. Rather than a planning application, Fabric is a unified data foundation — a place where all of it lands, stays current, and stays accessible at the grain you need.
The move is straightforward to describe: instead of piecing the picture together each cycle, you build the repository once. ERP data — inventory, shipments, deliveries, receipts, order history — flows in continuously. Then you augment it with the sources the ERP was never going to hold:
- Factory data — actual delivery-time performance, production schedules, lead-time reality versus lead-time promise.
- Retailer and customer forecasts — the spreadsheets that arrive by email, brought into the foundation instead of taped onto the side of it.
- External signal — web sources, commodity or freight indices, and anything else that improves the picture.
The result is a single, current, granular view of the supply chain — not assembled by hand each week, but maintained as a foundation. For the underlying structure — ingestion, storage layers, and modeling — see our work on building resilient data pipelines, which describes the same foundation in more technical detail.
Why the foundation is the point — not the destination
Here is where the argument turns, and where it matters most.
A data foundation like this is immediately useful on its own. It gives you visibility. It powers dashboards. It answers "where do we actually stand" without a multi-day fire drill. For many organizations that alone justifies the work.
But visibility is not the real prize. The real prize is that once the data is available, accessible, and granular, everything you would want a planning platform to do becomes something you can build on top of the foundation — rather than something you buy as a monolith.
Statistical forecasting, safety-stock optimization, replenishment logic, demand sensing, scenario modeling — none of these are magic locked inside enterprise software. They are analytical methods. What makes them hard to run in-house has never been the math; it has been getting clean, current, granular data into one place to run the math against. Solve the foundation, and the planning layer is buildable.
And this is exactly where AI changes the economics. Modern AI and statistical tooling can be assembled into the planning functions that used to require a licensed engine — provided they are sitting on a solid data foundation. AI without good data is a demo. AI on top of a well-built Fabric foundation is a planning capability. The order matters: data foundation first, intelligence layer second.
The economic argument, stated plainly
Put those together and the conventional path inverts:
| The old path | Buy an expensive planning platform, then spend months feeding it data so its engine can run. |
| The alternative | Build the data foundation first, then layer planning and AI on top — statistical tools, optimization, and AI functions built around your actual process. |
You get the planning outcomes — consensus demand, forward inventory positions, recommended buys, exception surfacing — without the cost and multi-year implementation of a monolithic platform, and without contorting your process to fit a generic system. The foundation is reusable, the intelligence layer is tailored, and the whole thing is built around how your business actually plans.
Data availability, accessibility, and granularity are the real keys to planning. Get those right, and the software question gets a lot smaller.
Where this fits — enterprise and growing brands alike
For large manufacturers and distributors, Fabric becomes the analytical foundation the ERP was never meant to be — the place where internal records and external signal finally live together at planning grain.
For growing apparel and consumer brands, the same principle is what makes modern planning affordable at all. These companies have outgrown spreadsheets but cannot justify a million-dollar enterprise system. A Fabric foundation that consolidates their ERP, retailer forecasts, and factory data — with a planning application built on top — delivers what used to require enterprise software, at a fraction of the cost and timeline. More on that in our piece on the mid-market apparel planning gap.
Different scale, same truth: the foundation is what unlocks the planning.
If you are still weighing licensed platforms against a build-on-foundation approach, our S&OP software selection guide walks through how BlueYonder, Kinaxis, o9, and Fabric compare, and our earlier piece on using Microsoft Fabric for S&OP covers where Fabric fits in the planning process itself. For a worked example of a lightweight planning cycle running on this kind of foundation, see agile S&OP for mid-market CPG brands.
Frequently asked questions
Does Microsoft Fabric replace my ERP?
No. Fabric does not replace your ERP — it sits alongside it. Your ERP remains the system of record for transactions; Fabric becomes the analytical foundation that pulls ERP data together with external sources for visibility, planning, and AI.
Does Fabric replace a planning platform like Blue Yonder or Kinaxis?
It changes the question. Rather than licensing a monolithic planning engine, you build the data foundation in Fabric and layer planning functions — statistical forecasting, optimization, AI — on top. For many organizations, especially those that cannot justify a seven-figure platform, that is a more affordable and more tailored path to the same outcomes.
What data sources can Fabric bring together?
Internal systems like your ERP (inventory, shipments, deliveries, order history), plus the sources an ERP was never built to hold: factory delivery-time and lead-time data, retailer and customer forecast spreadsheets, and external web or market signals.
Why is the data foundation more important than the planning software?
Because planning was never bottlenecked by the engine — it was bottlenecked by data that was scattered, stale, or too coarse. The analytical methods behind planning are well understood; the hard part is getting clean, current, granular data into one place to run them. Solve that, and the planning layer becomes buildable.
Can AI really do the planning that enterprise software does?
AI and statistical tooling can perform the planning functions that used to require a licensed engine — but only on top of a solid data foundation. AI without good data is a demo; AI on a well-built Fabric foundation is a genuine planning capability. Data first, intelligence second.