Back to Blog

Fabric IQ Planning Sheets - Budgeting and Forecasting Inside Microsoft Fabric

November 20, 2026•9 min read•Michael Ridland

Every year around March, I watch the same thing happen at a dozen Australian organisations. Finance kicks off the budget cycle for the new financial year. Someone exports actuals from the data warehouse into Excel. That workbook gets copied 40 times and emailed to department heads. Three weeks later, finance is stitching together 40 versions of the truth, half of which have broken formulas and one of which has a hardcoded number someone typed in to make the totals look right.

Then the board asks for a revised forecast in October and the whole thing starts again.

Planning sheets in Microsoft Fabric IQ are Microsoft's attempt to fix this inside the data platform itself, rather than through yet another standalone planning tool bolted onto the side. I've been looking at them closely because we get asked about budgeting and forecasting constantly, usually by clients who've already put their reporting into Power BI and Fabric and are wondering why planning still lives in email attachments.

Here's what planning sheets are, where I think they fit, and where I'd be careful.

What planning sheets actually are

Planning sheets are the planning component of Fabric IQ. You get a grid that looks and behaves a lot like a spreadsheet, sitting on top of your governed Fabric data, where business users can enter budgets, forecasts and targets directly. Those inputs are written back and stored in Fabric, so they become data you can report on, rather than numbers trapped in someone's laptop.

The pieces that matter:

  • A familiar grid experience. Rows, columns, hierarchies, subtotals, inline formulas. Finance people can use it without first sitting through a course on how OLAP cubes work.
  • Connected to your semantic models. Planning sheets read from the same semantic models your Power BI reports use. Actuals come from the same definitions everyone has already agreed on.
  • Write-back. Users enter values in the sheet and those values are persisted in Fabric. This is the bit that Power BI on its own has never done well.
  • Planning logic. Things like allocations, distributing a top-down number across lower levels, copying prior periods forward, applying growth percentages, and comparing scenarios side by side.
  • Versions and scenarios. You can keep a budget, a reforecast and a "what if revenue drops 10%" scenario next to each other without three separate files.

If you've seen the Inforiver style of planning and reporting inside Power BI, a lot of this will feel familiar. Microsoft has been building out the planning story in Fabric IQ, and planning sheets are where it lands for the people who actually do the work.

Why this matters more than it sounds

The obvious pitch is "replace spreadsheets." I'm sceptical of that framing because spreadsheets have survived every attempt to replace them for 40 years, and for good reason. Excel is flexible, everybody knows it, and it doesn't need IT to approve anything.

The more interesting point is about where planning data lives.

When budgets live in Excel and actuals live in Fabric, you get a permanent reconciliation problem. Every variance report needs someone to load the budget back into the warehouse, usually through a manual upload, a SharePoint list or a fragile Power Automate flow that one person understands. We've built more of those integrations than I'd like to admit. They work until the chart of accounts changes or someone adds a cost centre mid-year.

With planning sheets, the plan is stored in the same platform as the actuals, against the same dimensions. Budget versus actual stops being an integration project. It's just a measure.

For Australian organisations specifically, this hits a few common pain points:

  • The July to June cycle. Plenty of tools assume a calendar year and treat fiscal years as an afterthought. Because planning sheets sit on your own semantic model, your fiscal calendar comes along for free if it's already modelled properly.
  • Multi-entity groups. A lot of mid-market Australian businesses are a handful of entities that got acquired over time, each with its own GL. If you've already done the hard work of conforming those in Fabric, planning sheets let you plan against that conformed view instead of rebuilding it in a planning tool.
  • Regulated industries. Health, financial services and government clients care a lot about who changed a number and when. Having planning inputs inside Fabric, with Fabric's security model and audit capabilities, is a much easier conversation with risk teams than 40 emailed workbooks.

What we've seen work

A few patterns are coming through from the planning work we've been doing with clients on Fabric.

Start with one process, not "planning." The projects that go well pick one specific cycle, like the operating expense budget for a single division, and get that working end to end. The ones that struggle try to model revenue, headcount, capex and opex across the entire group in the first release. Headcount planning in particular has enough edge cases (on-costs, super, part-time FTE, start dates mid-month) to swallow a project on its own.

Fix the semantic model first. Planning sheets inherit whatever is in your semantic model, good and bad. If your cost centre hierarchy is a mess, or your actuals measure double-counts intercompany transactions, planning sheets will expose that immediately and loudly. In one engagement we spent the first two weeks just cleaning up the account hierarchy before anyone opened a planning sheet. That was the right call. Our Microsoft Fabric consultants spend more time on model design than on any planning feature, and that ratio is about right.

Give finance ownership of the sheet layouts. IT should own the data model, security and the write-back storage. Finance should own the layouts, the formulas and the allocation rules. When IT tries to own both, every change to a budget template becomes a ticket, and finance goes back to Excel within a quarter.

Keep driver-based logic simple at first. It's tempting to build a beautiful driver model where revenue flows from volume times price times seasonality factor. Do that later. In the first cycle, let people enter numbers directly and use simple growth percentages. Once people trust the tool, add the drivers.

What's still rough

I'll be honest about the gaps, because some of our clients have been burnt by vendors promising planning tools that do everything.

It's not a full EPM suite. If you're a large enterprise running complex consolidations, currency translation with multiple rate types, intercompany eliminations and statutory reporting, planning sheets aren't a replacement for a dedicated corporate performance management platform. They're very good at operational planning and forecasting close to the data. They're not trying to be your consolidation engine.

Excel habits die hard. Some finance users will want to paste a block of 500 numbers in from a workbook they've maintained for six years. Test copy and paste workflows with your actual users early. If the experience is clunky for the way your team works, you'll know quickly.

Capacity and cost. Planning sheets run on Fabric capacity. Planning cycles are spiky: nothing for weeks, then 60 people hammering the system in the last three days before the deadline. If you're on a smaller F SKU that's already busy with refreshes, you need to think about whether the budget crunch will throttle your reports. Plan the capacity before the cycle, not during it.

Licensing and availability questions. Fabric IQ features have rolled out unevenly across regions and some parts have sat in preview for a while. Check what's generally available in the Australian regions you use before you commit a budget cycle to it. I would not run a board-reported budget on a preview feature without a fallback plan.

Governance needs a design. Write-back is great until three people are editing the same cost centre for the same month. You need to think about who can edit what, when versions get locked, and how approvals work. The tooling gives you the building blocks. It doesn't decide your approval process for you.

How I'd approach it

If you're an Australian organisation already running reporting on Power BI or Fabric and you're still doing planning in spreadsheets, here's roughly how I'd sequence it:

  1. Pick a target cycle. Ideally the next reforecast rather than the full annual budget. A reforecast is lower stakes and smaller in scope.
  2. Audit the semantic model. Make sure your actuals, hierarchies and fiscal calendar are correct and that finance agrees with the numbers. If you have doubts, our Power BI consultants can review it quickly.
  3. Build one planning sheet with one department. Sit with the actual users. Watch them use it. Note every time they reach for Excel.
  4. Run it in parallel once. Keep the old process going for one cycle while the new one runs alongside. Compare the outputs.
  5. Then scale. Add departments, add scenarios, add driver logic.

The parallel run is where most of the real learning happens. You find out that a particular manager plans at weekly granularity when everyone else does monthly, or that the marketing budget is allocated across regions using a formula nobody wrote down.

Where this fits with AI

Planning sheets also sit nicely alongside the other parts of Fabric IQ. Once your plan lives in Fabric next to your actuals, you can point a data agent at it and ask "which cost centres are tracking more than 10% over budget this quarter, and why?" and get an answer grounded in governed data. That's a much more useful AI use case than most of the chatbot projects we get asked about, because the data is structured, trusted and already has business meaning attached. If you want to explore that side of it, our work on AI for business intelligence covers how we approach it.

My take

Planning sheets are one of the more practical things Microsoft has added to Fabric. They don't do anything a dedicated planning tool can't, but they do it where your data already is, with the security and models you've already invested in. For mid-market Australian organisations that have outgrown Excel-based budgeting but can't justify a six-figure EPM implementation, that's a genuinely good fit.

Just don't expect the tool to fix a messy data model or an undefined approval process. Those are still your problems, and they're the ones that decide whether planning projects succeed.

You can read Microsoft's documentation here: Planning sheets overview.