When Your Demand Plan Is Wrong Before the Ink Dries

At 500 SKUs and two retail channels, you can hold the demand picture in your head. The commercial team knows what's selling. Finance can reconcile trade spend manually. The supply chain lead talks to the planner because they sit three desks apart.

At 3,000 SKUs across eight channels with seasonal promotions, NPD launches, and a competitor set that moves weekly, that same approach falls apart. Not dramatically. Gradually, then all at once - usually around September when you're trying to land the annual plan while simultaneously managing Christmas stock builds.

What Actually Breaks in FMCG Planning

We've worked with consumer goods businesses from GBP 50M turnover to GBP 2B+. The breaking points are remarkably consistent.

Your demand forecast is stale within days. You build a statistical baseline, overlay commercial judgement, agree the numbers in a Monday S&OP meeting, and by Thursday the picture has shifted. A competitor promotes. A retailer changes a listing. Weather moves unexpectedly and your chilled category volumes swing 20% in a week. Your planning process assumes stability, but FMCG demand is anything but stable. So the forecast becomes a fiction that everyone politely ignores while making decisions based on gut feel.

Promotional spend disappears into a spreadsheet and never comes back. You're spending 15-25% of gross revenue on trade promotions. That's your single biggest controllable cost after COGS. And yet in most FMCG businesses we see, promo ROI is essentially a guess. The plan goes into a spreadsheet. Actuals come from a different system - sometimes weeks late. Nobody reconciles the two with enough granularity to know which promotions actually drove incremental volume and which just pulled forward sales that would have happened anyway. Finance reports total trade spend as a percentage of revenue. The commercial team reports volume uplift. Neither number tells you whether you actually made money on the promotion.

Supply chain and finance are looking at different numbers. The supply chain team plans in cases and pallets using a demand signal from the commercial forecast. Finance plans in pounds using a revenue build from the annual budget. These two views of the world should reconcile perfectly. They almost never do. The result: finance is surprised by margin variances that supply chain saw coming three months ago, and supply chain is building stock to a demand signal that finance has already revised downward. In one business we worked with, the gap between the supply chain volume forecast and the finance revenue forecast implied a 12% price increase that nobody had agreed to.

SKU proliferation makes driver-based planning nearly impossible. You've got 3,000 SKUs. Your planning model treats them all the same - or worse, it doesn't plan at SKU level at all, just at category level, which means your forecast accuracy at the level that actually matters for production scheduling and stock allocation is terrible. Every NPD launch adds complexity. Every range review is supposed to simplify things but rarely does because delisting a slow-moving SKU is politically harder than launching a new one. The result is a long tail of low-volume products that consume disproportionate planning effort and warehouse space.

Why Generic EPM Implementations Fail in FMCG

Here's the pattern we see too often. The CFO signs off on a planning platform. The implementation partner builds a standard P&L planning model - top-down revenue, cost centre budgets, headcount planning. It works fine for the back-office functions. Then they get to commercial planning and everything stalls.

Because FMCG commercial planning isn't a P&L exercise. It's a volume exercise. Revenue is the output of volume multiplied by price multiplied by mix, minus trade spend, minus waste, minus returns. The drivers are physical: cases per store per week, distribution points, promotional frequency, cannibalisation rates, seasonal indices. A standard financial planning model doesn't speak this language.

The other failure mode is trying to replicate the existing spreadsheet in the new platform. If your current process is a mess of disconnected files with inconsistent assumptions, rebuilding that mess in a more expensive tool doesn't fix anything. It just makes the mess harder to change.

Regardless of which platform you choose, the implementation has to start with the commercial planning process, not the financial reporting structure. Get the demand and promotional planning right, and the financial view falls out naturally. Start with the financial view and you'll never connect it properly to what's actually driving the business.

What Actually Works

Driver-based planning at the right level of granularity. Not at SKU level for everything - that's a trap. You need SKU-level planning for your top 200 products (which probably represent 70-80% of your volume). For the long tail, plan at sub-category level with a mix assumption. This keeps the model manageable while giving you accuracy where it matters. The drivers should be physical: baseline rate of sale, distribution, promotional uplift, seasonality. Price and margin flow from there.

Connected S&OP that actually connects. The demand plan, supply plan, and financial plan should be different views of the same model, not three separate spreadsheets that get reconciled once a month in a meeting. When commercial updates the volume forecast for a promotional change, supply chain should see the stock implication immediately and finance should see the margin impact in the same cycle. This isn't aspirational. It's achievable in 10-12 weeks with a focused build.

Real-time promotional tracking. Every promotion should have a planned volume, planned spend, and planned ROI before it goes live. Actuals should flow back within the same week - not the same month. After 12 months you'll have a dataset that actually tells you which promotional mechanics work, which retailers deliver genuine incrementality, and which promotions you should stop running. That dataset is worth more than the planning platform itself.

A rolling forecast that moves weekly, not monthly. In FMCG, a monthly forecast cycle is too slow. By the time you've collated inputs, reviewed assumptions, and published a number, the demand picture has moved. A weekly rolling forecast on the top 200 SKUs - updated by exception rather than from scratch - gives you a current view without drowning the team in process.

Signs You've Outgrown Your Process

These aren't hypothetical warning signs. They're the reality in most FMCG finance teams we talk to.

Your demand forecast accuracy is below 70% at SKU level, but nobody measures it formally because the number would be embarrassing.

You can't answer the question "what was the ROI on last quarter's promotions?" without two weeks of manual analysis.

Your S&OP meeting spends more time reconciling numbers than making decisions. Half the attendees bring their own version of the truth.

You've got a finance team that costs GBP 400K+ per year and they spend most of their time compiling data rather than analysing it.

The annual plan takes 16 weeks and is out of date before it's approved because the retail landscape moved while you were planning.

Your supply chain holds two weeks of safety stock "just in case" because they don't trust the demand signal. That's working capital you're burning because your planning process isn't good enough.