The old FP&A playbook for a consumer brand is simple. Build a big workbook. Paste in last month's numbers. Spend a week making them tie. Send the board a deck nobody can double-click.
It works. Until it doesn't. And the point where it breaks is usually right when the business gets interesting: more channels, more SKUs, a retail launch, a board that wants answers by Friday.
Black Friday is November 27 and Cyber Monday is November 30. Right after, most brands sit down to build the 2027 plan. Most will do it the way they did last year. That's the opportunity.
FP&A best practices are the habits that keep planning, forecasting, and reporting accurate and fast: one source of truth, driver-based models, rolling forecasts, scenario planning, quick variance analysis, and a reporting cycle you can repeat. Everyone agrees on the list. The difference is how you run it.
Before Drivepoint, immi's co-founder spent 10 to 20 hours a month copying financials across spreadsheets. Now he spends zero time updating the historical model. Same practices. Different mechanics.
Here are seven, each as the old way versus the new way, for a brand selling across Shopify, Amazon, and retail. I've seen this pattern across a lot of consumer brand models, and the gap is almost never knowledge. It's maintenance.
What are FP&A best practices for a CPG brand?
They are the same core habits every finance team is told to follow, applied to the specific mess of consumer brands: retailer data that lags, trade spend and deductions, SKU-level margin, and cash that moves with purchase orders and payment terms.
Key takeaways:
- Get to one source of truth for your data.
- Build driver-based models, not hard-coded ones.
- Run rolling forecasts and keep actuals current.
- Plan scenarios in one model, not one file per question.
- Track plan versus actual and explain variance fast.
- Make board and investor reporting a same-day task.
- Put leverage ahead of headcount.
None of this requires leaving Excel. We are Excel-native for a reason. The new way is Excel with live data and structure, not a rip-and-replace.
1. Get to one source of truth for your data
Old way: Exports from Shopify, Amazon, retailer portals, and accounting get pasted into tabs. Some tabs link to external files. Nobody is sure which version is right, and everyone has a different number.
New way: A connected data layer feeds the model, so actuals are current without manual actualizing. One set of numbers, one definition of net sales, one place to look.
Data quality is where most finance AI and automation efforts stall. Gartner's 2025 survey of 183 finance leaders named data literacy and data quality among the biggest obstacles. Fix the data layer first and every other practice gets easier.
Laundry Sauce pulls Shopify, Amazon, and QuickBooks into one hub and reports 98% forecast accuracy with Drivepoint. Their CEO put it plainly: if your model isn't accurate three months out, "you're living in fantasy land." See how to connect Shopify, Amazon, and accounting data in one place.
2. Build driver-based models, not hard-coded ones
Old way: Hard-coded numbers, hidden tabs, and a model only the person who built it understands. Changing one assumption means hunting through the workbook.
New way: Key drivers feed results. Velocity, doors, price, trade rate, and CAC sit in one place, and anyone can see what moves what. When someone asks "what changed?", the answer is visible.
If you want the method behind this, our guide to driver-based modeling for consumer brands goes deeper.
3. Run rolling forecasts and keep actuals current
Old way: The annual budget gets locked in January. Reforecasts happen by hand when someone has time. You end up with weekly snapshots and never really compare to actuals.
New way: The model rolls forward after close, and plan versus actual is always on. You reforecast because the numbers changed, not because it's quarter end.
SEEQ moved from quarterly planning to real-time planning. Forecasts that used to take two to three days are now instant, and reporting that ran 30 to 60 days behind now runs on live data. See how rolling forecasts update themselves.
4. Plan scenarios in one model, not one file per question
Old way: Copy the workbook, change a cell, and compare two Excel docs side by side. By the third question you have v7_FINAL and no idea which one the board saw.
New way: Toggle scenarios inside one version and see revenue, margin, and cash move together. What if the Target PO ships late? What if ad costs go up 20%? You get the answer in minutes, not a new file.
Lalo cut its time to scenario by 97%, from months to days, and the two founders save 40 to 60 hours per forecast cycle. Learn how to run scenarios in minutes.
5. Track plan versus actual and explain variance fast
Old way: Variance gets explained a week after close by whoever has time. By then the decision it should have informed is already made.
New way: Variance reporting is ready at close, and the drivers behind the miss are already visible: price, volume, mix, timing. You spend the meeting on what to do, not on why the numbers differ.
immi lowered its actual versus budget variance by 50% through better forecasting accuracy. As the co-founder put it: "Now, when month close happens, I'm not exhausted from accounting work."
6. Make board and investor reporting a same-day task
Old way: Weeks of prep, manual slides, and numbers that differ from deck to deck. Every board question starts a new analysis.
New way: Reports come straight from the live model, so the board can ask a follow-up and the answer is already there. Lalo went from weeks of prep to same-day board readiness. Explore same-day board reporting.
7. Put leverage ahead of headcount
Old way: Hire an analyst or a fractional CFO to maintain the workbook. Wait months for them to build infrastructure, then keep them busy keeping it alive.
New way: A platform handles the mechanics so your finance people spend their time on decisions. A platform does not replace judgment. It gives good judgment something solid to stand on.
Ibex's team now makes model changes in about half an hour, versus three to four hours when someone with the right expertise was available, and reports $314K in annual finance personnel savings. Platform first, people second. I wrote about why finance teams are going platform-first and how to think through every objection.
Old way vs. new way: the summary
| Practice | Old way | New way | What changes for a CPG brand |
|---|---|---|---|
| One source of truth | Exports pasted into tabs, external links | Connected data layer, live actuals | Shopify, Amazon, retail, and accounting agree |
| Driver-based models | Hard-coded, hidden tabs | Visible key drivers feed results | Velocity, doors, price, and trade rate in one place |
| Rolling forecasts | Annual budget, manual reforecast | Model rolls forward after close | Plan versus actual always current |
| Scenario planning | One file per question | Scenarios in one version | PO timing, price, and ad cost tested in minutes |
| Variance analysis | Explained a week after close | Ready at close | Price, volume, mix, and timing visible |
| Board reporting | Weeks of prep | Same day from the live model | Follow-up questions answered in the room |
| Leverage over headcount | Hire to maintain the workbook | Platform handles mechanics | Finance time goes to decisions |
A checklist for 2027 planning
- Data sources connected, with one definition of net sales and margin.
- Key drivers documented and visible in the model.
- Actuals current through the last close.
- Base, upside, and downside scenarios saved in one version.
- A named owner for variance explanations after every close.
- A board pack template that pulls from the model, not from a deck.
One honest note. If you run one channel with a handful of SKUs, a well-built spreadsheet can carry many of these practices for a long time. The signs you've outgrown it are multiple channels, retail terms, a growing SKU count, and investors or lenders asking for numbers you can defend. When those show up, the maintenance cost starts to exceed the cost of fixing the mechanics.
The best practices haven't changed. How much of your week they take has.
Frequently asked questions
What are FP&A best practices?
FP&A best practices are the habits that keep planning, forecasting, and reporting accurate and fast. The core set is one source of truth for data, driver-based models, rolling forecasts, scenario planning, fast variance analysis, and a repeatable reporting cycle. For a consumer brand, that also means handling trade spend, SKU-level margin, and multiple channels in the same model.
What are FP&A best practices for CPG and consumer brands?
The same practices apply, but CPG adds complexity: retailer data with sell-in and sell-through, trade spend and deductions, SKU-level margin, and cash timing driven by purchase orders and payment terms. A CPG model has to connect Shopify, Amazon, retail, and accounting data so actuals stay current without manual copying.
Should an FP&A team use Excel or software?
Both. Excel is a good modeling environment, and many teams should keep it. The problem is maintaining the data, links, and versions around it by hand. The better setup keeps the model in Excel and connects it to live data so actuals, scenarios, and reports update without copy and paste. A small brand with one channel can run many of these practices in a well-built spreadsheet.
How often should a consumer brand reforecast?
At least monthly after close, and whenever a major input changes, such as a retail PO, a price change, or a shift in ad costs. A rolling forecast that updates as actuals land makes this a routine step instead of a quarterly project.
What is modern FP&A?
Modern FP&A means connected data, live actuals, driver-based models, and scenarios that run in minutes, with finance time spent on decisions instead of data maintenance. It keeps the judgment of a good finance lead and removes the manual work around the model.
