Continuous forecasting: How finance keeps up with volatility

Key takeaways

• With calendar-based forecasting, you make decisions for this month based on last month's assumptions.

• Continuous forecasting treats forecasts as live inputs that update as activity changes.

• Airwallex's unified infrastructure across payments, FX, accounts, and spend gives finance teams the real-time visibility that makes continuous forecasting practical.

The first law of financial forecasting is that assumptions have a shelf life. No matter how rigorous your modelling process is, assumptions age the moment you commit them to a spreadsheet. A forecast built in October reflects October's reality. By March, that reality has already shifted, even if your budget hasn't.

Most teams work around this lag and try their best to accommodate. You might build in buffers, hold quarterly reforecasts, and spend board meetings explaining the gap between forecast and reality. These measures all manage around the problem without solving the real issue: traditional forecasts are stuck in the past and don’t update fast enough.

What we should be working towards are continuous forecasts that update themselves as the business moves.

Forecasting grew up in a slower, more predictable world

Calendar-based forecasting made sense in the nineties, when finance ran on paper ledgers and manual record-keeping, supply chains took weeks to adjust, and interest rates moved in careful quarter-point increments. You could spend the final quarter building a budget, lock it in December, and trust those assumptions would hold for 12 months.

The forecast wasn't built to track what was actually happening in the business. It served board meetings and reporting deadlines, following a rhythm borrowed from construction projects where everything moved in sequence: plan it, build it, deliver it, review it. Once a quarter, you'd gather the actual numbers, explain why they differed from the forecast, present it to the board, and move on. Few people questioned the lag between transaction and report because markets moved slowly enough that last quarter's snapshot remained relevant.

Even when spreadsheets replaced ledgers in the eighties and nineties, the philosophy didn’t change all that much. Plans were still developed top-down using fixed assumptions and best guesses instead of real transaction data. Budgeting season still ate up the final quarter of every year. The targets were outdated before the fiscal year even began, but updating them took so much work that quarterly refreshes seemed good enough. That world is gone.

Calendar-based forecasting can’t keep up with volatility

92% of finance leaders say accurate forecasting is now their biggest challenge.¹ The problem is structural, and here’s what that could look like in practice:

1. It’s Q3 forecast season. You've budgeted €180,000 for a Black Friday marketing push based on last year's returns, assuming inventory costs stay stable and conversion rates hold steady.

2. Week two of October: your primary supplier announces a 22% price increase due to shipping disruptions, adding €95,000 to your quarterly cost base.

3. Week three: Meta changes its algorithm and wrecks your ad performance. Conversion rates drop 40% overnight.

4. Your marketing director sends you a Slack message wanting to double the paid budget to compensate.

Do you approve the spend? The forecast says you have the runway, but you've lost margin on every unit and your acquisition costs have spiked. If you approve that budget increase, you'll be scrambling to make payroll in December.

This is typical of the systematic lag that calendar-based forecasting creates. By the time your next scheduled refresh arrives, you've already made decisions based on assumptions that no longer hold.

Forecasting works best as a live signal

You can probably see where this line of thinking leads to. Instead of building a budget that covers January to December and then starting over, with continuous forecasting you maintain a forecast that always looks 12 to 18 months ahead. As one month closes, you add a new month at the far end. Your planning horizon stays constant, but the assumptions refresh based on what's actually happening in the business.

The forecast becomes something you check daily, not quarterly. Sticking with the eCommerce example above, if your platform shows conversion rates dropping or an increase in chargebacks, the forecast reflects those changes immediately. You can reallocate marketing budget, adjust inventory orders, or pause hiring plans whilst you still have room to manoeuvre. Nobody's waiting for the quarterly review to discover the damage that has already been done.

The goal shifts from achieving accuracy on paper to making the forecast useful for the decisions you make today.

The practical upsides of continuous forecasting

Continuous forecasting delivers three operational improvements: you catch problems earlier, you spend less time rebuilding models, and you redirect resources based on the current reality instead of defending last quarter's variance.

Earlier visibility into cash flow and exposure problems

A quarterly cash forecast might show a healthy runway, but that forecast was likely built on assumptions from six weeks ago. If a major customer delayed payment or the euro moved against you mid-month, you're working with outdated numbers. Continuous forecasting shows you these shifts as they happen. Which is why finance teams using this approach report 77% confidence in their numbers, compared to far lower trust in annual budgets.²

Faster decisions with higher confidence

When your forecast updates monthly instead of quarterly, you can act quickly. A product line underperforms? Reallocate that budget within weeks. A new market opportunity opens? You can move resources before the window closes. Leadership makes faster decisions because they trust the numbers reflect this month's immediate reality.

Less time maintaining models, more time on strategy

Finance teams currently spend 46% of their time collecting and validating data.³ That's nearly half your week hunting down numbers and reconciling systems. Continuous forecasting automates the vast majority of that work through direct integrations. With Airwallex, payments, FX conversions, corporate card spend, and accounts all run through one platform that feeds directly into your accounting software. A customer pays, currency converts, someone uses a corporate card, an invoice gets marked as paid in Xero, and the forecast updates automatically.

That frees your team to focus on strategy, market analysis, and directing resources where they drive the most value.

Smarter resource allocation throughout the year

Although less common in tech, many companies still see departments rush to spend their allocated budgets in November and December. Marketing might push campaigns that could wait until Q1. Operations might buy software licences they won't use for months. Everyone's playing the same game: spend it now or lose it next year.

Continuous forecasting treats budgets as fluid. You could shift resources toward high-impact initiatives as conditions evolve, and adjust underperforming areas mid-cycle rather than defend it through quarterly reviews.

How to move to continuous forecasting

Moving to continuous forecasting requires changing how your organisation collects, updates, and acts on financial data. Here are practical steps to get started:

1. Decide how far ahead you need to see

Most businesses maintain a rolling 12-to-18-month forecast, though the right horizon depends on your planning cycle and capital requirements. SaaS companies with annual contracts often extend to 18 months for visibility into renewal cycles. eCommerce businesses managing seasonal inventory usually focus on 12 months to stay close to demand patterns.

2. Set a realistic refresh rhythm

Monthly updates suit businesses exposed to currency movements, seasonal demand, or rapid market shifts. Quarterly updates work for more stable industries. The rhythm matters less than maintaining it reliably.

3. Establish cross-functional data flows

Revenue forecasts require pipeline data from sales. Headcount projections need hiring plans from department leads. Spend forecasts depend on understanding upcoming commitments across teams. Set up regular touchpoints with operations, marketing, and HR, with clear expectations about timing and format. Get buy-in early, because some teams will push back on monthly updates.

4. Automate data consolidation

Manual consolidation is where most finance teams lose time. Data flows from payment systems, expense platforms, and accounting software directly into your forecast model without requiring CSV exports and spreadsheet rebuilds. This frees you to focus on what the numbers mean.

5. Bring in leadership early

Continuous forecasting changes how the organisation makes decisions. The board needs to understand the trade-off: less precision in long-range projections in exchange for accuracy in near-term decisions and the ability to adjust as conditions change. Some leadership teams struggle with that trade-off because they want both precision and flexibility, but you can't have a forecast that's both perfectly accurate 18 months out and responsive to changes that happened yesterday.

Continuous forecasting works best on unified infrastructure

Everything described above assumes something that most finance teams don't have: the ability to see your financial position update in real time. If payments run through one provider, FX through another, expenses through a third system, and accounts sit with a traditional bank, you're manually consolidating data days or weeks after transactions settle. The model doesn't matter if the data feeding it is always playing catch-up.

Continuous forecasting works best when cash, payments, FX, and spend flow through one system that updates your forecast as money moves. This unified view also sets the foundation for what comes next: AI in finance. Over the next few years, AI will shift from being a tool that analyses historical data to one that actively shapes financial decisions. AI tools are starting to transform finance operations, but they can only work with the data they can see. When transaction data is scattered across disconnected systems, AI analyses whatever you manually export and stitch together. That means it's learning from your approximations rather than your actuals. When every payment, conversion, and approval flows through unified infrastructure, AI can learn from actual patterns, surface early warnings about cash flow or FX exposure before they compound, and help finance teams shift from explaining results to shaping outcomes.

Airwallex is building this infrastructure now. Payments, FX, accounts, and spend operate on a single platform where transactions flow through local rails and settle quickly enough for forecasts to reflect current reality. Your finance team can spend less time chasing down numbers and more time deciding what those numbers mean for your business.

Sources:

1. https://www.pwc.com/us/en/executive-leadership-hub/library/election-insights-2024-cfo.html

2. https://fpa-trends.com/article/unlocking-strategic-value-fpa-guide-data-harvesting

3. https://fpa-trends.com/article/re-imagining-expense-fpa