Cash Flow Forecasting
for startups

Detailed cash flow projections for UK startups navigating growth phases and funding rounds. Strategic financial planning incorporating market conditions, seasonal variations, and realistic business cycles.

Most UK startups need two cash flow models running simultaneously: a 13-week rolling weekly model that gives operational warning of cash crunches before they happen, and an 18-month strategic model used for investor decks, hiring decisions, and runway calculations. The two models work together but serve different purposes, and neither is a substitute for the other.

Forecasts are only as good as the assumptions underneath them. Specialist startup accountants build models that bake in the timing realities founders typically miss: VAT payments, R&D credit cash receipts, payroll cycles, customer payment delays, and the working capital impact of growth itself.

The single most important cash flow tool for an early-stage UK startup is a 13-week rolling cash flow model, updated weekly. The 13-week horizon is operational - long enough to spot a cash squeeze before it becomes a crisis, short enough that the projections can be grounded in confirmed inputs (signed contracts, expected payments from named customers, scheduled payroll runs, known supplier invoice dates) rather than speculation. The model starts from the current bank balance, layers in expected receipts week by week with confidence indicators, layers in expected payments week by week with timing precision, and projects the closing bank balance for each of the next 13 weeks. The output is a single number per week: how much cash you'll have on Friday afternoon. Founders who maintain this discipline consistently catch cash issues four to eight weeks before they would have without the model, which is the difference between a planned action (negotiate supplier terms, accelerate a customer payment, draw down a facility) and a crisis (emergency bridge round at punishing terms).

The 18-month strategic model serves a different purpose. It's the model used for investor decks, hiring plans, runway calculations, and board reporting. The horizon is long enough to capture the full investment-round cycle (raise, deploy, achieve milestones, raise again) and the timing of major business events (product launches, geographic expansion, key hires, exit prep). The model is updated monthly with the previous month's actuals, with the projections refined based on what actually happened versus what was forecast. Variance analysis between forecast and actual is the most important output for the founder - it shows where assumptions were wrong and what to adjust. A company that runs the 18-month model in isolation, without the 13-week operational backstop, typically discovers cash gaps several weeks late because monthly granularity isn't fine enough to spot week-to-week swings caused by VAT cycles, payroll dates, and supplier payment terms.

Revenue modelling for early-stage startups is harder than expense modelling because revenue depends on uncertain external events (customer signings, payment timing, churn) while expenses are largely controllable internal decisions. The discipline that separates good founder forecasts from optimistic ones is sales pipeline grounding: each forecasted revenue line tied to a named customer (for B2B), a measurable acquisition channel (for B2C), or a contractual relationship (for marketplace or platform fees). Forecasts based on top-down market sizing ('we'll capture 0.5% of a £10bn market in year three') are not useful operationally - they're investor-pitch fiction. Forecasts grounded in named pipeline ('these eight prospects expected to close at this expected ACV with this expected timing') give the model the discipline to be useful. The best founders rerun the pipeline view weekly with sales updates and update the cash flow model accordingly.

Expense modelling needs to capture the timing realities that catch founders out. UK payroll runs typically on the 25th-28th of each month, with PAYE due by the 22nd of the following month. VAT returns are quarterly with payment due on the 7th day of the second month after quarter-end. Corporation tax is due 9 months and 1 day after year-end. R&D tax credits, when they arrive, typically pay 4-6 weeks after submission of the corporation tax return. SEIS and EIS investment closes typically deposit cash within 1-2 weeks of legal completion. Trade supplier terms vary widely but most B2B suppliers default to 30-day terms. The model should bake in all these timing rhythms so the cash position reflects the actual timing of inflows and outflows, not just the periods they relate to. A common founder error is to recognise revenue when invoiced and expenses when invoiced, ignoring the cash-timing reality that customers pay 30-90 days late while suppliers expect payment in 14-30 days.

Working capital modelling is where rapid-growth startups get into trouble. As a startup grows, it typically funds growing levels of accounts receivable (customers who haven't yet paid), inventory (for ecommerce or hardware companies), and prepaid expenses (annual SaaS commitments, advance hires). All of these consume cash even though they don't show up as expenses in the P&L. A company growing 30% month-on-month with 60-day customer payment terms will see its cash position deteriorate even while the P&L looks profitable, because the receivables grow faster than the payables can be stretched. The cash flow model should explicitly forecast working capital as a function of revenue growth, with assumptions about days sales outstanding, inventory days, and payment terms. The working capital line is often the single largest swing factor in growth-stage forecasts.

Sensitivity analysis on the cash flow model is what makes it useful for decision-making. The base case forecast tells you what's expected; sensitivity analysis tells you what's at risk. Standard sensitivity scenarios for a startup model include: revenue at 70%, 85%, 100% (base case), and 115% of plan; sales cycle 50% slower than plan; hiring freeze (no new hires after current roster); cost reduction (10% across the board); R&D credit delayed by three months. Each scenario's runway calculation tells the founder how much margin for error the current cash position provides. A forecast where the base case shows 18 months of runway but the 85% revenue scenario shows 10 months tells the founder that revenue is the critical variable and needs management attention. A forecast where every scenario shows 14 months or more says the company has comfortable runway and the focus should be on growth rather than survival.

R&D credit cash receipts are typically the largest single annual cash inflow for pre-revenue R&D-heavy startups. Modelling them correctly matters more than for any other line. The credit is typically claimed at year-end, with the cash payment from HMRC arriving 4-12 weeks after submission of a clean claim. For a startup with a March year-end, this means the cash typically arrives between July and October of the following calendar year, six to nine months after the qualifying R&D was performed. Founders frequently model the credit as arriving in the same month as the year-end, ignoring the 6-9 month lag, leading to cash shortfalls in months 4-9 of the new financial year. The model should always carry the R&D credit cash inflow at the realistic timing, with a contingency assumption for HMRC enquiry delay (which can extend the receipt by another six months).

Multi-currency operations introduce FX timing risk into the cash model. A UK startup with US-dollar customer revenues and GBP-denominated payroll is exposed to GBP/USD movements between invoicing and collection. The model should forecast revenue in original currency, apply an FX assumption (usually a sensible average rate or the forward rate where the company hedges), and convert to GBP for the cash position. Forecasting in mixed currencies without explicit FX assumptions creates a hidden volatility in the cash position that founders only notice when sterling weakens. Companies with significant non-GBP revenue should review FX hedging at quarterly intervals.

Annual subscription billing creates deferred revenue that distorts cash flow if not modelled carefully. A SaaS startup billing annually receives twelve months of cash upfront but recognises revenue monthly, creating a deferred-revenue balance that builds up on the balance sheet. The cash position is positive even when the P&L shows a small loss, because the cash has already been received but the revenue hasn't been recognised. Founders modelling cash flow from P&L (instead of from the cash bookings) miss this entirely and underestimate runway. The model should track booked cash separately from recognised revenue, with deferred revenue as an explicit line on the balance sheet roll-forward.

Investor capital calls and tranched funding need explicit modelling. Many SEIS/EIS rounds and most institutional Series A rounds are released in tranches based on milestones (e.g. £500k at signing, £500k on hitting an MRR milestone, £500k on closing the next round). Modelling the full round as a single inflow at signing overstates available cash and leads to overcommitment of expenditure. The model should reflect the tranche structure, with each tranche conditional on achievement of the milestone. If the milestone is missed or delayed, the model needs an explicit scenario showing the cash position without the missed tranche.