Bookkeeping tells you what already happened. Financial planning and forecasting is different: it's the forward-looking discipline of projecting revenue, costs, and — critically for inventory-heavy ecommerce businesses — cash needs, far enough in advance that decisions (how much inventory to buy, whether to hire, whether a new marketplace launch is affordable this quarter) are made deliberately rather than reactively. Many scaling sellers run excellent bookkeeping and still get regularly surprised by cash crunches, because bookkeeping and forecasting are genuinely different skills serving different purposes.

Why ecommerce forecasting is harder than it looks

A few features of ecommerce specifically complicate forecasting relative to, say, a services business:

  • Inventory has to be paid for well before it's sold. The gap between paying a supplier and collecting revenue on the resulting sales (and then actually receiving the cash, net of marketplace settlement delays) can be months, and it has to be forecast and funded in advance.
  • Marketplace payout timing lags actual sales. Revenue recognized this week isn't cash in the bank this week — settlement periods, reserves, and payment holds all create a gap between the P&L and the bank account that a pure revenue forecast misses.
  • Seasonality can be extreme. Many categories see a large share of annual revenue concentrated in a short window, which means a naive month-over-month forecast badly misjudges both the revenue ramp into peak season and the cash needed to fund the inventory buy that precedes it.
  • Multi-channel complexity. Different marketplaces have different fee structures, payout schedules, and growth trajectories, so a single blended forecast often obscures what's actually driving (or dragging on) results.

Building the core forecast

A workable forecast for a scaling ecommerce business generally has three linked layers:

  1. Revenue forecast, built bottom-up by product/SKU and channel where possible (unit sales × price), rather than top-down as a single growth-rate applied to last year's total. Bottom-up forecasting forces you to account for real constraints — inventory availability, ad budget by channel, known seasonality by product category — that a top-down percentage growth assumption glosses over.
  2. Cost forecast, split into cost of goods (tied directly to the unit forecast), marketplace/payment fees (see The Full Cost Stack of a Marketplace Sale), fulfillment/shipping, advertising, and fixed operating costs (payroll, software, rent if applicable).
  3. Cash flow forecast, translating the revenue and cost forecast (which are typically built on an accrual or "when it happens" basis) into an actual cash-in/cash-out timeline that accounts for marketplace settlement lag, supplier payment terms, and any financing (see Cash Flow Management at Scale for the specifics of this layer).

Scenario planning, not a single number

A single-point forecast ("we'll do $X in revenue next quarter") is less useful than a small set of scenarios, because it gives you decision-ready answers to "what do we do if growth is slower/faster than expected" before you're in the middle of finding out:

  • Base case — your best honest estimate given current trends
  • Downside case — meaningfully lower growth, or a specific risk materializing (an account health issue, a key product going out of stock, a competitor's price war)
  • Upside case — meaningfully higher growth, which has its own risk (can you actually fund and fulfill it if it happens?)

The upside case is worth taking seriously — a common failure mode for growing sellers isn't running out of money because sales are too low, it's running out of cash because sales grow faster than planned and there isn't enough working capital to buy the inventory needed to keep up.

A practical forecasting cadence

  • Weekly: A short cash position check — current balance, known upcoming payables (supplier payments, ad spend, payroll), and expected incoming settlements.
  • Monthly: Revisit the forecast against actuals, and roll the forecast forward another month. This is also the natural checkpoint for reviewing product/channel-level profitability.
  • Quarterly: A deeper review — reassess assumptions (growth rate, new product/channel launches, fee changes), and refresh the scenario set.
  • Annually: A full annual plan, including major initiatives (new marketplace launches, new product lines, team growth) mapped against the cash and revenue implications of each.

Common mistakes

  • Building the forecast once a year and not revisiting it. A forecast that isn't updated against actuals stops being useful within a quarter or two, especially for a fast-growing business.
  • Forecasting revenue without forecasting cash separately. Profitable-on-paper growth can still create a cash shortfall if the timing of inventory payments and marketplace payouts isn't modeled explicitly.
  • Applying a single blended growth rate across all channels/products instead of building the forecast bottom-up, which hides which specific channel or product is actually driving (or dragging on) the numbers.
  • Ignoring the upside case. Planning only for the downside leaves a business unprepared (out of inventory, out of working capital) if growth comes in ahead of plan.
  • Not involving whoever owns fulfillment/inventory in the forecast, so the sales forecast and the inventory buy plan drift out of sync with each other.

Best practices

  • Build the forecast at the SKU or product-category level where feasible, not just as a single company-wide number — this is what makes it actionable for inventory and ad-spend decisions.
  • Track forecast accuracy over time (how far off was last quarter's forecast, and in which direction) and use that to calibrate how much of a buffer to build into future forecasts.
  • Tie the forecast explicitly to inventory purchasing decisions and hiring decisions, rather than keeping it as a separate finance exercise that doesn't inform operational choices.
  • Use a rolling forecast (always projecting some fixed number of months or quarters ahead) rather than a static annual budget that becomes stale by mid-year.

FAQ

Do I need forecasting software, or is a spreadsheet enough? A well-built spreadsheet is genuinely sufficient for most sellers below a fairly substantial revenue and SKU-count threshold. Dedicated forecasting/FP&A software becomes worth the cost once the spreadsheet becomes too unwieldy to maintain accurately or too slow to update — that's a "when it hurts" decision, not a fixed revenue trigger.

How far ahead should I forecast? A rolling 12-month forecast is common, with the near-term months (next 1-3) built in much more granular detail than the far months, which are necessarily more approximate.

How does this differ from budgeting? A budget is typically a fixed plan set for a period (often a fiscal year) used to control spending against a target. A forecast is a continuously updated best estimate of what will actually happen. Mature finance functions use both — the budget as a spending discipline, the forecast as the current best read of reality — but a scaling seller building this for the first time can start with just a rolling forecast and add formal budgeting later.