# Budgeting and forecasting for a small business: zero-based, incremental and rolling Source: https://sarang.pro/budgeting-forecasting-small-business/ Author: Sarang | Published: 25 October 2026 Many small businesses either have no budget at all or have one that was built in January and forgotten by March. Neither helps much. A budget is useful when it sets a clear plan for spending and earnings, and a forecast is useful when it tells you, honestly, where you are likely to end up. You need both, and you need to compare them with what actually happened. This guide explains the main budgeting methods, how a rolling forecast works, and how to use variance analysis to learn something each month rather than just record a miss. ## Budget and forecast are different tools A budget is a plan for a fixed period, usually a financial year. It sets how much you expect to sell and how much each part of the business may spend. Its main jobs are allocating money and holding people to account. A forecast is an estimate of what will happen, updated as new information arrives. It does not need to match the budget. If a large customer leaves in April, the budget stays as it was, but the forecast should change straight away. Keeping the two separate matters. When managers are judged against a forecast they also write, they tend to make it easy to hit. When the budget is never updated, it stops being useful for decisions. Most small businesses do well with one annual budget for accountability and one rolling forecast for steering. ## Incremental budgeting Incremental budgeting takes last year's budget or actual spend as the starting point and adjusts it up or down. Marketing was £30,000 last year, so it becomes £33,000 this year to allow for growth and price rises. Its strengths are speed and stability. It needs few assumptions, it is easy to explain, and it suits businesses whose costs do not change much from year to year. Its weaknesses come from the same source. Because each budget builds on the last, past waste carries forward. Teams learn to spend their full allocation so they do not lose it next year. And little money is freed up for new ideas, because almost everything is already committed to existing activities. ## Zero-based budgeting Zero-based budgeting (ZBB) starts every line at zero. Each cost must be justified afresh for the coming period, whatever was spent before. The method was developed in the 1970s by Peter Pyhrr at Texas Instruments, and the US federal government later used it for a time. ZBB forces useful questions. Do we still need this software subscription? Could this service be delivered more cheaply? Which activities would we fund if we were starting today? It can expose spending that has survived only because nobody looked at it. The cost is time. Building every line from nothing takes far more effort than adjusting last year's numbers, and managers can still dress up favourite projects as essential. A practical middle path for a small business is to run ZBB on a few cost areas each year (for example software and subscriptions one year, marketing the next) and use incremental budgeting for the rest. IncrementalZero-based Starting pointLast year's figuresZero for every line EffortLowHigh Finds wasteRarelyOften Best forStable, predictable costsDiscretionary spend, turnarounds, new strategies ## Rolling forecasts A rolling forecast always looks the same distance ahead. A common format is twelve months, updated monthly. When a month closes, you drop it from the front, add a new month at the end and revise the months in between using the latest information. This fixes the main problem with an annual plan, which is that by the third quarter you are only looking a few months ahead. With a rolling forecast, in October you can still see next September. Keep it light. Forecast the few drivers that move your results (orders, average order value, conversion rate, headcount, key supplier costs) rather than every line in the ledger. Use less detail for months further away. And add a downside scenario, because the most useful thing a forecast can show you is a problem early enough to act on it. ## Variance analysis: learning from the gap A variance is the difference between what you planned and what happened. A result better than plan is favourable (F). A result worse than plan is adverse (A). A single revenue variance often hides two separate stories: you sold a different number of units, and you sold them at a different average price. Splitting the two tells you where to look. An example (illustration only): an online shop budgeted 2,000 orders in March at an average of £20, which is £40,000 of revenue. It actually took 1,900 orders at an average of £19.50, which is £37,050. The total revenue variance is £2,950 adverse. VarianceCalculationResult Volume(1,900 minus 2,000) multiplied by £20£2,000 (A) Price(£19.50 minus £20.00) multiplied by 1,900£950 (A) Total£37,050 minus £40,000£2,950 (A) Two-thirds of the shortfall came from fewer orders and one-third from a lower average price. Those point to different causes. Fewer orders might mean weaker traffic or conversion. A lower average price might mean more discounting or customers buying cheaper items. Each has a different owner and a different fix. Apply the same thinking to costs. If delivery costs came in over budget, check whether the cost per parcel rose or you simply sent more parcels. More parcels because sales were higher is not a problem. A higher rate per parcel is. ## Setting it up: a monthly routine - Build the annual budget from your sales drivers first, then costs. Decide which cost areas get a zero-based review this year. - Set up a twelve-month rolling forecast in the same format, so the two can sit side by side. - Close the month within the first week or so: get actuals from your bookkeeping software. - Compare actuals with budget and forecast. Split large variances into volume and price (or rate and usage for costs). - Agree a materiality threshold, for example any line more than 10% or £1,000 away from plan, and only investigate those. - Write one sentence of explanation and one action for each material variance. - Roll the forecast forward and update it with what you learned. ## Common mistakes and when to use what - Budgeting costs in detail but guessing revenue. Revenue drives almost everything else, so build it from customers, orders and prices. - Explaining variances without splitting them. "Sales were down" is not an explanation. "We took 100 fewer orders because paid traffic fell" is. - Updating the budget every month. Then nothing is ever off budget. Update the forecast instead. - Forecasting profit but not cash. A good profit forecast can still hide a cash shortfall, so link it to a cash forecast. - Making the process so heavy that nobody keeps it up. A simple forecast reviewed every month beats a detailed one reviewed twice a year. Use incremental budgeting when costs are stable and time is short. Use zero-based reviews when spending has crept up, when strategy changes, or when cash is tight. Use a rolling forecast in every case where the future is uncertain, which for most small businesses means always.