Sales forecasting is much easier than you think, and much more useful than you imagine. It’s not about guessing the future correctly. We’re human; we don’t do that well. Instead, it’s about assumptions, expectations, drivers, tracking, and management. You review and revise your forecast regularly. Since sales are intimate with costs and expenses, the forecast helps you budget and manage. You measure the value of a sales forecast like you do anything in business, by its measurable business results.
Don’t ever think you can’t do it yourself. I speak from experience, as holder of a fancy MBA degree and once-upon-a-time Vice President of a market research firm charging thousands of dollars for sales forecasts. If you know your business, then your educated guess is better than any sophisticated forecasting model. And if you don’t know your business, no offense intended, but get a clue. Partner up with someone who does. Stop reading this and deal with that problem first.
Step 1: Find the right level of detail
Don’t be misled by accounting reports that give you infinite detail of sales broken down product by product, or customer by customer; for looking at the future, you need a manageable level of detail. Forecasting, although it often results in tables that look like accounting reports, doesn’t work in too much detail.
Lines of sales
In my experience, breaking your forecast into two, five, or maybe 10 lines of sales is common and works pretty well; but 20, 50, or 100 is way too much. For example, a restaurant ought not to forecasts sales for each item on the menu; instead, lunches, lunch beverages, dinners, and dinner beverages. That’s about the right level of detail. It’s way easier to forecast total lunches than ham sandwiches vs. bowls of soup. We’re human. That’s the way we think.
A bicycle store ought not forecast sales by item, but rather by type, like you see in this illustration. New bicycles, accessories and parts, clothing, repair and service, for example, are the standard sales items of the bicycle shop included in our sample plans at bplans.com.
If you’re an existing business, you should be able to get the information you need from your accounting. Most accounting systems used in small business have categories in the chart of accounts and reports that summarize by line of sales. It’s good to think in terms of the accounting system from the beginning because the goal is management, not brilliance or clairvoyance, and management means tracking, review, and revision.
And for startups, you have the advantage of thinking about lines of sales and forecasts at the beginning, as you set up your accounting.
Although you don’t want the sales line to match the detail of your accounting item by item, you do want the lines to match information you get from your accounting. If your accounting tracks sales by channel but not by product type, for example, then try to forecast sales by channel. And if your accounting doesn’t give you information you can summarize and aggregate into meaningful categories, then maybe it’s time to review your chart of accounts. Sales forecasting as management requires an easy way to track results and see how they match the plan.
Months and years
Whether or not you use a spreadsheet for your forecast, it’s useful to take a minute to think of your sales forecast as a spreadsheet with rows and columns. The rows are line items or groups of items, with a label in the leftmost column. The columns are months, quarters, or years, and which you choose depends on your business needs. I always recommend you keep a sales forecast as a rolling 12 months ahead, and years beyond that. I discourage having monthly columns beyond the next 12 months, because more detail gives you diminishing returns on the effort. I don’t think guessing that far ahead in that much detail works. However, you know your business better than I do, and there are some contexts in which 24, 36, or even 60 months are appropriate. If your boss, your bank, or your investors require that much detail that far ahead, I don’t envy you, but you will have to serve the business need.
Step 2: Break it into meaningful pieces
Think of this as divide and conquer at a conceptual level. There are so many advantages to breaking the sales forecast into meaningful pieces. For example, it’s way easier from the beginning to think of total sales as “units x prices” than just as total sales.
Don’t assume you don’t have units just because you sell a service. I made a living with management consulting for more than 10 years, and my units were days. I projected consulting days per month, and a price per day. That worked. Accountants and lawyers have either hours or quarter hours.
What you miss, if you don’t break sales into units and prices, is the benefit of the plan vs. actual analysis later. Of course your forecast will be wrong—they all are—but think how valuable knowing whether the difference was in how many units you sold or in the average price per unit. That tells you want to change.
And you can use the same logic for direct costs too. Just like you project an average revenue per unit, so too you can estimate an average cost per unit, which gives you a simple way to forecast direct costs in your sales forecast (which is the best place to put them).
Whether you use a spreadsheet or not, think of your sales forecast in five units, each of which is 12 or more months and three or more years. First comes units, then revenue per unit, then sales (units x price), then unit direct costs, then direct costs (units x per-unit direct cost). And that’s a sales forecast you can use.
Step 3: Educated guesses
If you have past data already:
Those of you looking to forecast sales for existing businesses, which have past sales data available—you’re a fortunate special case. You have the luxury of taking data from the recent past, projecting it forward, and then fine tuning to deal with what you expect will change for the near future. Don’t build a sales forecast from scratch.
Do, however, reflect on the previous point about summary and aggregation. If your past sales data isn’t grouped in a way that simplifies doing a sales forecast, then consider revising your chart of accounts and grouping your sales in ways that make forecasting and tracking easier.
And if you don’t:
So you’re working on a startup? A new business, for which you have no past data? Don’t fall into the trap of thinking not having past data justifies not forecasting sales. Every startup began without past data, but you can still estimate sales drivers using reasonable assumptions.
Don’t suggest you can’t forecast because it’s so new. That doesn’t fly. Focus on what will drive sales and especially because the sales forecast is a management tool, break it into the factors you can identify that drive sales up or down.
Units and prices per unit are obvious and important. Don’t think you can’t forecast by unit just because you don’t charge by units. When I was a consulting I was billing by the job, not by hour or day. But I planned using the consulting day as my unit, and an average revenue per day as my price; having that breakdown helped me see, month by month, when my billing days or average billing amount per day differed from my expectations. A lot of professional services, such as attorneys and accountants, bill by units of time.
Most businesses have other drivers to work with. Our software business, for example, depends on web traffic, conversions, new subscriptions, and renewals. In my pure consultant years, my business depended on repeat business and it was relatively easy to categorize consulting jobs as either repeat business or new business. I imagine a restaurant business depends on meals served, and that a retail store depends on traffic in the store. Identify your drivers. If you can, build them into your rows.
A Sales Forecast Example
With apologies for miniaturization, the illustration below shows the full sales forecast for a bicycle shop, for 12 months and three years, units, average revenue per unit, sales, average cost per unit, and direct costs (and you can click on the image to see it a larger version).