Category Archives: Sales Forecast

How to Forecast Sales

LivePlan Users Click HereYes, you can forecast your sales. Don’t think you need to have an MBA degree or be a CPA. Don’t think it’s about sophisticated financial models or spreadsheets. I was a vice president of a market research firm for several years, doing expensive forecasts, and I saw many times that there’s nothing better than the educated guess of somebody who knows the business well. All those sophisticated techniques depend on data from the past. And the past, by itself, isn’t the best predictor of the future. You are. So let’s look at how to forecast sales, step by step.

Graph Chart Diagram Drawing

Your sales forecast won’t accurately predict the future. We know that from the start. What you want is to understand the sales drivers and interdependencies, to connect the dots, so that as you review plan vs. actual results every month, you can easily make course corrections.

If you think sales forecasting is hard, try running a business without a forecast. That’s much harder.

Your sales forecast is also the backbone of your business plan. People measure a business and its growth by sales, and your sales forecast sets the standard for expenses, profits and growth. The sales forecast is almost always going to be the first set of numbers you’ll track for plan vs. actual use, even if you do no other numbers.

If nothing else, just forecast your sales, track plan vs. actual results, and make corrections; that’s already business planning.

Match Your Forecast to Your Accounting

It should be obvious: Make sure the way you organize the sales forecast in rows or items or groups matches the way your accounting (or bookkeeping) tracks them.

Match your chart of accounts, which is what accountants call your list of items that show up in your financial statements.

If the accounting divides sales into meals, drinks, and other, then the business plan should divide sales into meals, drinks, and other. So if your chart of accounts divides sales by product or service groups, keep those groups intact in your sales forecast. If bookkeeping tracks sales by product, don’t forecast your sales by channel instead.

If you’re planning for a startup business, coordinate the bookkeeping categories with the forecasting categories.

Get your last Income Statement (also called Profit & Loss) and keep it in view while you develop your future projections.

  • If you don’t have more than 20 or so each rows of sales, costs, and expenses, then make the rows in the projected statement match the rows in the accounting.
  • If your accounting summarizes categories for you – most systems do – consider using the summary categories in your business plan. Accounting needs detail, while planning needs a summary.

If your categories in the projections don’t match the accounting output, you’re not going to be able to track plan vs. actual well. It will take retyping and recalculating. And you’ll lose the most valuable business benefit of business planning: management, steering your company.

The math is simple

Normally your sales forecast will group sales into a few manageable rows of sales and show projected units, prices, and sales monthly for the next 12 months and annually for the second and third years in the future. Here’s a quick example from the bicycle retailer named Garrett I’ve used in other examples (with columns for April-November hidden on purpose, to make viewing easier):

sample sales forecast for bicycle retailer

Sample sales forecast for bicycle reteller

The math for a sales forecast is simple.

  • Multiply units times prices to calculate sales. For example, unit sales of 36 new bicycles in March multiplied by $500 average revenue per bicycle means an estimated $18,000 of sales for new bicycles for that month.
  • Total Unit Sales is the sum of the projected units for each of the five categories of sales.
  • Total Sales is the sum of the projected sales for each of the five categories of sales.
  • Calculate Year 1 totals from the 12 month columns. Units and sales are sums of the 12 columns, and price is the average, calculated by dividing sales by units.
  • The numbers for Year 2 and Year 3 are just single columns; unless you have a special case, projecting monthly results for two and three years hence is overkill. It’s a problem of diminishing returns; you don’t get enough value to justify the time it takes. Other experts will disagree, by the way; and there may be special cases in which extended monthly projections are worth the effort.

Estimate Direct Costs

A normal sales forecast includes units, price per unit, sales, direct cost per unit, and direct costs. Direct costs are also called COGS, cost of goods sold, and unit costs.

Most people learn COGS in Accounting 101. That stands for Cost of Goods Sold, and applies to businesses that sell goods. COGS for a manufacturer include raw materials and labor costs to manufacture or assemble finished goods. COGS for a bookstore include what the storeowner pays to buy books. COGS for Garrett are what he paid for the bicycles, accessories, and clothing he sold during the month. Direct costs are the same thing for a service business, the direct cost of delivering the service. So, for example, it’s the gasoline and maintenance costs of a taxi ride.

Direct costs are specific to the business. The direct costs of a bookstore are its COGS, what it pays to buy books from a distributor. The distributor’s direct costs are COGS, what it paid to get the books from the publishers. The direct costs of the book publisher include the cost of printing, binding, shipping, and author royalties. The direct costs of the author are very small, probably just printer paper and photocopying; unless the author is paying an editor, in which case what the editor was paid is part of the author’s direct costs.

The costs of manufacturing and assembly labor are always supposed to be included in COGS. And some professional service businesses will include the salaries of their professionals as direct costs. In that case, the accounting firm, law office, or consulting company records the salaries of some of their associates as direct costs.

The illustration below shows how Garrett uses estimated margins to project the direct costs for his bicycle store. For the highlighted estimates, the direct entry for bicycles unit cost is the product of multiplying the price by 68 percent. The total direct costs for bicycles in January are the result of multiplying 30 units by $340 per unit.

sales forecast estimating direct costs

estimating direct costs

 

Some Quick Notes About Standards

Timing Matters

Standard accounting and financial analysis have rules about sales and direct costs and timing. A sale is when the ownership of the goods changes hands, or the service is performed. That seems simple enough but what happens sometimes is people confuse promises with sales. In the bike store example, if a customer tells Garrett in May that he is definitely going to buy 5 bicycles in July, that transaction should not be part of sales for May. Garrett should put those 5 bicycles into his July forecast and then they will actually be recorded as sales in the bookkeeping actual sales in July when the transaction takes place. In a service business, when a client promises in November to start a monthly service in January, that is not a November sale.

Direct costs also happen when the goods change hands. Technically, according to accounting standards (called accrual accounting), when Garrett the bike storeowner buys a bicycle he wants to sell, the money he spent on it remains in inventory until he sells it. It goes from inventory to direct costs for the income statement in the month in which it was sold. If it is never sold, it never affects profit or loss, and remains an asset until some day when the accountants write off old never-sold obsolete inventory, at which time its lowered value becomes an expense. In that case it was never a direct cost.

What’s Accrual Accounting and Why You Care

Business accounting is either cash basis or accrual. I hate how attractive “cash basis” sounds, because accrual is way better, and easier to manage too. Cash basis accounting only works right if you absolutely always pay immediately for every business purchase, and you never buy something before you sell it, and all of your customers pay you in full whenever they buy something from you. So accrual is better.

Here’s why, in a few obvious examples.

  1. You make a sale when you deliver the goods. If the customer doesn’t pay you immediately, in cash basis nothing is recorded. The sale doesn’t even show up in your books until the customer pays. In accrual, you record the accrued amount as Accounts Receivable, so you keep track of the amount, the date, and the customer who owes it to you. It’s obvious that unless you never sell without immediate payment, accrual basis is better.
  2. You order some goods. When you receive them, you don’t pay for them. You owe the money. You have an invoice to pay. In cash basis, nothing happens until you pay up. In accrual basis, you record the accrued amount as Accounts Payable, along with the date, a record of what you bought, and who and when you are supposed to pay. So cash basis is better only if you pay everything immediately; all normal businesses need accrual.

I hate the fact that the accounting standards set a few generations ago chose to call it “cash basis” when you don’t record money owed into your books until it’s paid; or money you owe until you pay it. It’s a terrible idea to keep that information in your head instead of in your bookkeeping. That causes many mistakes as we business owners fail to keep track and remind ourselves of these outstanding obligations. And yet, ironically, they call that “cash basis” accounting. I do wish that the right way to do it, which is accrual accounting, didn’t have such an off-putting name.

Gross Margin

Once you have sales forecast and direct costs, you can calculate your estimated gross margin. Gross Margin is sales less direct costs. Gross Margin is a useful basis of comparison between different gross-marginindustries and between companies within the same industry. You can find guidelines and rules of thumb for different industries that give you an industry profile or average gross margin for different industries. For example, industry profiles will tell you that the average gross margin for retail sporting goods is 43%. Every business is different, but knowing the standards and averages gives you some useful comparisons.

The distinction isn’t always obvious. For example, manufacturing and assembly labor are supposed to be included in direct costs, but factory workers are paid sometimes when there is no job to work on. And some professional firms put lawyers’ accountants’ or consultants’ salaries into direct costs. These are judgment calls. When I was a young associate in a brand-name management consulting firm, I had to assign all of my 40 hour work week to specific consulting jobs for cost accounting.

Garrett can easily calculate the gross margin he’s projecting with his sales forecast. The illustration below shows his simple calculation of gross margin using his sales and direct costs.

calculating-gross-margin

Calculating Gross Margin

 

How do I know what numbers to use?

But how do you know what numbers to put into your sales forecast? The math may be simple, yes, but this is predicting the future; and humans don’t do that well. Don’t try to guess the future accurately for months in advance. Instead, aim for making clear assumptions and understanding what drives sales, such as web traffic and conversions, in one example, or the direct sales pipeline and leads, in another. And you review results every month, and revise your forecast. Your educated guesses become more accurate over time.

Use experience and past results

  • Experience in the field is a huge advantage. In the example above, Garrett the bike storeowner has ample experience with past sales. He doesn’t know accounting or technical forecasting, but he knows his bicycle store and the bicycle business. He’s aware of changes in the market, and his own store’s promotions, and other factors that business owners know. He’s comfortable making educated guesses. In another example that follows, the café startup entrepreneur makes guesses based on her experience as an employee.
  • Use past results as a guide. Use results from the recent past if your business has them. Start a forecast by putting last year’s numbers into next year’s forecast, and then focus on what might be different this year from next. Do you have new opportunities that will make sales grow? New marketing activities, promotions? Then increase the forecast. New competition, and new problems? Nobody wants to forecast decreasing sales, but if that’s likely, you need to deal with it by cutting costs or changing your focus.
  • Start with your best guess, and follow up. Update your forecast each month. Compare the actual results to the forecast. You will get better at forecasting. Your business will teach you.

How to Forecast a New Business or New Product

What? You say you can’t forecast because your business or product is new? Join the club. Lots of people start new businesses, or new groups or divisions or products or territories within existing businesses, and can’t turn to existing data to forecast the future.

Think of the weather experts doing a 10-day forecast. Of course they don’t know the future, but they have some relevant information and they have some experience in the field. They look at weather drivers such as high and low pressure areas, wind directions, cloud formations, storms gathering elsewhere. They consider past experience, so they know how these same factors have generally behaved in the past. And they make educated guesses. When they project a high of 85 and low of 55 tomorrow, those are educated guesses.

You do the same thing with your new business or new product forecast that the experts do with the weather. You can get what data is available on factors that drive your sales, equivalent to air pressure and wind speeds and cloud formations. For example:

  • To forecast sales for a new restaurant (there is a detailed example coming in the next section), first draw a map of tables and chairs and then estimate how many meals per mealtime at capacity, and in the beginning. It’s not a random number; it’s a matter of how many people come in. So a restaurant that seats 36 people at a time might assume it can sell a maximum of 50 lunches when it is absolutely jammed, with some people eating early and some late for their lunch hours. And maybe that’s just 20 lunches per day the first month, then 25 the second month, and so on. Apply some reasonable assumption to a month, and you have some idea.
  • To forecast sales for a new mobile app, you might get data from the Apple and Android mobile app stores about average downloads for different apps. And a good web search might reveal some anecdotal evidence, blog posts and news stories perhaps, about the ramp-up of existing apps that were successful. Get those numbers and think about how your case might be different. And maybe you drive downloads with a website, so you can predict traffic on your website from past experience and then assume a percentage of web visitors who will download the app (The following sections on Sample Sales Forecast for a Website and Sample Sales Forecast for Email Marketing offer more examples).

So you take the information related to what I’m calling sales drivers, and apply common sense to it, human judgment, and then make your educated guesses. As more information becomes available — like the first month’s sales, for example – you add that into the mix, and revise or not, depending on how well it matches your expectations. It’s not a one-time forecast that you have to live with as the months go by. It’s all part of the lean planning process.

Sales forecast depends on product/service and marketing

Never think of your sales forecast in a vacuum. It flows from the strategic action plans with their assumptions, milestones and metrics. Your marketing milestones affect your sales. Your business offering milestones affect your sales. When you change milestones — and you will, because all business plans change — you should change your sales forecast to match.

Sample Sales Forecast for a Restaurant

LivePlan Users Click HereMagda is developing a lean plan for a café she wants to open in an office park. She wants a small locale, just six tables of four. She wants to serve coffee and lunches. She hasn’t contracted the locale yet, but she has a good idea of where she wants to locate it and what size she wants, so she wants to estimate realistic sales. She assumes a certain size and location and develops a base forecast to get started.

Establishing a base case

She starts with understanding her capacity. She does some simple math. She estimates that with six tables of four people each, she can do only about 24 sit-down lunches in an average day, because lunch is just a single hour. And then she adds to-go lunches, which she estimates will be about double the table lunches, so 48 per day. She estimates lunch beverages as .9 beverages for every lunch at the tables, and only .5 beverages for every to-go lunch. Then she calculates the coffee capacity as a maximum of one customer every two minutes, or 30 customers per hour; and she estimates how she expects the flow during the morning hours, with a maximum 30 coffees during the 8-9 a.m. hour. She also estimates some coffees at lunch, based on 3 coffees for every 10 lunches. You can see the results here, as a quick worksheet for calculations.

Sales Forecast Assumptions

Where do those estimates come from? How does Magda know? Ideally, she knows because she has experience. She’s familiar with the café business as a former worker, owner, or close connection. Or perhaps she has a partner, spouse, friend, or even a consultant who can make educated guesses. And it helps to break the estimates down into smaller pieces, as you can see Magda has done here.

And, by the way, there is a lesson there about estimating and educated guesses: Magda calculates 97 coffees per day. That’s really 100. Always round your educated guesses. Exact numbers give a false sense of certainty.

Café monthly assumptions

She then estimates monthly capacity. You saw in Illustration 7-2 that she estimates 22 workdays per month, and multiplies coffees, lunches, and beverages, to generate the estimated unit numbers for a baseline sample month.

So that means the base case is about 1,500 lunches, about 1,000 beverages, and about 2,000 coffees in a month. Before she takes the next step, Magda adds up some numbers to see whether she should just abandon her idea. At $10 per lunch and $2 per coffee or beverage, that’s roughly $15,000 in lunches, $2,000 in lunch beverages, and $4,000 in coffees in a month. She probably calls that $20,000 as a rough estimate of a true full capacity. She could figure on a few thousand in rent, a few thousand in salaries, and then decide that she should continue planning, from the quick view, like it could be a viable business (And that, by the way, in a single paragraph, is a break-even analysis).

From base case to sales forecast

With those rough numbers established as capacity, and some logic for what drives sales, and how the new business might gear up, Magda then does a quick calculation of how she might realistically expect sales to go, compared to capacity, during her first year.

Estimating monthly sales

Month-by-month estimates for the first year

Month-by-month estimates for the first year

All of which brings us to a realistic sales forecast for Magda’s café in the office park (with some monthly columns removed for visibility’s sake).

Cafe Sales Forecast

Notice that Magda is being realistic. Although her capacity looks like about $20,000 of sales per month, she knows it will take a while to build the customer base and get the business up to that level. She starts out at only about half of what she calculated as full sales; and she gets closer to full sales towards the end of the first year, when her projected sales are more than $19,000.

Important: these are all just rough numbers, for general calculations. There is nothing exact about these estimates. Don’t be fooled by how exact they appear.

Notice how she’s working with educated guessing. She isn’t turning to some magic information source to find out what her sales will be. She doesn’t assume there is some magic “right answer.” She isn’t using quadratic equations and she doesn’t need an advanced degree in calculus. She does need to have some sense of what to realistically expect. Ideally she’s worked in a restaurant or knows somebody who has, so she has some reasonable information to draw on.

Estimating direct costs

We’ve seen direct costs already, in the previous section. They are also called COGS, or cost of goods sold, or unit costs. In Magda’s case, her direct costs or COGS are what she pays for the coffee beans, beverages, bread, meat, potatoes, and other ingredients in the food she serves.

Just as with the sales categories, forecast your direct costs in categories that match your chart of accounts.

So, with her unit sales estimates already there, Magda needs only add estimated direct costs per unit to finish the forecast. The math is as simple as it was for the sales, multiplying her estimated units times her perunit direct cost. Then it adds the rows and the columns appropriately.

cafe-projected-direct-costs

Here again you see the idea of educated guessing, estimates, and summary. Magda doesn’t break down all the possibilities for lunches into details, differentiating the steak sandwich from the veggie sandwich, and everything in between; that level of detail is unmanageable in a forecast. She estimates the overall average direct cost. Coffees cost an average of 40 cents per coffee, and lunches about $5.00. She estimates because she’s familiar with the business. And if she weren’t familiar with the business, she’d  find a partner who is, or do a lot more research.

email marketing sales assumptions

Sample Sales Forecast for Email Marketing

Sample Sales Forecast for an Email Marketing

The idea with forecasting something new is start with something that’s easy to guess, then go on from there. That’s what Magda does in the example above, going from restaurant layout with chairs and tables, to times of day, and days per week. This next example projects unit sales from email marketing. Here again, the key is to track the assumptions. So here’s a sample sales forecast for the projected unit sales of the first few months of a product to be marketed via email.

email marketing sales assumptions

email marketing sales assumptions

Warning: This is really simplified
May the email marketing experts forgive me for making it look this simple. It isn’t; but the basic numbers follow these basic principles.
  1. It starts of course with how many emails get sent. The assumption here is that the marketing department sends out 20,000 emails the first month, 25,000 the next month, and so forth. And let’s remember that while it’s easy to type numbers into a spreadsheet, execution requires an effective email message, design and formatting, and a good list of email addresses of real prospects. Targeting is essential.
  2. We put assumptions for how many people open the emails into the second row. And the assumption shown for January, by the way, is amazingly high, and quite unrealistic. A business would have to be sending emails to a list of opted-in email addresses for customers or prospects who like this sender a lot. Available information on average emails opened, from MailChimp and other vendors of email services, runs more like 15% to 25%. The numbers here are high.
  3. We use the third row for our assumption for how many people click the link on the email. There too, this example is very optimistic. Normal rates rarely get above 2%.
  4. Next is website views. With emails sent, emails opened as a percentage, and clicks as a percentage, we can project how many people click an email link and arrive at a website. In January, for example, we take 20000*.35*.08 = 560. Here again, the math is simple. The business behind it — a good email list, a good email, subject line, text, and links, and offering — is not simple.
  5. Then we project a conversion rate, which is how many people who see the offer on the web choose to buy. The 0.5% (one half of one percent) assumption here is not unusually low. Actual conversion rates depend on how well targeted the people are who arrive at the website, how attractive the offer is, and many other marketing and sales variables.
  6. Finally, in the last row, we arrive at projected sales. The indication here is that sending 20,000 emails produces the small unit sales shown here in the bottom row.

From here we would take the unit sales resulting from these assumptions to the main sales  forecast, with the structure we use for the sample sales forecast above: units, prices, sales, direct costs per unit, and direct costs. The spreadsheets would look a lot like the ones for Magda, above; and Garrett the bicycle retailer in How to Forecast Sales.

sample sales forecast for a website

Sample Sales Forecast for a Website

The next sample sales forecast shows assumptions for a web business. Here too, we look at the key assumptions that lead to the unit sales forecast. The point is we don’t pull a forecast out of thin air, we base it on some sales drivers that we can predict, and, to some extent control — or at least track and revise. We can look at these in detail below.

sample sales forecast for a website

Sample sales forecast assumptions for a website

First, estimate the drivers for web traffic

Clearly the web business sales assumptions depend on web traffic. In the first two rows of the forecast, we project reasonable numbers of web visits based on past web experience, search engine optimization (SEO), links that we can predict. In this case we break them into two categories:

  • First, website visits from organic search, based on the site, its contents, the SEO, and so forth. This projection may be optimistic because getting 200 people per month at the outset isn’t as easy as writing numbers into a spreadsheet. It takes marketing. Still, it’s an assumption we can track.
  • Second, website visits from social media. This assumes active engagement, posts, links, and updates on Facebook, Twitter, and other social media sites.

More about the pay-per-click assumptions

As you can see in the bottom two rows of the forecast, pay per click web traffic depends on two factors: how much you spend on pay-perclick advertising, and how much you pay for each click. A click in this case means somebody who was browsing on some other website, or who did a web search for some specific search word or phrase, clicked a link that went to your website. If you are not familiar with this kind of online marketing, there’s a good summary in Wikipedia under “pay per click”.

I base my assumptions here on bid-based pay-per-click systems, such as what Google uses. As you set up the campaign, you use a system where you bid for how much you’ll pay for each click you get from the paid area of search results when a web search requests a specific keyword. For example, the illustration here shows what happened when I searched for the term “restaurant in Eugene OR.” Two businesses have paid for the ad placement at the top. One is a restaurant supply business, the other a yellow-pages index. If I clicked on either one, I would go to that website and the business would be charged the pay per click amount. The rest of the search results are Google’s favorites, based on Google search algorithms, as the most useful.

Google search showing pay -per-click ads.

Google search showing pay -per-click ads.

Conversion Rate and Projected Sales

The row labeled “Website conversion rate” holds the very important assumption for the percentage of website visitors who choose to buy the product. That assumption is half a percent (0.5%) for the first month, increasing to six tenths of a percent (0.6%) in the second month. The total unit sales estimate in “Total unit sales” comes from multiplying the conversion rate in “Website conversion rate” by the estimated web traffic in the row labeled “Total website visits.” So, for example, the projected  13 units for January is one half of one percent of the estimated 2,550 web visits.