Can you accurately forecast sales?

I have been working on two new projects over the last 7 days that have involved preparing forecasts around new business plans. Both of the forecasts are very different. One is a projection for a new business with no historical data to work from, and the other for an established business with historical accounts where sales data can be extracted from their Management Accounts.

It's generally pretty easy to put projections together until you reach sales, which are always difficult to forecast because you don't control your customers. You can't predict their buying behaviour but instead make educated best guesses and therefore there is always the temptation, because we're optimists, to over estimate our sales targets.

I remember a friend at school in my GCSE business studies class who realised his business plan for a tea shop in Cleethorpes didn't stack up and was forecasting a loss. His solution was to increase the sales figures until he realised the desired profit that would surely lead to an 'A' grade. The only problem was that the uplift in sales required a x.10 increase in cups of tea!

Unfortunately I've seen that approach taken in real life too. Sets of projections with sales targets that are simply unrealistic and better suited to fiction books than business plans. To ensure your projections are robust it's important to spend time trying to generate realistic sales forecasts.

If you have a sales pipeline or confirmed bookings these are an excellent starting point. If not you need to adopt a more logical approach. If you have historical data then you can review sales figures for the last 12 periods and set a target based on past performance. Using your knowledge of the industry you can then make upward or downward adjustments. Once you have identified your base annual target you can then fine tune this by looking at the profile of sales over each of the last 12 months against the last 12 months total sales. Re-applying this profile to your adjusted annual sales target then gives you a good base from which your initial forecasts can be based. Finally you can tweak the monthly figures for any specific promotions or plans you have or for any external events that might have either a positive or negative effect on sales, such as the Olympics or Diamond Jubilee extra bank holiday!

It's much more difficult to forecast sales when you don't have a history of sales to work from. As part of your business planning process you might have conducted some research with your target audience and have some market size statistics. The market research stage of your planning might provide an opportunity to generate some valuable information that could be used for sales forecasting.

Perhaps you could collect data on the size of the market and the percentage of target customers. Using your research you could then apply a percentage of target customers who said they would definitely buy the product or service and finally a percentage market share that you think you might win.

So for example your market could be 100,000 people. Within that market target customers could be 40%, a group of 40,000 people. Your research may have indicated that 20% of people will buy, 8,000 in this example and you feel you can win a 10% market share giving 800 customers.

You might then be able to use freely available data about which months people search for keywords associated with your products in search engines to profile these sales before applying your sales price.

Irrespective of which method you use to forecast sales in your projections it's important you re-visit them on a regular basis to compare with your actuals. By monitoring the sales figures and underlying drivers you can compare to the assumptions you made and if necessary adjust.