Bottom-Up Forecasts vs. Top-Down Forecasts

In a startup, a valuable concept is the understanding of bottom-up forecasts vs. top-down forecasts and when to use either method of forecasting.

Top-down forecasts

Here is an example for a top-down forecast, borrowed from Guy Kawasaki:

• We want to sell internet access in China.
• There are 1.3 billion people.
• 1 percent want internet access.
• We'll get 10 percent of that potential audience.
• Each account will yield $240 per year.
• 1.3 billion people × 1% addressable market × 10% success rate ×$240/customer = $312 million.

That might work for a huge established company, trying to forecast its estimated market share when entering a new market or segmenting an existing market—looking back at an extensive history of success, an existing sales apparatus, and reliable data points to justify the estimates ("Sure, look at how conservative they are!") above.

Such a forecast will not work for a startup. In fact, you can always pick a big enough market to fool yourself that success won't be that hard to achieve.

Bottom-up forecasts

A startup has neither the resources nor the extensive operating history a huge established company has. A startup is an experiment. It needs to bootstrap and build bottom-up models to arrive at realistic assumptions for market sizes.

Here is an example for a bottom-up forecast, again from Guy Kawasaki:

• Each salesperson can make ten phone sales calls a day that get through to a prospect.
• There are 240 working days per year.
• Five percent of the sales calls will convert within six months.
• Each successful sale will bring $240 worth of business.
• We can bring on board five salespeople.
• Ten calls/day × 240 days/year × 5% success rate × $240/sale × 5 salespeople = $144,000 in sales in the first year.

Such a bottom-up forecast is not only much more realistic in the case of a startup. It has the additional benefit of being testable. The above assumed success rate can be proven wrong immediately while the validity of the 10% market share assumption in the top-down forecast example can at most be assessed asymptotically, after years or even decades of operation.

Conclusion

Use top-down analysis to estimate the market opportunity for your company—to decide whether you want to try it in the first place. Use bottom-up analysis to get a realistic estimate of what you can achieve within the first years of your startup.