Among the numerous things I got wrong when I launched my business, my initial income prediction was at the top of the list. Not only was my forecast incorrect, it was incorrect by a large margin.
Instead of the $1 million in income that I had predicted for the first year, we ended up with about $100,000. My board of directors was unsurprised by my prediction miss, but also unsatisfied.
Most startups underperform their original projections, and I was no exception.
Barry, the chairman of the board of directors of my company, had a witty remark regarding startups and their forecasts. It goes as follows: “Most startups underperform their forecasts, and you’ll find that we’re no exception.”
Barry’s statement is amusing when it is not applicable to you. However, in my instance, it was my problem, and I needed to resolve it.
My board requested me, properly, to explain what was going on.
You must accept responsibility for forecasting errors, regardless of the cause.
Aren’t we entrepreneurs an optimistic bunch? This explains a portion of why I was off on my forecast, but not entirely.
Your initial income prediction is almost certainly a simulation.
When I founded my company, I had worked in the analog semiconductor sector for about two decades. I was familiar with how a typical product’s revenue rose over time, and so I utilized that as the basis for my model’s revenue growth.
Recognize that understanding the income stream is simply one component of your forecasting model. There are a few other components required. The first element you’ll need is an estimate of how long it will take to produce a product, feature, or revision to an existing product, as well as the number of engineers required per product.
I drew on my experience and that of Jeroen, my co-founder and VP Engineering, to deduce the second component of the model. We put a 50% success rate for “working first silicon” into the model, so we anticipated having to redo half of our devices.
The third component is your recruitment strategy for engineers and support workers. This would include test engineers, application engineers, and layout engineers in our scenario.
These three limitations define your model for the number of goods and features that can be introduced in a given time period.
Then you translate your model into an estimate of what will occur.
The following phase in our procedure (which you may be able to avoid, but we couldn’t) was to transition from generic to specialized products. We would employ the general model for unannounced future items.
This was the revenue prediction our board of directors received.
Then you take a deep breath and realize how inaccurate your model is.
It’s one thing to have real-world experience, as I did, and quite another to see what would transpire in your new reality, with a new team, and at a new firm. That is where the excitement begins.
After one year of income, we were far behind schedule, and our board wanted to know why. The exercise was quite beneficial for us to complete.
To begin, you can attribute the model’s mistakes to each of your initial assumptions.
As I already stated, this is where the adventure begins.
Are you familiar with the term “force majeure” in French? This term refers to “unforeseeable events that preclude a party from performing under a contract.” To put it mildly, we encountered some unexpected circumstances.
In our first year, we had two product families that would provide all of our revenue. We were really ahead of schedule with our initial product family, but it returned from our foundry, TSMC, with an inverted mask layer due to an error committed by TSMC.
The error delayed the launch by four months, resulting in a $300,000 revenue shortfall in the first year. The second product family was two months late, costing us approximately $200,000, meaning that $500,000 of our $900,000 shortfall was related to product delays.
The remaining $400,000 was a forecasted shortfall. Why did we fall short of the forecast?
You must adjust your forecast in light of reality.
You never know how the market will react to a product until it is actually sold. The good news is that clients who were aware of our products, and the crucial term here is “were aware,” truly enjoyed them.
More significantly, our items were being purchased by customers. Thus, what resulted in the omission?
I had forgotten that we were engaged in a two-front conflict, with the second front focused on generating interest in our products. I adjusted our prediction to match our current location.
Waterfall charts can be used to visualize your forecast’s changes.
If you’re unfamiliar with waterfall charts, they’re an excellent method to visualize the changes in your forecasting. If you’ve never seen a waterfall chart, here’s an example:
The one above is for administrative costs. The top line represents the initial budget. The real numbers are represented by the gold boxes that run diagonally. The rows represent the year’s monthly plan, while the columns represent the month for the plan.
The idea is that you and your board of directors can track how your plan has evolved over time for various aspects of the business. As you can see, we began the year with a budget of $3.233 million and are currently spending around $3 million.
After that, you rinse, repeat, and repeat.
Accurate forecasting is all about using what you’ve learned from past forecasts, adapting, learning, and repeating the process. That is all.
Then develop a practice of underpromising and exceeding sales projections. Take the forecast that you and your team create and slash it by 20%.
Assign this new number to your board of directors. This provides a buffer as you work to get your team back to the initial number.