TLDR: For startups, 3 years is the most common. 5 years if you are in a later stage (Series B+) or if your business is more predictable, or if your startup potential can only be expressed in a longer time horizon. 10 years or more for edge cases like infrastructure plays, solar panels, pharma, and real estate. You are trading off showing more of your startup potential vs having credible forecasts.

If you are a founder, you’ll be all too familiar with the pain of making your financial projections for an investment round. It starts with the modelling, and the time it takes to get the simple-in-your-head connections translated to the devil machine that is Excel. It continues with assumption setting. How can you forecast the square meters of office your company is going to need?

And it ends up with third party considerations. Are these numbers believable? Have I been too conservative? How will investors react?

When fundraising, projections are a necessary evil. However, when done right, they can be produced efficiently and have valuable informative outcomes, which make it worth the effort.

We are not going to cover everything about projections in this article, if you are new to the topic, check out:

This article is going to focus on a technical aspect of financial projections, how far in the future should you forecast?

First principles

For valuation, but also for communicating your startup’s potential, the longer the forecast, the better. The largest companies took some years to develop into the world leading giants that they are today, and you likely want to show that your company is going to be among them in the next 10 years. You should then forecast for a time horizon long enough to allow the company to become a world giant, correct?

Not quite. The trade-off that you are making is with credibility. Yes we can forecast that the ice cream shop that still needs a logo will become the next largest franchise business in the world, but we are including so much risk in that forecast that it almost becomes a worthless exercise.

How long should we forecast then? As long as we can, up until that trade-off becomes negative.
There is no mathematical formula to determine when this happens, but my rule of thumb when adding an additional year of projection is:

  1. Am I adding information to the future potential
  2. Is the information I’m adding credible enough not to be offset by the increased risk?

Let’s unpack those.

Am I adding information to the future potential

Adding an additional year is not really useful if it doesn’t add additional information. This means that if the added year is just another extension of your model, it’s business as usual but with a greater level of uncertainty. Iit doesn’t add any information to the reader, and could thus be skipped.

If the extra year means your projections now reflect a key milestone, such as planned international expansion or a key product release, then there’s a very clear argument to include it in your projections.

Is the information I’m adding credible enough not to be offset by the increased risk?

If there is information in that added year (e.g. the aforementioned expansion or product release), is that information believable enough to be worth inserting?

For example, I speak with a lot of entrepreneurs that will launch additional products in the future. These additional products are already conditional on the current products working.

Let’s say that you are a furniture studio. You’re including in your projections the current tables and chair lines that you are selling. Your customers are all asking you to add drawers, so you’ve designed drawers and you are raising money to produce the first batch. Some customers asked about stools, you haven’t designed those yet, and the demand seems to come from edge cases, but you’ll start making them in two years, so you include them in your projections. Then you are excited to add lamps, based on a comment someone made to you. Lamps could be a new segment in your business, but given the cost, you can’t start it before three years; you have no idea if anybody would want your lamps, or how to make them and design them.

Should lamps be in your projections? Should stools? The case for inclusion gets slimmer and slimmer.

A good test for including these would be: would they be stronger as a possibility?

Let me explain, with an example of how you might frame that:

  • And in year 3 we are adding lamps, their sales and margins will be very large as you can see here. We still don’t know how to make them or who will buy them but we are confident in a big market with large growth.
  • And think about this, the stools, drawers, tables and chairs are not only products, they are creating a brand. If they go as planned, we haven’t even touched the profits and growth we could have by expanding to the nearby market of lamps.

Which one of the two is stronger? Which one is more likely to increase valuation?

This is just an example, but adapt it to your own circumstances and it can be very useful. Some examples of trade-off inclusions that you could be facing:

  • International expansion
  • New product lines
  • Integrating with your supplier/customers
  • Additional services
  • Additional revenue streams from the same customers/products
  • Enough future to show initial proof of business model

Then if the startup has only one product and it’s very simple, I can just forecast for 1 year, correct?

Well, you could, but you would not allow your business model to express itself. Your projections should allow a reader to have an understanding of what initial success means for you and for the business.

Can the company be profitable with the current idea when it gets a little bigger? how much would be the free cash flow that the company will either be able to distribute or to use for growth?

In practice

All this theory generally converges into some tried and tested startup practices. As always, these are broad averages and that is why it’s important to understand where they come from.

Most startups converge on 3 years

Startups raising up to and including Series A that don’t belong in the next niche category are generally forecasting for 3 years. 3 years should be enough for an agile software startup to show it’s initial cash flow traction and if they have the correct model to make an initial profit and cash flow.

Some startups plan for 5 years

In some cases, startups should extend their forecasts beyond 3 years. This should be done carefully as the uncertainty added in year 4 and 5 is substantial. However this could be justified in two cases:

  • Startups with predictable business models
    These startups have business models that are more “predictable than the average startup”. This is a big value judgment but let’s say that the company has been going for 3-4 years and have a stable customer base and thus very predictable revenue. Or if their customers are large institutions that signed already for a longer period of time. These and other rare cases should fall in this category. The next one is probably more common though.
  • Startups that will take more than 3 years to express their full business cycle
    There are many startups for which 3 years might not be enough to express a full business cycle or to display at least the initial potential of their business idea. These could be deep tech startups, biotech, pharma, real estate, banking (because it requires authorizations). It could be that the big assumptions of these businesses can’t be achieved within 3 years. Thus forecasting for 5 gives a better picture of the potential of the company.

Also in this case, the best practice would probably be to try to squeeze everything into 3 years and if it is really difficult, expand the forecast to 5 years.

Series B+ startups mostly forecast 5 years

When your growth engine becomes predictable, 5 years forecasts become more and more the norm. From Series B onwards, especially of startups that will roughly expand on the same business model and product, 5 years would allow a more refined valuation that will contain more of the specific future of the startup. When we cut these projections to three years, we are assuming that afterwards, the startup will grow as an average company. In the case of these companies, the scale seems to be tipped towards the 5 years forecast.

Special cases can forecast for 10 years or more

These should really be edge cases, with business models that are peculiar and fixed. If you belong to one of these industries, you should already know it.

The specific businesses here are, for example, energy and infrastructure plays (solar panels, bridges, oil rigs) and patented drugs (that have a patent cycle of 20 years and thus generally forecast for as long). Real estate could be included here if the average hold on a property is 7-8y plus. If the turnaround is faster, shorter forecasts will still allow the business model to be expressed.

In conclusion

As with most things in valuation, the length of the forecast is a balancing act. We need to trade off real-life uncertainty against modelling uncertainty. As always every case should be considered unique and requires a different application of the principles outlined in this article.