Anyone who has ever written a business plan for a new venture knows the feeling: you’ve just done a great job of articulating the underlying proposition, how it works and why the customer will be interested in it, but now you need to project forward and show why it will be a worthwhile venture in financial terms. It would be nice if you could just rely on investors to have some imagination and go with the “yeah, feels like it could be big” sort of estimate, but it usually doesn’t work out that way.
We need some numbers for our forward financial projections. Investors say they are happy with estimates and maybe all the more so because we might be talking about promoting a product that is not yet built to a market that does not yet exist. The problem is that numbers feel precise and no matter how often we insert the words “estimate” and “rough” and “guess”, we know that at least one of the investors is likely to come along in a couple of years and say “whoa – I thought you said you’d hit $100m revenue in year 3 without further funding!”
To make this problem worse, if we take the obvious, linear approach, we might try to build a bottom up model showing how we expect the business to grow step-by-step. This makes it feel even more precise and opens up many, many more questions (How do you know what the sales cycle is going to be? How can you even guess pricing at this stage?).
I’ve had this particular discussion three times already this week so I thought it would be worth writing up some rough notes here to explain how I like to approach this issue. I’m not claiming this is any kind of unique rocket science, just a few thoughts that might provide a useful perspective.
In these situations, it is often useful to approach this from a different angle and work backwards.
Our objective is to create a financially interesting business, so the starting point is to define what an interesting business would look like. Let’s say:
- We are an enterprise-focused SaaS business and we have a few trial customers signed up and generating $5k per month so our current revenue run rate is $60k. Let’s say last twelve months’ revenue is $50k. Small stuff, but in the early days it is all significant (unless you are sub-letting desks).
- We expect the next 12 months to be a major step change as we begin selling in earnest while we also devote effort to marketing and account management with the goal of expanding the initial accounts through adoption of additional functionality and wider usage across the organisation.
- In the coming 12 months, Y1, we expect to hit $500k of revenue and be on $80k per month recurring revenue by the end of Y1, but more importantly, we expect to have built a pipeline of customers that we will then farm to deliver results in Y2 onwards. We project very aggressive, but achievable growth as shown in the table below:
…this looks interesting. Forget about cash requirement for the moment. If you can get this business to $7m revenue in Y3 and be on $10m revenue run rate at the end of that period, the business is quite interesting.
But is it believable? How easy is it to believe in these numbers and in the underlying assumptions?
We can approach believability in from two directions: top-down and bottom-up. The bottom-up view involves considering the mechanics of acquiring customers and delivering the solutions. The top-down view involves asking if there are enough customers in the world for us to achieve these numbers with reasonable assumptions around pricing and market share.
To continue with our example here, let’s make the bottom-up argument as to why these numbers are believable. Using the same example and knowing the detail I’d be able to argue the believability as follows:
- We’ve got 20 customers on a $3k per annum trial level, so a lot of work to do to hit that Y1 target, but the majority of clients have come via referrals and we feel that it would be easy to ramp this up.
- Evidence suggests we can up that initial charge to $5k per annum and we expect to reach 100 clients at that level during Y1. Against this target, we would are starting Y1 with a pipeline of over 300 clients of which 100 have already expressed a strong interest. Feels do-able.
- We expect 20 of the clients on our starter level to upgrade to the multi-user level with full flexible use and adaptation of the core module. This will also include some professional service fees which we will initially do in-house, but increasingly look to hand over to partners. These 20 accounts will generate annual revenue of $25k each. We can point to nearly 10 discussions already heading towards this level of engagement.
- We expect 10 of the clients on the starter level to expand usage and adopt a second module focused on interaction with suppliers. These 10 accounts will generate annual revenue of $50k each. We have 5 clients already doing early trials with this second module and although it needs some further development work, it is already functional.
- We expect 5 enterprise-wide licenses with unrestricted usage of both core modules and these will generate $100k per annum. Again, we have expressions of interest and intent and one client in a large organisation working towards this.
- Finally, we expect to sign 2 clients onto the enterprise + analytics solution where the analytics capability comes from a partner, but is tuned so that the results feed into our system. For these accounts we’d be leveraging the commercial strength of a proven partner in terms of both sales and delivery. Both of these would generate $250k per year.
If we work through these numbers, all of which we can track back to our current pipeline, we get to a monthly recurring revenue of nearly $200k, equivalent to annual revenue of more than $2.5m from license fees. Sure there is a huge amount of execution involved in actually getting it done, but before we start the clock on Y1 I can point to the pipeline that should get me beyond the Y2 revenue number. Those numbers are feeling believable, particularly if we factor in the virality that we are seeing drive in-bound interest.
We can go further with this and unpack the detail on what would be required to execute and then look at how believable those elements are – can we actually hire (and pay for) the required number of sales and delivery people?
Now let’s look top-down and think about the believability of the Y5 numbers. $18m revenue? We’d need a total of around 800 customers with the same distribution across the distribution of price points. This would give us MRR of £1.5m. We are starting with a sharp vertical and geographical focus for our first 12-18 months. We already have interest in other verticals and geographies, but for the sake of argument, we work out that there are 5,000 prime target customers for our solution within our constrained focus. We’d need 20% of those as customers. That is relatively high and would be more believable if we only needed 2%. If we consider that in Y2-3 we will be active in a second, larger market, we see the number of target customers swell to nearly 20,000, so we need 4% – believable from a top-down perspective if we can indeed target that second market.
If we go a few thousand feet higher, we can ask the $100m question: what would I have to believe to accept that this business could be generating $100m of revenue? We’d need more than four thousand customers with the revenue profile outlined above. That is still only around 20% market share in our target vertical. We could argue the case for establishing a winner-take-call position in our target vertical and aim to capture more than 20%, or we could develop a presence in 2-3 parallel vertical sectors and do well based on having just a modest share of the market. The $100m revenue level feels believable.
At this stage we have a set of projections that we agree are interesting and we’ve looked at the believability of those projections in terms of our existing engagement with the market and from the perspective of the overall market size.
Of course we can debate endlessly about *how* believable the various numbers are, but that is exactly what we’d like to be discussing. Instead of saying that we will have $19,347.23 of revenue in Y5 because we are hiring X sales people and each sales person will sign up Y clients per month and each client will pay on average $Z, we are saying that our existing data points and our knowledge of the overall market suggest that revenue of $18m – $19m should be achievable within the 5 yr timeframe. We’ve also created a framework to look at other scenarios: how believable would a $10m rev projection in Y2 be?
The example used here is an enterprise SaaS business, but the application elsewhere is obvious. For example, maybe you have a mobile gaming company with $5m of revenue. Interesting = $20m of annual revenue within the next 3 years. To get your $5m revenue you typically have 5m monthly active users of which 1.25% are spenders with an ARPPU of just under $7 per month. How difficult is it to believe that you can grow ARPPU by maybe 10%, grow the % of spenders to 1.5% and get MAUs up to 15m? Beyond those numbers, what would be the expected churn rate and therefore how many users would you need to be acquiring, and at what cost in order to reach and maintain that 15m number? How do those cost of acquisition and retention numbers compare with industry peers?
At an early stage you might only need to explore the believability of the revenue line, but it is useful to go beyond this and give some thought to how the revenue flows through your system and ultimately (hopefully) gets converted to cash. Look at a set of comparable companies, pick 3-5 companies that are public and reasonably close to what you are doing, identify the key metrics of the business and extract an average figure for each of those key metrics. At the simplest level, maybe we find that the average gross margin for those businesses is 87% and the average EBITDA margin is 35%. Using just those numbers you can build a believable budget that is driven by your revenue line and have a view on how much you should be thinking of budgeting for the various line items in your operational plan.
Each type of business will have some unique metrics that will be worth considering. A few examples:
- Revenue per employee: In the SaaS example, looking at a broad set of companies we’d find a median figure of around $185k-$190k of annual revenue per full time employee. Looking back at our projections above, $18m revenue in Y5 would give us about 95 staff.
- Cost of Acquiring a Customer (CAC) and time to recover CAC: maybe we set a target to recoup our CAC in the first 12 months with the goal of moving that down to 7-8 months over time. Looking at our example here, the basic SaaS customers are paying $5k per annum and so we can consider spending $5k to acquire each new customer.
- Lifetime Value (LTV) and multiple of CAC: maybe in our SaaS example, we find that healthy companies look like they are running at LTV of 3x CAC so we can set that as our target and aim to shift it up to 5x over time. For our basic SaaS customers at $5k per annum, we want to generate $15k of value over the customer lifetime and this means we need to focus on Y2 and 3 renewals as well as account development and up-selling. This shouldn’t be too bad because, of course, our churn rate thus far is 0 and we have no reason to doubt that it will change in the future!
- Valuation: Enterprise Value as a function of the last twelve months’ revenue and EBITDA. This will give you a sense of how the value of your company will be estimated externally as you develop, but the valuable insights often come when you look at the spread of results and look at what underpins the difference between achieving 3x EBITDA value vs 8x EBITDA value. Often it is these insights at an early stage of a company’s development that can help to design a valuable business.
I could go on, but hopefully this very general framework is clear. Beyond being a handy approach to crafting projections into the unknown, I’d argue that being able to think through propositions in this way is critical for any entrepreneur wanting to establish themselves as a credible CEO.