Got you excited with that title, didn't I.
If you're not a "numbers person" and the thought of an excel spreadsheet makes you queasy, the last thing you’ll want to do when starting up a business is to create a financial model. My goal, by the end of this blog post, is to convince you to put together a basic financial model and equip you with three key concepts to help you do so.
While you may have decided how much you want to charge for your product, this by itself is not a financial model. You might have an idea of how many customers you think you can acquire in your first twelve months, and while this will indicate the revenue potential of your business it is again, not a financial model.
Why do you need a financial model?
Having a decent grasp of the dynamics affecting how your business makes money is crucial when starting a business, especially once you start investing your (or someone else's) money and time into it. For a business with a recurring revenue (predictable repeated revenue over a length of time), such as a SaaS business, building a basic financial model early on is super useful for many reasons including the following:
- It forces you to define and question your assumptions around growth, retention and churn
- It gives you an idea of where you might be in future years, based on some educated guesses. It’s as close to a crystal ball as you’ll get
- It gives you a useful tool to help inform decisions today (e.g. What happens to my end of year revenue if we charge £10 more per month? How much more cash will we need by the end of year-one if we decide to give our first 100 customers free access for a year?)
- If you're having discussions around shares with your team, it gives you a meaningful basis to project what your share percentage might be worth in 3-5 years' time
- It gives you something more robust to talk about with potential investors than a dream, and might just help get your foot in the door
- It gives you an idea of where you might be in future years, based on some educated guesses. It’s as close to a crystal ball as you’ll get
To avoid confusion, this is what I mean by growth, retention and churn:
- Growth: the rate at which you get new customers
- Retention: the rate at which your customers repeat their purchase with you
- Churn: the rate at which your customers leave you
The key word here is ‘basic’. You don’t want a complex model that you don’t understand and won’t be able to explain to your team and invite them to use.
Throughout this post, I'll be talking about a fictional SaaS company called Bants. Bants offers a chatbot service on your website on a subscription basis. Use of Bants costs £25 per company per month.
In "napkin maths" terms, one could say, if Bants aims to get 10 customers per month, to have 120 customers by the end of the year, paying £25 a month, they could expect to have £36k by the end of the year (25 x 120 x 12).
The £36k figure is close to what potential investors would be interested in as it indicates revenue potential of the business. However, Bants' actual first year total revenue is likely to be closer to £12k. Read on to understand why.
Below are three key concepts and misunderstandings that often come up when I speak with startups about modelling recurring revenue. They might help you build and test your own company's future potential.
1. Assuming day-1 sales (or, silly optimism)
Most businesses have sales cycles. A sales cycle, simply put, is the process that begins with the customer first becoming aware of your product and ends with the customer purchasing your product (hopefully).
Sales cycles vary by type of business and product. They can last anything from one month to a year. It matters less how you define your sales cycle. More important is having a common shared definition of your sales cycle by everyone in your business. And when it comes to modelling financials, a mistake I see made when projecting revenue is to count the revenue on day-1 of the sales cycle rather than at the end.
For most products today, there are two common routes to market, digital and field sales, each with its own typical sales cycle.
Digital sales cycles can be quick. Provided you already have a website and/or a landing page for your product, you need to put some ad creatives together, set up some digital ads on a channel such as Google or Facebook and then wait for people to sign up. Your ads and copywriting might need some tweaking before you get the right people signing up. This process of design, set up and optimisation might take a few weeks to a few months. During this time, you will be incurring costs without making a penny. Your model should account for this.
For field sales, you need sales people. Those sales people need to find potential customers (leads), reach out to them, get the conversation going and convince them to try your product.
Usually, companies adapt a mix of both. They may use digital means to get qualified leads and have a sales person convert the lead to a trial or a sale. Their sales cycle might look something like this:
Bants plans to run social media marketing directing potential customers to their website where they can register their email address for a demo. A sales person will then give them a call, explain the product and get them to sign up as a paying customer. (Note: this is not the most effective sales strategy; however, this post is not about sales strategies.) Bants expects this sales cycle to take one month on average.
Your model should allow a runway of sorts, for the average duration of the sales cycle, often 1 - 3 months for simple B2B products. If your model starts on month-1, your ad spend and sales wages would start leaving as costs from month-1 but your first sales revenue should not arrive until a later time.
In practical terms, you may choose to spread your total revenue across the cycle duration. For example, a company that has a sales cycle of 2-3 months may decide to model their revenue projections by counting 20% in month-1, 20% in month-2 and 60% in month-3. This would suggest that, if they expected to sell to ten customers per month, two would sign up within one month, two would sign up in the second month and six would sign up in the third month following their first interaction with the company. Or, to keep things simple, you could delay all revenue for three months and count 100% in month-3.
2. Recognising all your revenue up front (or, counting all your eggs before they’ve hatched)
A second mistake I come across is when companies don't spread their revenue projections consistently across the customer lifetime, and then making further projections based on a shaky base. This does not apply to businesses that only have a monthly payment option. If you only accept monthly payments and are not considering quarterly or annual up-front payments, you can skip ahead to point 3.
There are two financial concepts that are important to understand here, MRR (Monthly Recurring Revenue) and ARR (Annual Recurring Revenue). Definitions, courtesy of the Corporate Finance Institute, are as follows:
MRR: Monthly recurring revenue (MRR) is the "revenue that a company expects to receive on a monthly basis from customers for providing them with products or services." Or, the company’s normalised monthly revenue.
ARR: "Annual recurring revenue (ARR) is normalised on an annual basis revenue that a company expects to receive from its customers for providing them with products or services."
"The only difference between the two metrics is the period of time at which they are normalised (year vs. month). Thus, ARR provides a long-term view of a company’s progress, while MRR is suitable for identifying its short-term evolvement."
The term ‘normalised’ can be taken to mean ‘in a normal situation without one-off variances’. Generally, ARR is calculated by multiplying the MRR by 12. The ARR will grow as the MRR grows, as you gain more paying customers.
Back to our fictional company Bants; use of Bants costs £25 per month and you can get a 20% discount if you pay for a year up front - so £240 ((25 x 12) x 0.8) for the year. In a high-growth scenario, Bants expects to acquire around 40 new users per month and have roughly 50% of them choose the annual payment option. Bants does not offer a free trial period. All customers are paying customers.
The cash from these new users in a month that gets paid into Bants' bank account would be £5,300. This is broken down as follows:
Note: To keep things simple, we’re only focussing on revenue from new customers in the month. In reality, your total bank balance for the month will include monthly payments from customers on-boarded in previous months, your new customer payments (mix of annual and monthly) and any annual renewals from customers who have already been with you a year and are happy to renew for another.
Let's imagine Bants expects this rate of growth to continue, with an additional £5k or so landing in their account every month. So, when asked about ARR, the CEO recalls £5k as the monthly growth figure, multiplies it by 12 and arrives at an annual revenue estimation of £60k (5 x 12). However, as explained previously, the £5k per month is the sum of 50% of customers paying monthly (genuine MRR) and 50% of customers paying for the whole year up-front (not MRR).
It is important to identify the correct amount that should be recognised as ‘normalised monthly revenue’ without including what can be considered ‘pre-payments’ to arrive at your MRR. Then, you can multiply this by 12 to arrive at your ARR.
The illustration below would be a more sensible approach to Bants' MRR and ARR calculation in this high-growth scenario.
Correctly breaking down the monthly and annual payments results in an ARR of £11k, not £60k.
Incorrect modelling may result in wildly inflated future growth projections. Understanding the difference between MRR and ARR, as well as having a good idea of your customers’ subscription preferences will help you avoid making mistakes when forecasting your revenue growth.
3. Incorrectly modelling churn (or, more of that eternal optimism)
The final common mistake made is with churn. Or, the rate at which your customers leave you.
Churn can be tricky to model and the biggest obstacle in my experience is understanding the difference between monthly and annual churn, and deciding which to use for your model. The correct churn rate to use will depend on the way your customers usually pay (monthly, annually or 3-yearly sometimes!) and the normal or expected churn behaviour of your customers.
It’s a major downer to think about your customers leaving you when you’re starting out, so you may leave the science of churn on the back burner (aka bury your head in the sand). However, without including any churn, your model will be bloated and overly optimistic because, well, that’s not how humans work.
First off, whether you have a minimum contract period or notice period will affect how your churn is modelled. Many modern SaaS businesses don’t have a minimum period (customers can leave when they want) and this is very much the model I would advocate but that’s a discussion for another day.
If you allow customers to cancel at any time, before their next billing date, like Bants, you will want to use a monthly churn figure. Bants expects 1 out of 10 customers to give their notice within their first paying month. Their estimated monthly churn rate is 10% (1 ÷ 10).
Fast-forward 12 months; assuming the churn rate stays roughly the same for 12 months; this is what customer numbers might look like for Bants at the end of the year (allowing for the 1-month sales cycle mentioned in point 1):
So, a monthly churn rate of 10% ends up being an annual churn rate of 45%.
Alternatively, if you have a minimum contract term of one year, then you may want to use an annual churn rate from the start. Say, you expect 4 out of 10 customers to stop using your product after their first year, your annual churn would be 40% (4 ÷ 10).
Going back to Bants' original ambitions, in order to have 120 active paying customers by the end of year-1, they have to onboard over 200 customers and make sure their monthly churn does not exceed 10%.
In reality, your churn rate won't be consistent. You would expect the number of new customers to grow as the business gets better at selling. You would also expect the churn rate per month to decrease as the business gets better at delivering value.
In addition, you should consider the realistic expected lifetime of the customer’s use of your product. Is your product something that a customer uses seasonally or all-year round? Is your product for a certain stage of life only (e.g. children’s educational products or menopause-related products) or is its use perpetual? Answers to these questions should feed into your churn assumptions.
Having arrived at your best educated guess at how many of your subscribers will leave you - and at what frequency - you need to apply this to your financial model. Once you're comfortable with the fundamentals of how churn works, you will need to model churn based on groups of people with similar expected churn behaviour, which are referred to as 'cohorts'.
A sensible financial model will show a steady flow of new customers joining and some customers leaving. The hope is that more customers join and stay than the ones that leave and over time, the churn rate reduces as the product improves and satisfies the customers’ needs better.
But, why does Bants only make £12k in year-1 again?
Thanks for reading so far. In order to tell the full story we return to how, if the Bants team achieves its target of 10 new customers per month, they might end their first year with around £12k revenue and not £36k as the "napkin maths" might suggest.
To simplify the illustration, I will assume that all of Bants' year-1 customers are paying monthly. Based on the customer growth and churn rates assumed in point 3, monthly revenue growth will be as follows:
By the end of the year, Bants will have acquired 110 new customers, lost 48 and retained 62. Churn rate will have stayed flat at a monthly churn of 10%, resulting in an annual churn of 45%. Total cash generated from customer payments will be £12k at the year end, although they will probably have spent a lot more. They will end the year with a Monthly Recurring Revenue(MRR) of £1.7k, translating to an Annual Recurring Revenue(ARR) of £20.6k.
To summarise, having a basic understanding of how recurring revenue models work is essential for any founder looking to build a business that relies on regular, repeated revenue. It lets you make growth assumptions that can stand up to scrutiny and helps reduce the number of surprises you get down the road, once you start spending your precious cash.
The three common mistakes I come across when helping companies build business models are:
- Assuming day-1 sales and not accounting for the sales cycle
- Recognising all your revenue up front when dealing with annual or pre-payment options, and then incorrectly modelling MRR and ARR
- Incorrectly modelling churn
Each of these mistakes, whether based on unrealistic expectations or a misunderstanding of the way these metrics work, are likely to result in inflated revenue projections which may look good on paper but won't help your business.
So, the next time you hear someone say the sales team will smash sales from day-1 and none of your customers will ever leave you…
As a side note, I recognise that understanding these concepts and building models is hard and that considering all these factors may run contrary to the optimism required to arrive at an exciting future projection and valuation. My advice is always to set stretch goals and challenge your team; however, you need to first understand what you are dealing with and start from a solid foundation.
It is much better to put some hard work in at the start in order to avoid being in situations where your projections don't stand up to scrutiny by investors or hawk-eyed team members. Worse yet, you may run out of cash three times faster than you thought you would, find yourself unable to pay salaries and lose control of the fate of your business.
On that bright note, thank you for reading, I hope you found it useful. :D