Financial Modeling for Startups: An Introduction | Hacker News Reader

Financial Modeling for Startups: An Introduction

fivecastfinancial.com

276 pointslhh33 comments

Finance person here, this is a good grounding of the basics. The hardest part to take forward is working out the timing of things. A company is constantly owed and owing money, and this is the real trick to working out your funding requirements.

On top of the model every business needs an operational cash flow forecast going out say 3 months at least. For every day you enter the brought forward balance from yesterday. Then you add and subtract all of the line items of cash inflow and outflow for the day to forecast a closing can balance. It is more than possible for your financing model to show profitability and yet to be insolvent, because you are paying money out before it comes in. Like maybe a big customer pays on the 28th but payroll goes on the 25th...

Cash is king as they say, and a daily cash-flow forecast is the main tool that a financial controller would use to maximise it.

I've heard that it's rather common to be technically profitable (i.e.: a company has a greater income than expenses), but nonetheless insolvent due to bills coming due before clients pay their invoices.

From what I was told, this mostly affects supply-chain heavy companies; software companies are mostly spared this kind of consideration.

What are some of the red flags that founders should be aware of when reading their own cashflow statements?

No, this also affects software companies, especially those that grow faster than you would organically. The cost in software companies is a cost in people. These people need to be paid. If you are not paid in advance for your software, then you have a cash flow issue: you need to have enough reserves to bridge the gap between paying your developers and you getting paid for your outgoing invoices.

Think of it like this: if you are providing a service, but are not getting paid for this in advance, you are essentially giving the customer a short term loan. You cannot infinitely provide those short term loans considering that you have bills/salaries to pay. It might still be money that belongs to you, but if you do not have it in your bank account when your bills are due, you are insolvent.

Things to be very wary of when looking on your balance is having very low ratio of liquid cash in respect to your debtors post (e.g. unpaid outgoing invoices). This means that if your debtors are going to pay later than expected, you have very little runway to cover that. Now, what risk this poses to your company depends a lot on how many customers you have, whether they pay their invoice automatically, and what their history of late payments is. Furthermore, the same applies for you on the purchasing side: if the majority of your costs are on the purchasing side and you can afford to pay those bills later without getting your servers shut down, then being illiquid is less of a risk. If the majority of your costs is in employees, then you are in trouble: not paying employees is a big no-no, so that increases risks and allows for less wiggle room.

I'm currently handling bookkeeping for software companies and one thing that's often overlooked is your clients consistently making late payments on their invoices. Make sure that you know who those client are and schedule accordingly.

DSO (Days Sales Outstanding) is one of the most important numbers to track for any company. If you are a small company it is arguably the most important one to manage cash flow

And it is a notoriously fickle metric, most people track it at month end, and yet their main sales receipts come in after... leading to a rather alarming result. DSO or debtor days as it is often called in the UK, needs to be read with an understanding of the payment patterns in that business.

In the UK the big risk is sales tax (VAT). It needs forecasting on top of sales receipts, and every quarter you have to pay this sum you have collected over to the exchequer. VAT is notorious for taking down businesses who spent the receipts!

Otherwise you should model what happens when the sales come late or not at the level you want, Braintree hold your cash, etc. If you can't flex your overheads to stay within your cash facilities, then you are risking insolvency.

Sales receipts, payroll and sales tax are the big numbers.

It seems like modeling could help get an idea of the size and frequency of each cash shortfall and thus inform how large a short-term credit line you needed?

Indeed negotiating a line of credit often requires a model.

Hi @jimnotgym, [Using a throwaway account] We built an active cash flow management tool to bridge the gap between finance and non finance people. We would love to hear your feedback on our product. Can you help?

The issue is forecasting revenues.

Zoom in and the issue is forecasting unit sales.

Zoom in and the issue is forecasting new unit sales.

In the example, this line does a lot of the work in the model: "Forecasting New Subscriptions (line 10). We've just entered hardcodes here for simplicity, but these could be the result of calculations related to a marketing / sales funnel"

I submit that this single assumption will carry more weight than the rest of the model, and is the most difficult to forecast.

Broadly speaking you should have a MQL -> SQL -> Deal model (i.e assumptions for ratios between the three) and assumptions for CPL and SQL per SDR/AE.

With these you can tie sales forecasts to marketing spend and sales hires.

Obviously these won't be perfect, but when you're off target you can see why (i.e which assumption was false) and then either try to fix it or correct the false assumption giving you a more accurate model going forward.

Yes, that is the unknown variable. However, your costs can be steered pretty accurately in software startups through hiring and firing. This means you can easily track whether your sales are still hitting the targets you expected, and if not, how much reduction you can accept into on the costs side before you need to look into getting additional capital investments.

Edit: For existing businesses this metric is much more predictable by the way, but especially in B2B it might be obfuscated because the finance department does not know how much value has been provided for which there was not an invoice created for it yet.

Indeed, but when your sales don't meet expectations in month one the model gives you a very strong clue as to what to cut or defer in future months. The model is not the business, it is something to measure the business against.

Won't argue with that, but it's nice to be able to put bounds on the result by using best- and worst-case guesses here.

This is great. If you like this sort of thing you can go one step down the modeling path and take a great coursera course called “Model Thinking” [1] which totally gave me a different appreciation for spreadsheet nerdery (you use lots of different tools).

1. https://www.class-central.com/course/coursera-model-thinking...

Why is it linked to an affiliate website and not coursera itself?

EDIT: here the original coursera link: https://www.coursera.org/learn/model-thinking

I was on my phone and couldnt copy the target link without the app opening the page instead. :/

When I had to produce my first financial documents like this, I went out and looked at public companies that had similar enough business models to mine, and read all their annual reports. Then I figured out what metrics are worth tracking or that I wanted to track, and pretty much just copied their methodology (which they described in their statements).

Another thing worth considering, if you hate The Sheet, or if it’s getting out of control, considering putting your data into a database and reporting with redash. It’s a also a very convenient way to share data internally.

Also, in my experience, the thing that mattered most was customer acquisition cost to lifetime value ratio.

Decent article, thanks for writing it. I think more founders should do a bit of financial modeling.

That said, I think most founders should not be forecasting salary expenses on a per-position basis, even if they're under 100 employees. In my experience, you definitely won't know which positions you'll be hiring for further out than 1 year. If you're trying to impress investors it might work, but it will have limited utility for you personally.

Instead, you should group salaries by function (e.g. sales, engineering) and then make explicit your assumptions about labour efficiency. In the model described in the article, these assumptions are also there, but spread out over 40 rows in a table - not good! Assumptions in models should always be explicit.

For example, you could say that, in order to maintain your projected growth, you need to spend 10% of your revenue on sales staff. Or if you're aiming to be funded, you might instead work out how much labour it might take to make one sale, and then extrapolate based on how many sales you intend to make in the year.

You could look at engineering and decide you need 1 person in your engineering team (disregarding job title) per 100 clients. Then extrapolate, once again, based on number of projected clients. Obviously, software is meant to be scalable, so this all depends on how much up-front development you intend to do and at what pace you intend to add new features, so you might want to factor your growth targets in too.

Now, organisations normally bring in layers of management as teams grow. Do you need to account for this? Probably not. Remember, we're focusing on labour efficiency. These managers might increase your costs, but the idea is that they also help your teams function in a scalable way. And if you have good managers, the average tenure at your company should increase, leading to higher productivity.

The benefit of the above approach is that, now that all your assumptions have been made explicit, you can easily tweak them to see how they impact your model, rather than having to dig through many rows of data.

Lastly, and this is nit-picking, but ignoring income tax means this model should only be used to forecast up to periods where the company is not profitable. As soon as there's profit it will be completely wrong. Although it's a nice simplifying assumption if all you're trying to model is your road to break-even.

I prefer the bottoms up expense buildup of people as opposed to a broad percent of total calculation as it introduces more discipline to the forecast and reduces risk of weird situations where sales go up by x% and all of a sudden you’re hiring 3 half people (in models I expect to see expenses go up in steps because that’s how they work in practice). Also gives insight into the mindset of the forecaster. For example do they assume 2 sales people can cover 300 accounts while building a pipeline without SDRs? I am an operator, not a VC so perhaps it’s an audience preference.

I think it would be good to have a bottom up forecast for the first 12 months. Apart from that, a disciplined forecast will clearly show how the calculation was derived and what assumptions went into it. You should not have to dig through a bunch of data to gain insight into the mindset of the forecaster - the forecaster is supposed to put that info in the model! If they can justify year 1's sales expenses by mentioning that it's for 2 sales people + an SDR, then great! Just don't expect that granularity in year 3.

It's quite easy to avoid the "3.5 people" issue by making it a step function (i.e. rounding). Once again your assumptions become explicit, which is good. E.g. you might decide that one person can do the work of 1.3 employees (people can do this for a while when it's needed!) and round everything above that up to 2.

However, when forecasting 3 years in advance as in the article, the fact that your model has you hiring fractional people becomes less important.

> That said, I think most founders should not be forecasting salary expenses on a per-position basis

It is still very useful for variance analysis. Like, I made 100k, expected 120k...because x person cost more than expected and x person was hired early. It's nothing to get upset about, but it aids your understanding.

If you're forecasting at such a granular level 3 years in advance, your variances will be all over the place, which is not terribly useful to analyse. It won't be as per your example. It will be "I thought I'd hire one of person x, but instead ended up hiring 2 of person y, and delayed hiring z to compensate". The aggregate variance is what matters when you're doing a long term forecast.

A granular 12 month forecast is very useful for the reason you described, but we're discussing longer time horisons here.

Pros and cons to each approach. I tend to recommend that it in the early days companies build up by employee, and then look at the metrics coming from the other direction as well to make sure everything makes sense (e.g. if you're forecasting that sales will triple, can you accomplish that with the hiring you've projected).

Main reasons for this are:

- It enforces discipline. It's easy to make hand-wavy assumptions like forecasting costs as a % of sales and calling it a day

- Small changes to the hiring plan can have a drastic impact on a startup's finances, including the timing of those hires

- If you tether all your expenses to sales, you can't really explore the downside, because you'll always be showing consistent profit margins

To your point though, I agree that it's important to look at expenses both bottom up and top down.

I was just thinking about using stats in startups. For sure you need to know your runway (and burn rate) - but then for SaaS (and similar business with recurring revenues) the first useful statistics is churn. Customers quiting is the most important info when looking for product market fit. Customer groups with lowest churn are your target group (market). Before you start scaling business you need first to decrease churn - because you don't want to grow the holes in your system together with it (this metaphor is not mine - but I cannot now find where it came from - but I googled this article instead: https://pakman.com/churn-is-the-single-metric-that-determine...).

Sales statistics become important after you have satisfying churn rate (i.e. after product market fit). With advertising you can quickly grow sales - but if your churn is too big - then the LTV (https://en.wikipedia.org/wiki/Customer_lifetime_value) will be too low. And generally customers that come from advertising will have a bigger churn than those who learned about the product organically.

Nice writeup. A previous company had us participate in a simulation exercise. Split into teams, move around a board similar to monopoly except you were collecting accounts payable from customers, paying rent and utilities, paying payroll, paying taxes on xyz, lawsuits, etc. CashFlow was king. Losers all were paying bills as they arrived rather than on due dates. They had the money, on paper, but not on time. Timesheet-driven contracting - it showed how important it is to fill-in timesheets so customers can be billed. Miss the cutoff and your 48 hrs gets invoiced two weeks later. Your company doesn't thrive when it's borrowing money to pay you for time it can't bill yet because you forgot to sign your timecard.

Good brief summary but sometimes high level is just as important. eg. Using the tools to explain how you prioritize spending money raised and connecting it to material signposts is very helpful for funder communications as well as testing out your own ideas in a cohesive framework.

Good basics but the article is too much on the simple side for me.

"There are no hard and fast rules on how to categorize your expenses"

Pro tip: Loosely aligning them within recognized tax lines can help simplify things in the early days and reduce work for your accountant.

In my experience with early stage startups, these models are often pure fiction and completely divorced from reality.

They look nice to investors. Reality is very, very different.

(Ex VC here)

Before you launch, yes (ish) - it can give you an idea of where you'd like to start charging and why. Models come into their own post launch. They provide a clear structure on how to optimise the economics of your business. E.g. If you are currently selling at $20 and losing $3/sale due to support costs and returns, it provides a great structure to focus on a) increasing the price, b) reducing support costs / order and c) reducing returns.

A dream for any VC is a startup with fantastic economics at day 0, but this is rare. The majority of high quality companies have negative unit economics during their infancy. We liked investing in companies with negative unit economics with founders that understood the drivers of their economics deeply and were optimising them aggressively on a weekly/monthly basis (and doing all this inside of a H-U-G-E market). (The best ones did it on a weekly basis.)

An interesting example is Just Eat, one of Europe's best performing startups (IPOed at £1.5bn). This company had negative economics for ~3 years, but the market was very big and investors could see a pathway to positive unit economics through optimisation. This allowed them to fund the company through the negative UE period.

As they say, "Plans are worthless, but planning is everything."

nice post. also in the 'financial modeling for startups' space: https://foresight.is/

Reply on news.ycombinator.com