Johnnie Ball, Co-Founder and Chief Data Officer, Fluidly
Short-term forecasting is essential, yet difficult to achieve. Producing an accurate look at the cash position of a business seven, 30 or 90 days into the future is notoriously complex. Most short-term cashflow tools rely only on forecasting bills and invoices. This can lead to substantial inaccuracies and can actually be dangerous for a business, providing a false sense of security.
82% of small business failure is due to poor cashflow management and 69% of small business owners have been kept up at night due to cashflow worries. For the health of both small businesses and the people who run them, accurate short-term forecasts are vital.
Producing a good short-term forecast comes down to two critical factors:
- Understanding when money is really going to come in to the bank account
- Predicting your biggest outgoings, like rent, salaries and tax. These outgoings aren’t generally invoiced, which adds to the complexity of predicting them accurately
Let’s consider the cash inflows and outflows in turn:
Short-term forecasting cash in
There are lots of ways to get money into a business, but for short-term cash it usually comes down to:
a) when are your customers really going to pay
b) what other cash might be collected
The latter is especially important if a business is taking cash or credit card payments without issuing an invoice. This type of cash collection covers a whole host of business types, from subscription businesses to shops to e-commerce.
When are your customers really going to pay their invoices?
Most short-term forecasting tools are overly dependent on the date that invoices are due. If a business has 30 day terms, forecasting tools assume that all outstanding invoices will be paid exactly 30 days after they are raised. And no wonder; it’s relatively easy to line up the invoices for a business and fill in when they are due. But given the amount of invoices that get paid late, it’s a very inaccurate way of predicting real payment behaviour. As a result, it’s a dangerous way to cashflow forecast.
The effect of late payments on small businesses has been widely examined in recent years. According to Xero in 2019, 48% of invoices issued by small businesses are paid past their due date.
Forecasts based on due dates don’t take into account that half of all invoices are not paid on time, so at least half the predictions are wrong from the start. This problem isn’t easy to solve, as it’s more difficult to predict irregular payments that don’t follow a pattern. It’s something we’ve worked hard to overcome at Fluidly and why our cashflow forecasts are not based on invoice due date, but on the actual historic behaviour of each customer.
Our machine learning algorithms are trained on this living, breathing invoice payment activity. So, the customer that always pays 10 days late, or the one that doesn’t pay until chased are properly accounted for within the forecast. This avoids producing a misleading forecast that shows cash coming into the bank, when it really isn’t likely.
How to predict cash coming in if you’re collecting money without an invoice?
When you’re operating without sales invoices, it can be difficult to predict cash movements. This applies to many businesses, especially those that collect payments via cash, direct debit or credit cards like restaurants and shops.
In this scenario, it’s important to look at what’s happened previously in order to build a picture of what might happen in the future. This is where machine learning is really useful. It can pick up historical patterns in actual cash collections and build that into a forward-looking forecast that factors in not just the invoices raised, but also cash incoming that isn’t invoiced.
Short-term forecasting cash out
Beyond bills: predicting the most important cash outgoings
Many forecasting tools predict when bills are going out based on the supplier invoices received. But a significant proportion of a businesses outgoings are rent and salaries, and these generally aren’t invoiced. This is a big limitation of forecasting that’s linked only to invoices, as it doesn’t begin to cover all the outgoings. If you’re missing a large percentage of money you owe, this renders the forecast useless, even dangerous, to rely on – providing a false sense of security.
Forecasting like this can lead to businesses operating on unfounded assumptions. There are also other payments that would be missed, like recurring non-invoice payments and direct debits, to name a couple.
Fluidly accepts that invoices and bills only ever give a partial picture. It’s crucial to look at the actual payments in and out of a business, and understand the patterns behind this. Fluidly builds a holistic picture of when all money moves in and out, not just relying on what is already present in the general ledger.
Fluidly can link predicted payments with invoice numbers when they have already been raised, but it also covers transactions that although not invoiced, are still key cash movements, such as salaries.
Intelligent Cashflow forecasting
Short-term cashflow is an intricate problem, that requires a sophisticated technological solution. In order to save small businesses from failing due to poor cashflow management, it’s critical that we produce the most accurate possible short term prediction.
When considering a cashflow tool, take the time to understand:
- How it handles invoice predictions – whether this is based on optimistic due date assumptions or real-world payment behaviour
- How it incorporates the critical non-invoiced transactions like cash sales, rent, salaries and direct debits
To find out more about Fluidly, click here.
