Liquidity management – Generating cash effects through order-to-cash processes

Definition of the order-to-cash process

The O2C process usually begins with customer requests for goods and services, subsequently continues into the provision of those goods and services and finishes up with receipt of the payment from the customer. During this process the same business transaction passes through various items in the P&L, balance sheet and cash flow statement. Sales and liquidity that are anticipated in the forecast turn into actual sales with real cash effects that are reported. Outstanding receivables turn into real cash inflows.

Please note: A well-structured O2C process will secure the operating cash flow, i.e. the liquidity in the business.

Liquidity management objectives

Besides receiving payment, the process goal consists in calculating a realistic liquidity forecast of incoming payments. For many companies, efficiently and effectively organising an O2C process with respect to both goals will pose enormous challenges. These arise out of, among other things, large numbers of customers, the amount and quality of daily data updates as well as the involvement of various specialist departments. In particular, answers to the following questions will be of key significance:

• When do the customers actually pay?
• Which customers have to be encouraged to pay and how?

The processing of enormous amounts of data sometimes gives rise to intensive deployment of personnel as a problem that has to be solved. It is important to have a holistic approach to the large number of customers, their invoices and other transactions (e.g. invoice correction, dunning notices, partial prepayments). This is why unstructured processes can result in negative cash effects.

In order to forecast incoming payments, usually, customer receivables including the agreed payment targets as well as open items from dunning runs are processed. A forecast for incoming payment can admittedly be derived from this basis. However, that alone will not result in satisfactory answers to the above questions.

Recommendation: The management must also have an interest in automating many process steps and also obtaining reliable projections for liquidity forecasts.

Approaches to problem solving that are focused on automation and improving quality

There are tried and tested approaches available for the required automation and the desired quality improvements. These can be executed, for example, in the following steps (cf. Fig 1).

(1) Customer clusters – To answer the question when customers actually pay, among other things, historic customer behaviour as well as other appropriate master data can be considered. For example, on the basis of historic receivables and when payment was actually received it is possible to derive conclusions about future incoming payments in the case of regular customers. By contrast, for new customers so-called peer groups can make this possible. With the help of statistical cluster analyses it would thus be possible to create customer groups (cf. illustrative example in Fig. 1) in terms of realistic incoming payments. Statistical methods and support provided by an IT system enable customers to be grouped on the basis of facts. These days moreover there are sophisticated technologies available for this.

(2) Automatically derive measures – In the case of receivables management using large amounts of data, deploying resources efficiently and effectively specifically means that transaction costs have to be reduced and incoming payment maximised. In doing so, statistical procedures replace manual processing of lists of open items (for instance, simply from top to bottom or along a decreasing size of receivable). Automation will enable you to correlate receivables with additional data (such as, for example, historic payment behaviour and previous success of measures) and determine a weighted payment probability and default rate per customer. With such information for each customer or customer cluster you can automatically derive recommendations for actions that can be implemented. The recommendations make it possible to focus on the cases that require undivided attention.

By way of illustration, the example presented in Fig. 2 shows that the underlying statistical procedure can provide valuable information. According to line 4, there is an 80% probability that Cluster Z1,…,p will pay up when an e-mail including a firm reminder are sent. Sending a direct e-mail is thus better since this increases the expected value of the incoming payment by € 40k without tying up further resources. When the amount of receivables is € 1,500k then a cluster with € 50k (3.3% of share of overall volume) would conventionally have low priority in terms of processing. The cluster would have been examined last of all and processed according to the plan (e.g. call first).

(3) Continuously improve forecast quality – The quality of forecasts can be improved if the actual situation that has been realised is routinely compared with the forecast values. It is, in turn, then possible to draw conclusions about the statistical model from these data in order to continuously improve the forecasts. Besides manual checking, it may be appropriate to use artificial intelligence.

Benefits at a glance

• Introducing more structure and more automation to O2C processes will allow precious employee time to be reduced as well fact-based liquidity forecasts to be prepared.
• Moreover, the age structure of the receivables and the administration costs can be reduced and positive contribution margin effects generated. Furthermore, by using customer clusters combined with expectations based on payment track records it would be possible to automatically derive effective measures.
• In this way, companies will be able to purposefully guide actions towards risk issues in order to cut delays in payment together with the process costs.