This issue’s question:

“Forecasting and liquidity planning is often cited as a major challenge for treasurers. What would help gain the information needed?”

It is treasury’s responsibility to obsess about liquidity and to ensure that stakeholders from across the business understand why effective cash forecasting is so crucial to the financial sustainability of the company. However, a finance team’s efforts can sometimes be disproportionately skewed toward profit planning at the expense of cash forecasting.

To help redress the balance, and hopefully improve the quality of the forecast, it may help to focus on the following:

Obtain leadership buy-in. The support of the Finance Director is a must. Success lies in ensuring that sufficient resource and attention is dedicated to cash planning and forecasting.

Benchmarking a firm’s liquidity provision against industry peers and stress testing annual forecasts at their cyclical ebb are useful techniques to help quantify the exposure. Hopefully it will show how well positioned your firm is, but if not, this type of simple analysis can help bring to life the liquidity risk that exists so it can be mitigated.

Actively engage with internal stakeholders. If the teams that you depend on for information don’t appreciate why accurate cash forecasting is so important, they will inevitably de-prioritise it in favour of activities they think will add greater value. Proactive engagement with teams that feed into the forecast process is therefore critical.

A brief discussion at a functional team meeting can help de-mystify the world of treasury and provides an opportunity to evangelise the importance of effective liquidity risk management.

Find an enthusiastic analyst. Most finance teams are brimming with talented analysts, eager to immerse themselves in an interesting change project. Some tips that might support them in this endeavour are as follows:

• Source at least two years’ worth of daily cash flow data and ‘clean’ it to remove any one-off events (such as disposal proceeds or national events) that might skew the underlying data.
• Identify any predictable events (for example payroll, tax, VAT or rent) and remove them from the data set until you are left with a cash profile that is essentially a trading cash flow.
• Obtain the latest budget/plan and overlay the monthly cash flow figures into the model so it produces a phased daily forecast by month – adding back known periodic or unique events on the appropriate dates.
• Use your judgement to sense-check outcomes and listen to your analysts – they will doubtless have some great ideas that will enhance the quality and presentation of the forecast.
• With this baseline long-term forecast you can add payment data from your ERP system to create a more accurate near-term view.
• Track actual daily performance against the model and continually refine.

To ensure that a business can actively manage its liquidity risk, you need good data which requires support from across the business. If internal stakeholders appreciate the consequences of getting it wrong (bankruptcy!) and understand how valuable their contribution is, they will want to help you get it right and things will improve.

Forecasting is part art, part science, part human intelligence, part machine intelligence. While forecasting is done differently across industries, in all cases, treasury first must understand the sources of data and how they impact corporate operations, and then apply technology to make meaning of the data.

Relevant information used for forecasting can be found within the siloes of an organisation and beyond the enterprise, creating complexity, limiting data sharing, and resulting in inefficient data gathering processes. Costly searches and manual sorting lead to data integrity problems, which compound as organisations expand and acquire businesses locally and globally.

Once sources of data are understood, the next step is to centralise and normalise the data into a single source of truth. This can be accomplished through imports/exports from many source applications to an open, cloud-based TMS platform, using APIs and robotics to automate this process seamlessly, without human intervention. System-to-system communication eliminates human errors, and by removing manual information gathering, teams have the time and the analytical power to perform strategically, rather than operationally. This is beneficial especially as market conditions change.

When data is centralised, it must be flexible enough to manipulate. By default, treasury reverts to using Excel because users can quickly change the data. However, spreadsheets do not allow management to audit or identify trends, and they only provide static views for manual comparison. While management wants to give their teams autonomy to amend data as their business changes, they must still be able to measure changes in forecasts and monitor patterns dynamically. Treasury systems today automate this process, and if they are SaaS-based, they allow everyone to work consistently off the same application, regardless of where they are globally.

After treasury has extracted relevant data, centralised it on a common treasury management platform, and has been able to amend, analyse and archive different versions of forecasts, the TMS can perform the heavy lifting. Treasury management systems should be able to apply sophisticated simulations to predict financial viability over time using historical events. Advanced calculations, AI and machine learning is the science behind the art of accurate and efficient forecasting. Treasury management systems that can identify data patterns can bring organisations one step closer to predicting the future. Once implemented, statistical analytics become any organisation’s true competitive advantage.

For more than a decade, liquidity and forecasting have remained in the top three challenges for CFOs and treasurers globally. This begs the question: why has this been a perennial challenge for so long? The reason: treasury operations today are, for the most part, a series of unintegrated systems, spreadsheets and silos between groups and other departments.

Companies are often faced with multiple ERPs, many entities, and different currencies. These make the task of managing liquidity a major challenge, not to mention a significant manual effort involving many people. The result: lots of time spent gathering and validating data while still not having a full, transparent view into the numbers. The volume, variety, velocity and veracity of data generated each day has made traditional analysis – using spreadsheets, for example – obsolete. It is just not possible to manually aggregate and analyse that much data with sufficient speed to be able to gain insight, and then turn that insight into action.

What should a treasurer or CFO be asking themselves?

• Can you identify all your sources of data that you need to make a cash flow forecast? Eg ERP (how many do you have, are they all on the same instance), CRM, bank statements, trend analysis, manual data (such as budgets).
• How often do you refresh your short-term/mid-term cash forecast? (Daily, weekly, monthly, quarterly, or I don’t make a cash forecast).
• How do you ensure no mistakes happen in your data capturing/consolidation?
• How do you incentivise your subsidiaries? Local subsidiaries and users typically download information from their ERP, and upload in other types of files to HQ, or in SharePoint, or they will just send Excel files from all over the globe to HQ, which means it’s 100% manual. There’s no real alignment of the processes across subsidiaries and no audit trail at the local level.

What to consider?

• Companies should ensure their information is system-based. In other words, they have full integration with their ERP, so they don’t have to manually download data (it should flow automatically).
• Any augmentation of data should have an audit trail so that, ultimately, the group treasurer can see who did what, and when they did it.
• Automate the process and deploy alert functionality, such as reminders for subsidiaries to post their local forecast, and for the group treasurer to look for it.
• Ensure bank connectivity to enable comparison of actuals with forecast figures.

I have the data. Now what?

With this data, treasurers should now be able to answer these four key questions: what happened; why did it happen; what will happen; and what should be done?

• Descriptive analytics answers the question, “what happened?” This is the most basic form of big data analytics, and provides a picture of past events.
• Diagnostic analytics, “why did it happen?” Diagnostic analytics enables you to perform root cause analysis and use that information to prevent future repetition of events.
• Predictive analytics, “what will happen?” Predictive analytics uses advanced algorithms – often with artificial intelligence and machine learning – to forecast future events.
• Prescriptive analytics, “what should I do?” Prescriptive analytics tells you what the best steps are to achieve a specific result. Prescriptive analytics requires advanced machine learning capabilities.