In an age of business uncertainty, liquidity forecasting should be high on the agenda - but there’s also relentless pressure on costs, particularly for perceived ‘back office’ tasks. So any chance to improve performance and efficiency should be welcomed, especially if it can be achieved at low cost/effort, and free up resources for more valuable activities.
This article looks at some areas that are often below the radar, but may be serious inertial drags on achieving better results – and is intended to prompt you to consider whether there are practical opportunities for improvement. You may not currently be losing sleep over any of these questions, but if you haven’t reviewed them recently, now could be a good time...
Q 1. Do we have the right approach to supporting technology?
A large majority of organisations use spreadsheets as the main, or only, technology for liquidity forecasting. The reasons appear obvious – spreadsheets are very adaptable, easy to use, and effectively free (as a sunk cost).
But there are real costs to reliance on spreadsheets, which are often overlooked. They are indeed flexible, but this comes at the price of significant manual effort. They are not inherently designed to support process flow, requiring an additional (usually manual) layer of process control. And they are prone to operational error, from simple input error to fundamental compromises in functional integrity (e.g. unknowingly deleting a row in a vital decision-making tool). These factors increase process effort, especially in relation to data manipulation, data analysis and reporting.
One solution is to use automated data extraction from your financial systems – this relies on the availability of reliable AR/AP data, and can involve significant costs in design and configuration. Another is to leverage your TMS, as most have liquidity forecast modules that support seamless consolidation of business inputs into the central treasury positions, and the incremental cost is generally low – they tend to have limited analysis and reporting capabilities for non-financial flows. A different option is to consider one of the growing number of dedicated applications, which now include process flow controls, systems data interfaces, analysis and reporting, and modelling functionality (CashAnalytics, CashForce and Exidio are examples that give a flavour of what is available) – they do have a cost, but this should be compared to the value of process security, and the time freed up to deliver insight and strategy. After all, that is what you employed your resource talent for, rather than spreadsheet maintenance, isn’t it?
Q 2. Can we reconcile our liquidity forecasts to the cash forecasts produced by FP&A?
This may not appear to be a major issue – after all, liquidity and FP&A forecasts usually have very different uses, and infrequent points of comparison. And as a treasurer, you may think that if there is a difference between the forecasts, it’s not your problem, since the liquidity forecast is likely to be more accurate, given its short horizon.
The CFO may not agree! Since it is the FP&A numbers that are provided to investors and the market, it is effectively the orthodox view – and as the quarter or year-end approaches, if there is a significant divergence between the two forecast outcomes, explanations (often detailed) will be required in short order. In the worst-case scenarios, treasurers can find themselves potentially jeopardising customer and/or supplier relationships in order to manage actual liquidity in line with FP&A expectations. This is avoidable effort, at a busy time.
FP&A projections are normally focussed on net debt and prepared in an indirect format – which may be difficult to relate to a receipts and payments forecast of bank liquidity, except on a total cash outcome basis. However, this should not be an excuse to file the problem under ‘too difficult’. By identifying big ticket items that can be directly compared (e.g. financing, tax/VAT, payroll, capex) the reconciliation can be narrowed down to working capital movements – and working back from the P&L/Balance Sheet, it’s possible to compare receipts at an overall level. Having a regular disciplined process to compare and at least sense-check the forecasts will reduce the chance of nasty surprises, and may also improve FP&A forecast accuracy, by providing greater insight into working capital cash cycles.
Q 3. Are we investing the appropriate level of effort in the process?
Routine tasks tend to become fossilised, and liquidity forecasting is no exception – ‘we do it this way because we’ve always done it’. Worse, ad-hoc additions (e.g. management requests for more detailed reporting to address one-off circumstances) tend to become embedded in the regular process, just in case they are needed again. This legacy burden can result in an unnecessary level of effort, at a time when there is a squeeze on resource.
The acid tests to apply to forecast outputs are a) do they provide insight? And b) are they actually used for anything? If the answer to either is uncertain, it’s time to question whether the effort justifies the benefit. In an ideal world, you could perform a thorough review of existing processes, in the context of overarching business objectives, required levels of accuracy, and supporting technology – and re-design outputs so they meet business needs at an acceptable level of effort.
However, few treasury teams have the time for such a drains up exercise. Pragmatically, it’s easier to focus on two aspects (forecast horizon and level of line item detail), using the acid tests above. The forecast horizon especially can be challenged – thirteen weeks is a common default, but most businesses have supplier/customer terms under 90 days, which means that the full forecast horizon can’t be supported (and therefore automated) by systems data. Completing the forecast horizon requires a supplementary hybrid/modelling process, which is effort intensive and often opaque. Moreover, few organisations measure accuracy over the full horizon, which means forecast reliability is unknown (so is it useful?).
Q 4. Do we engage regularly with people in the business?
Even the simplest processes can be enhanced by committed people with appropriate expertise – but if they are not engaged, much of the potential talent will remain untapped. Businesses are also dynamic – operations, customers, suppliers and terms all change, and key personnel move. But you can’t be aware of what you don’t know, so if you don’t have regular communication channels, you can lose sight of the underlying cash cycles across the organisation, and worse, not be sure of who actually does have that insight.
Variance reporting is a vital component of the forecasting process, and should provide meaningful insight into any gaps in expected liquidity outcomes. However, it is a blunt instrument when compared with a more personal two- way relationship – and you never know when you may need to ask a favour from someone in the business, to meet an urgent information demand from senior management.
Establishing forums for ad-hoc and regular communication (not just reporting), across the business allows you to tap into local knowledge, promote common understanding and spread good practice – e.g. through monthly practitioner calls to periodic team-building conferences. This creates an environment where the star performers in your organisation can actively help you in improving overall performance – a potentially significant long-term saving of effort.
Q 5. Are we visibly recognising and rewarding good forecasting performance?
Business teams are faced with increasing workloads, so routine tasks tend to be prioritised on a basic hierarchy – from tasks that are not formally monitored/incentivised (treated as least important by the ‘doers’) up to tasks that are both measured and feed directly into business incentive schemes (seen as the most important). Unless cashflow is in crisis, liquidity forecasting is a routine process, and often falls into the unfortunate category of tasks that are monitored, but not part of formal incentivisation – if it goes wrong, management wants to know why, but there’s no pay-off if it goes right (i.e. all stick and no carrot).
Given this, you shouldn’t be surprised if people do the minimum to get the job done – after all, there are plenty of competing priorities. This results in lower than optimum performance (in terms of accuracy, timeliness and insight) – and attempting to counter by constant management repetition of the importance of the task will only have a short- term impact, unless good performance is reinforced with positive benefits – without them, it’s just more ‘stick’.
It’s not straightforward to embed liquidity forecasting into formal incentivisation (defining balanced metrics is a challenge), but there are initiatives that you can implement locally, including ‘naming and faming’, mini-rewards, and role model/SME status, both at a team and individual level. And it’s surprising how far a little bit of ‘carrot’ can improve morale and performance – try it and see!
None of the issues above are individually earth-shaking – but by addressing even a few of them, you could improve performance and reduce process friction, with little or no additional cost. That should be worth some consideration.
Cash flow forecasting should not just rely on Payables and Receivables entries from the ERP files
Brian Shanahan from Informita has found in his long career in working capital management that to produce accurate short term cash flow forecasts, it is not enough to rely on the actual dates in the A/R and A/P files
Forecasting risk is now a major concern in today’s companies, survey
Increasing integration of FP&A and risk management data collection, etc. needed to improve analysis
Cashflow forecasting: new opportunities for significant improvement
How Integrated Business Planning can overcome fragmented planning, budgeting, and forecasting processes to achieve highly accurate forecasts, and dramatically reduce time and cost