Predictive Analytics with TIP allows you to significantly automate your cash flow forecasting and standardise predictions and all of the underlying assumptions. This provides you with amazingly accurate forecast data and saves no end of valuable time.
It goes without saying that a computer is capable of processing enormous volumes of data and performing highly complex analyses that you couldn’t manage with Excel alone. Computing power and statistics alone, however, are simply not enough when it comes to generating reliable forecast data.
That’s why we have developed a multi-level procedure which significantly improves the precision of predictive analytics and therefore makes this concept fit for everyday treasury use. The treasury information platform TIP relies on this procedure to individually determine the best calculation method for every one of your forecasting categories in order to deliver amazing results. In combination with the expertise of your treasury department, this allows you to highly automate your forecasting and to improve the quality of your forecast figures.
During this process, simply use TIP to define and calculate information about all the influencing factors relevant to your cash flow development (Fig. 2) and therefore digitalise your cash flow forecasting. This is the key not only to flexibly tailoring your forecasting to various requirements throughout your corporate group but also to making it independent of specific individuals. The value of this becomes clear when personnel leave the company or are on vacation or parental leave.
This digitalisation of the process is the greatest advantage offered by predictive analytics. Why? Because it enables you to simultaneously take numerous influencing factors into account in your forecasting. This is nearly impossible to do manually. Statistical models and machine learning methods, however, can do this and therefore also generally make more accurate forecasts possible. Read the following article to find out more.
Many treasurers rely on Excel to analyse the cash flows of previous years in order to identify trends and patterns as a basis for being able to predict future developments. Predictive analytics is based on the same principle but on a far more technologically advanced level and with a significantly higher level of automation.
The most important ingredient for accurate predictions is your historical data. These data are analysed for trends and patterns using a statistical computation model and projected into the future. For this to work, the system needs to calibrate the model in advance so that it can categorise recurring events (e.g. the impact of public holidays) and ignore data outliers (Fig. 1).
While this might sound simple, it is actually a complex process during which the system needs to consider numerous variables and parameters and to both review and update these at regular intervals. There are also nearly twenty different statistical models available for the calculations and the suitability of these models varies significantly depending on the forecasting category and data quality. The selection of the right model is therefore equally as important as its modelling and needs to be reviewed just as often so that the calculations can improve, in terms of accuracy, over time.
The first step is for us to jointly analyse which data are relevant for your predictions. These data need to span a period of at least three years if they are to be used for predictive analytics. As soon as the data landscape has been clarified, TIP automatically integrates these data from existing upstream systems into its cash flow forecasting.
In addition, it of course has to be defined which currencies, over which forecasting horizon (on a daily, weekly or monthly basis) and for which future period the forecasts are needed for every forecasting category.
An in-depth understanding of your group’s business model is fundamental to the analysis. Which external and internal factors influencing your cash flows will you need to calculate in treasury in the future? The assessment of these factors is ultimately the key to reliable and robust forecast data. That’s why our data scientists discuss all known influencing factors (Fig. 2) with you at the outset, how these can be modelled, and everything that needs to be taken into account. This initial step is decisive because statistical methods can only deliver meaningful results if they have ‘learnt’ in advance not only how to identify only one-off data outliers but also regular dependencies on other factors.
General influencing factors exist which, even if they may be very different in terms of their impact, are felt in every industry and every company. Public holidays, for example, vary not only internationally but also nationally, from state to state, and their impact on the retail sector is completely different from that on the tourism industry. The same of course also applies to seasonal variations and trends. There is one thing however that all of these factors have in common: they are relevant to nearly every company, have massive impacts on cash flow developments and, therefore, have to be modelled before the first test run. While the effects of public holidays can be easily modelled by uploading a calendar, seasonal effects and trends are mainly identified by analysing historical data. This step is handled automatically by TIP in an analysis of whether seasonal effects, trend shifts and catch-up effects after public holidays exist in the various forecasting categories.
Specific influencing factors impact your industry or even just your company. The data landscape and the preparation of the data for modelling purposes therefore also need to take these factors into account. TIP models external facts using general economic data, such as exchange rates, which can usually be sourced from data providers. We upload internal factors, such as fixed payment deadlines, so that these can be integrated into the computation models. Many specific factors are initially not even known or the data landscape is insufficient in order for these to be modelled. The longer you use forecasting, however, the greater the reservoir of data which can be used for re-adjustments. This means that your forecasts will become even more reliable over time.
The key challenge in dealing with predictive analytics is determining which statistical model is suited to every single forecasting category.
We currently rely on eight model groups, each of which has different strengths and weaknesses. Regression models, for example, are able to deal easily with missing data, which are a challenge for models based on ARIMA. Neural networks are very good at identifying patterns but require a large volume of data as a starting point; data which don’t exist for a forecast on a monthly basis. An automated analysis of the existing data is all it takes to identify which models can immediately be rejected and don’t even need to be subjected to a test run. This is extremely important since numerous forecasting categories and models, as well as their modelling, require considerable computing power, meaning resources both in terms of computation capacity and money. Minimising the scope in advance to promising models is key to the efficiency of this procedure.
The data analysis generally delivers a pre-selection of five to six models for the test run. TIP calibrates the models using the influencing factors already identified and then calculates a test forecast for the most recent existing twelve months on the basis of actual historical data (at least three years’ data are necessary for this). Deviations of this third year to the actual data are subsequently analysed for every model and the best model is then selected.
As soon as the parameters for the forecasting and the necessary data have been defined, the entire process, from the analysis of the data to the pre-selection, calibration and validation of the models as well as the ‘final’ selection of the model and the calculation of the first forecast, is largely automated. Only data outliers need to be considered individually and adjusted for in order to improve the accuracy of forecasts. You can change the selection of the best model automatically based on even minor deviations or adjust this manually based on tactical considerations. This might make sense if, for example, the historical data are dominated by extraordinary influencing factors (e.g. an economic crisis, natural disasters or pandemics) for longer periods and therefore do not represent normal developments. In such cases, it is possible that a particular model performs well while another one would nonetheless be better in the long term.
The forecasting data generated provide a sound basis for Treasury to leverage its specific expertise to swiftly prepare a professional cash flow forecast. For some forecasting categories (e.g. salaries), this works so reliably that TIP can even directly generate a cash flow forecast if wanted. For categories which are often dominated by exceptions, it makes more sense to display the calculated figures in the standard forecasting screen as proposals. In this case, the existing data act as a guide for manual fine-tuning and save no end of time (Fig. 6).
Not only the forecast data themselves but also the estimates and observations based on these are extremely important. TIP gathers all of the relevant information in a report which is generated along with the forecast data:
This information can help every treasurer to quickly understand how the forecast figures were generated. This can be helpful not only in swiftly identifying forecasting errors but also in understanding which factors are responsible for a sudden change. This delivers not only transparency but also makes it easier to perform checks.
The first run still requires some attention and manual tasks but, as soon as these have been performed, the system takes over more and more of these steps itself. And the calculations become increasingly accurate over time as further actual data in the right degree of granularity become available. That’s why TIP checks before every new forecast whether the model selected is still the best option or whether another model would deliver better results. The selection process here essentially remains the same, with the only real difference being that previously modelled influencing factors are applied for the calibration completely automatically which, in turn, means that the test calculations become increasingly accurate over time. It goes without saying, however, that Treasury can calibrate the models with additional or better influencing factors at any time to improve the predictions.
The more volatile the cash flows of a company become, the less likely it is that the models will be able to identify useful patterns in the historical data. That’s why this procedure is best suited for the B2C sector and for subsidiaries with a broad customer base since the development of cash flows in these cases is not dominated by major one-off payments. Nonetheless, this procedure also offers advantages for small subsidiaries with major B2B projects. Even if the data in such cases need to be carefully checked and manually adjusted, this procedure provides the advantage of digitally documenting known influencing factors and can therefore be easily referred to, making forecasting digital. Even when key personnel leave, their knowledge stays and the quality of predictions is safeguarded.
The aim of this summary is to as transparently as possible set out and explain the advantages of predictive analytics. We know that the process as a whole is detailed and therefore complex, which is why we would welcome the opportunity to discuss the issue and advise you personally. Hit us with your questions and we can then clarify together whether predictive analytics could also pay off for you.
The 8th Cash Management Campus of DerTreasurer was held digitally for the first time as a result of the COVID-19 restrictions but was nonetheless well received by participants. Hardly surprising given that it is currently more important than ever to inform yourself about developments in the market and, not least, about solutions for system-based cash flow forecasting. Our client Corestate was also among the presenters. During his presentation, Tobias Wriedt (Head of Treasury) explained how Corestate has reorganised its treasury system landscape and which factors were important here. Corestate Capital Holding S.A. is a publicly-listed real estate investment manager and co-investor with around €28bn in assets under management and is therefore one of the leading providers of integrated real estate investment solutions in Europe.
Among other factors, it was important for Corestate that the new treasury system landscape be able to ensure the smooth exchange of data both internally and externally. The aim was to integrate upstream systems which were already deployed, as well as third parties, without complex workarounds. It soon became clear to Corestate that neither an SAP-integrated solution nor an all-in-one solution fitted the ticket.
“We decided against an SAP-integrated solution from the outset because we work with four different ERP solutions and committing to SAP would have excluded the others,” says Wriedt summarising the decisive argument. For similar reasons, they also opted not to go for an all-in-one solution. While one of these might be a good option in the long term, it would have required very carefully conceived and therefore time-consuming specialist concepts during the preparation phase in order to implement all of the requirements in terms of cash flows, forecasting, reporting and credit management in order for the various areas of treasury to be fully integrated. Corestate wanted fast, reliable results, and that’s why they decided to rely on experts in the various areas of treasury.
The treasury information platform TIP is responsible for reporting and cash flow forecasting and is closely integrated with data captured by TIS. Both systems have ‘known’ each other for several years and are closely aligned. The TIS interface allows Corestate to access its group-wide financial status and cash flows in its forecasting at the click of a button in TIP. Besides this, daily ECB exchange rates can be imported and any cash flows which haven’t been captured in upstream systems can easily be imported via Excel.
Cloud-based or on-premise solution? Wriedt explains why the cloud-based solutions offered by TIS and TIPCO were important for Corestate: “In addition to issues such as data protection and data security, a major advantage is the fact that a cloud-based solution means that we don’t need to reserve any in-house resources for support, administration or the like.” His assessment is one which many of our clients share and which has been proven in practice many times.
TIP went live at the end of 2019 and has since been used intensively. A short-term cash flow forecast had already been planned even before the COVID-19 outbreak and was quietly implemented at the beginning of the year.
“From our point of view and in terms of what we need, we opted for the best providers you can get for these three areas of treasury.” A great compliment and one that we are delighted about and happy to share with TIS and our Austrian colleagues at LANG Finanzsoftware.
The full version of the presentation on YouTube:
Most of our clients have been reliably forecasting their cash flows with TIP for years. In stable economic times, all you need for this is long-term forecasting horizons; but in time such as the COVID-19 pandemic, there needs to be a rethink. In this context, increasing numbers of clients are supplementing their previous forecasting with a short-term cash flow forecast covering 14 days and/or 13 weeks.
The volatile economic situation at present leaves little room for manoeuvre and, at the outset of our first such project, it was soon clear that a solution was needed ASAP, and one delivered from home office to home office. Since 16 March, Austria has been in lockdown and we of course have also sent our staff home. The question of how such an ambitious project could work without face-to-face meetings was one we had to first clarify by means of a dry run. The outcome of this was clear: it hardly makes any difference. The only thing that several colleagues missed was the office cookie jar.
But why did it work so well? TIPCO supports over 130 corporates worldwide, the majority of which are based in neighbouring Germany, but also in the UK and even in the UAE there are also group treasurers who enjoy the benefits of TIP on a daily basis. Online meetings with our clients are the rule rather than the exception, which is why we are so well versed here. On top, there are also perfectly configured hosting environments which we not only use internally but which are likewise offered to our clients.
As a result, the first home-office implementation of a short-term cash flow forecast went like clockwork. The parameters were quickly and efficiently defined with the client in an online kick-off meeting, after which a forecasting grid was prepared and immediately made available to the client’s various subsidiaries. These subsidiaries, based all around the world, were able to enter their forecast data just 24 hours later, which the majority actually did, and therefore automatically reported these to the parent company. Even the group treasurer was surprised by the high degree of compliance.
The current circumstances present major challenges for our clients. The situation is serious and, even if they don’t wear their business outfits when working from their home offices, they are nonetheless fully committed to and concentrated on the task of safeguarding their subsidiaries’ liquidity in the coming weeks and months. That is why we fully understand our client’s decision not to be identified by name in this case study. We are delighted with the success of this project but nonetheless are aware that our software is mainly being used at the moment to manage the ongoing emergency.
We would also like to say a big “Thank You” to all those colleagues who cooperated so well from their home offices. We are convinced that there are many other companies out there that we can help to expand their cash flow forecasting, and working from home offices isn’t going to get in the way. In fact, we are already working on other implementation projects.
If you have any questions on this, please get in touch, we’ll be happy to help.
This webinar shows you how cash flow forecasting can learn from your historical data, rely on existing data in your systems and combine both to deliver dependable cash flow forecasts. We will also show you how user friendliness and immediate responses on the quality of cash flow forecasts can win over your subsidiaries. This not only improves the quality of forecasting but also your communication with subsidiaries.
Easy access to current market data is a core requirement of any Treasury department. Incomplete data sources and inadequate automation cost time and lead to errors.
Good news: In cooperation with our partner vwd Vereinigte Wirtschaftsdienste GmbH, from now on, we offer our clients a reliable and cost-efficient integration of current market data in TIP. No matter if you require current rates for cash visibility, cash flow forecasting or risk management – our cooperation with vwd enables seamless access to the market data you need at highly competitive prices.
Your advantages at a glance:
Infront and vwd have joined forces as of July 2019 to become a leading European full-service provider of financial market solutions. Infront offers a powerful combination of global market data, news, analytics and trading tools. vwd provides financial information and IT solutions for the investment industry. Together the two companies are a strong partner for our customers, with a comprehensive and holistic solution offering that is unique in Europe. Infront and vwd product coverage includes data and feed solutions, real-time financial market data solutions, news, analytics and trading platform solutions, solutions for portfolio management and advisory, regulatory compliance as well as publication and distribution solutions.
At TIPCO, we are experts in Treasury digitalisation. Our web-based Treasury Information Platform TIP leverages data from various source systems, supplements these with information provided by local subsidiaries and offers a comprehensive toolbox for data analysis and reporting. By combining IT-skills with many years of treasury expertise, we have made TIP the solution of choice for some of the leading companies in Europe across various industries. More than 120 clients trust in TIP and in our ability to provide market-leading treasury innovation.
We attach particular importance to teamwork because we believe that an excellent working environment is the key to excellent work. We are proud of our employees as every one of them is the basis for the satisfaction of our clients and the success of TIP. We encourage further education and training and we are willing to give professional newcomers and lateral entrants a chance to discover the world at the interface of treasury and technology.
Please, note: Our job ads are in German as you need to speak our corporate language fluently.
Find more TIPCO jobs at karriere.at.
FIRDAY 15 NOVEMBER
13:00 – 13:30
Room: BOSTON 11
Speaker: Dr. Mark Blatt | Dräger
We are proud to present one of our latest projects together with Dr. Mark Blatt, Strategic Projects
Finance & Controlling at Dräger. Visit our product demo at the DACT at 13:00 and meet our team at the TIPCO stand.
Everybody does it. FX hedging today is common practice for Group Treasurers. Target hedge ratios like ‘80% of USD’, ‘75% of CHF’ can be found all around. But how do you ensure that your hedge ratios and hedge portfolio match your risk appetite? How do you base your decisions on solid statistical and mathematical models and thorough analysis without breaking Excel or spending fortunes for change requests to your TMS vendor? How do you ensure that your hedging portfolio is optimal in terms of reducing your overall risk, optimal in terms of its’ cost or optimal in terms of its’ efficiency (e.g. minimum number of hedge deals)? Let us show you how!
In this product demo, together with Dr. Mark Blatt of Dräger, we will walk you through an FX risk management solution that provides you with concrete proposals for hedge portfolios that not only ensure you stay within a pre-defined target ‘at-risk’ limit but are also optimized for minimum cost, minimum deals or minimum hedged currencies.
The outcome: A best-in-class FX risk management solution generating fewer deals and less cost.
On November 15, Dr. Mark Blatt (Dräger) and Alexander Fleischmann (TIPCO) present what this looks like in practice:
Dräger manufactures medical and safety technology products. In so doing, the company protects, supports, and saves people’s lives around the world in hospitals, with fire departments, emergency services, authorities, and in mining as well as industry.
Founded in Lübeck in 1889, Dräger has grown into a worldwide, listed enterprise in its fifth generation as a family-run business. Dräger has more than 14,000 employees worldwide and is present in over 190 countries around the globe. The company generated revenues of around 2.6 billion euros in 2018.
About the Treasury-Information-Platform TIP
TIP empowers Treasury departments to digitalise their processes and do away with manual data capturing and endless e-mail exchanges. Whether the focus is on cash visibility, cash flow forecasting, risk management, guarantee management or bank fee analysis, TIP leverages data from various source systems, supplements these with information provided by local subsidiaries and offers a comprehensive toolbox for data analysis and reporting. This avoids the need to capture data manually, which is one of the most common causes of inaccurate data. No matter if requests for intercompany financing, guarantees or bank accounts: TIP workflows engage the right people and the right systems at the right time, resulting in a smarter and more efficient Treasury. The Austrian software provider TIPCO has made TIP the solution of choice for some of the leading companies in Europe across various industries. More than 130 clients trust in TIP and in TIPCO’s ability to provide market-leading treasury innovation.
On 7 November, we will be taking part in a table talk about “Predictive analytics in cash flow forecasting” at the 1st ACTA – TMS day in Vorarlberg. Predictive analytics is currently one of the most promising digitalisation issues in treasury. The objective is to massively reduce the manual work involved in cash flow forecasting, while at the same time maintaining or improving quality relative to manual forecasting. Find out exactly how predictive analytics works, which preconditions need to be met and how you can benefit from it.
On 27 November (Wednesday), we will be presenting four treasury issues on the TIPCO stand at the 15th Structured Finance event, each in just ten minutes. These short demos will be taking place during the breaks between speakers and there is no need to register in advance.
TIP offers smart solutions in the areas of cash flow forecasting, reporting, risk management, guarantees, derivatives, automatic bank fee monitoring and workflows. Regardless of whether you have already deployed a treasury management system or still prepare reports manually with Excel, TIP is the key to saving you valuable time and increasing the quality of your data.
Want to find out more? Take a look at our case studies to speedily discover how our prestigious clients already deploy the TIP treasury information platform.
We look forward to your visit to our stand (A27) in Stuttgart.
On 17 September, we will be taking part in two table talks at the Future of Cash Management in Mannheim. This event, organised by Schwabe, Ley & Greiner, is again taking place at Congress Centre the Rosengarten venue.
Predictive analytics is currently one of the most promising digitalisation issues in treasury. The clear objective is to massively reduce the manual work involved in cash flow forecasting, while at the same time maintaining or improving quality relative to manual forecasting. We invite you to attend this table talk with Hubert Rappold at 14:20. Find out exactly how predictive analytics works, which preconditions need to be met and how you can benefit from it.
Optimised workflows support you in your daily processes, add structure to your working and approval procedures, and inform the right people at the right time about what they need to do next. During this table talk, Martin Winter will be presenting a successful example based on guarantee management. We will show you how it works – from the internal request, via the electronic request, up until the guarantee lifecycle finally ends in the system.
Today, we would not only like to congratulate our four new partners but, above all, also ourselves. Why? Because Daniel Richter, Herbert Schager, Gerald Dorrer and Kurt Szenftner are a clear indication that we have been able to attract top new players to our partner ranks. Over recent years, TIPCO has been able to grow significantly not least due to their efforts. Around 50 employees in the meantime belong to our team. This is reason enough for us to wish Daniel, Herbert, Gerald and Kurt all the best and thank them warmly.