9NOVEMBER 2020work with data sources becomes a specialist skill. Only for the hardened spreadsheet whizz, but even then, there's hours and hours of work.I recently joined a finance department where we have a reporting, forecasting, and consolidation spreadsheet called "The Uber". It got its name because it's become an all-encompassing model that we cannot work without. Several versions per month, hundreds of versions built up over the years. Every finance department has an "Uber" even if it has a different name. And every Uber needs an Uber driver. A skilled Microsoft Excel expert who can follow links, formulae and look-ups.What about if instead of having an Uber spreadsheet, we had a tool which was designed to cope with the complex structure we're attempting to report. Which, while needing an experienced driver, was mostly self-driving? The tool that I've used for the last few years which fulfils this purpose is called LucaNet--a multi-dimensional financial database tool. Once the experienced driver has taught LucaNet to interpret the data from those complex general ledgers and foreign subsidiaries, it will take the heavy lifting of data manipulation.This leaves me, as the driver, free to interpret what the data is telling me about the business, but much sooner than if I'd taken an Uber. If you work in private equity as a CFO, like I do, then you're used to the demands of quality of earning (QofE) or due diligence work. It always starts with years of historical financials, going back to the general ledger, then into financials, and forecasts.When we have to track numbers from the general ledger into an Uber, through a complex web of error ridden spreadsheets, where taking a view of the data a different way takes time and is unreliable--what's the potential cost of that on the confidence in our numbers? And what's the impact of that loss of confidence on the exit multiple? Then there are the other systems and data sources besides the messy general ledger. How many times do we get an extract of non-financial data from our ERP and other systems and mash it together using spreadsheets?I know that I've been doing that for my entire career. I've been able to use my excel skills to gain some amazing insights into what both our financial and non-financial data is telling me about a business. But those excel models are dead on arrival. They are only ever going to be as current as the data sucked out into a CSV file from our systems. They are only as strong as the data set that they are built from which is usually limited by excel's capabilities to handle large volumes. And they are only as good as their driver.What if we were able to connect to a complete data source, teach our computer to replicate the logic we apply in our manipulation, and connect multiple data sources together. All of a sudden we have a really powerful view into our business, which can be used to make data driven decisions in real time.I've found that Microsoft Power (BI) delivers this ability. LucaNet is built to allow us to access the newly structured financial data using Power BI. Connect this to non-financial data using Power BI and we can tell you anything about a business, quickly, accurately, and repeatedly. The only restriction on our ability to answer any question about the business is when we don't have the data. If the data exists we'll be able to answer your question using Power BI.But progress obviously doesn't stop there. Next up for me will be using AI software to replicate tasks performed repeatedly by humans. Not so that we can replace humans with computers, but so that we can use humans for what they are designed for: interpreting information. If you're a spreadsheet expert and your business runs on a web of spreadsheets, you are not alone. But you may well be alone soon if you don't move forwards. I mean, who would dream of producing an extended trial balance using a pencil and paper nowadays? What if we were able to connect to a complete data source, teach our computer to replicate the logic we apply in our manipulation, and connect multiple data sources together.
< Page 8 | Page 10 >