• Matija Nakić

How We Are Taking Excel Down From its Financial Modeling Throne

Updated: 7 days ago


The Golden Girls Cover - With old planning tools name over them - Excel, Oracle Hyperion, IBM Cognos, SAP BPC

Believe it or not, Oracle Hyperion was founded in 1981.

💡
Farseer is a low-code SaaS for financial modeling, planning, and analytics. Our vision is to become the go-to software for financial modeling and planning, the go-to financial planning software for companies overgrowing spreadsheets. In this post, you can read more about our journey.

The Golden girls


For the last 35 years, Excel is the synonym for financial modeling and planning. Companies of all sizes have turned this personal productivity tool into financial planning and reporting "software" with tens or even hundreds of users handling tasks such as budgeting, cash flow projections, or consolidation.

While CRM, project management, and accounting have migrated to modern cloud applications, financial planning is done the same way it was done 30 years ago. Despite the plethora of FP&A (Financial Planning & Analytics) solutions, over 80% of companies still rely exclusively on spreadsheets for financial planning 🤯

Oracle, SAP, and IBM created the first FP&A solutions and today hold over 50% of this 5 billion dollar market. Believe it or not, one of our Golden girls - Oracle Hyperion, was founded in 1981. predating even Excel 👵🏻

Implementing these legacy solutions requires a small army of consultants, costing an arm and a leg. The result is often a clunky and user-hostile solution, which makes users swiftly regress to their old Excel habits. Both SAP and Oracle receive over 50% of their revenue from support, proving that user autonomy is not high on their priority list.



The second generation of planning solutions


The second generation of cloud-based FP&A solutions emerged in the 2000s led by Anaplan, Adaptive Insights, and One Stream. Their double-digit revenue growth is steady for the past few years and is fuelled mostly by large international companies migrating from Excel or legacy FP&A solutions. As robust and feature-packed solutions these solutions are, they inherited a similar approach and price-tag from the previous generation. The business model of the second generation involves a broad network of partners charging hefty project fees to customize the software according to the client's financial model and processes. The power is there, but the user autonomy stays elusive.



Farseer


The Farseer journey started a few years back intending to solve our planning problems in a telecom company. Approximately 70 people were involved in the planning process. It was done in humongous Excel documents and managed by email and sporadic meetings next to the water cooler. The simple program from the telecom era grew into a solid MVP during the next few years and in 2019. we started with customer development and startup programs, winning a few major pitch competitions during that incubation year. The message from the market and the startup community was clear. Nobody has managed to create a powerful planning tool that is simple to use, and we would love to see you go for it.

So in September 2020. together with a few amazing people working on the project part-time, we took a leap of faith and started pursuing our vision of becoming the go-to financial planning software for companies overgrowing spreadsheets.



Farseer vs. Excel - the biblical fight


The main driver behind leaving well-paid jobs and own successful companies to plunge into this David and Goliath battle was a challenge big enough to fill the room.

⁉️
We asked ourselves: How is it possible that the entire business world runs on flat-file database sheets stitched together by cryptic formulas and macros?
gif

Excel is ubiquitous, familiar, and hard-wired into the minds of finance folks around the globe - making strong corporate inertia its ultimate advantage. Going against the status quo is extremely hard, but at the same time, fuels our mission with burning defiance. Excel is the best nemesis we could have wished for since it's making us work relentlessly on our 10x features that will break the spell of inertia.



Financial modeling as our main competitive advantage


Beating Excel in a company-wide, bottom-up budgeting process is not such a difficult task, since Excel was not built for collaboration. Farseer and all other relevant FP&A platforms support user roles and permissions, advanced cell auditing, and managing the planning process.

Financial modeling is creating an abstract representation of a financial problem, usually to forecast the business' financial future. In the world of finance, Excel has made its fortune and fame through the ability to model just about anything. Ironically, a lot of Excel's problems lie exactly in that bush. You probably encountered the Excel horror stories in which Fortune 500 companies endured million or even billion-dollar damages caused by a faulty Excel model.

Modeling in Excel does allow just about anything, but it's error-prone, highly manual, and has a steep learning curve. This is only natural since the ultimate trade-off in software is power vs. ease of use. When users have power, they can also do stupid things.

So the thetical question of our battle is: Can we make software that gives the power to the user but minimizes the manual work and the chance of error that goes with it?

Here is a list of concrete ways we are beating Excel in financial modeling and a few areas we have to improve.



How we are beating Excel in financial modeling


Our models run on serious technology


💾 Excel has a 1 million row limitation but starts breaking and becoming very slow as you reach the hundred thousand numbers. Some of our clients had Excel financial models measured in gigabytes, and opening them was a good opportunity to make a cup of coffee before the data loads.

✈️ Farseer is based on leading-edge web technologies and stored on Google cloud. Our proprietary graph calculation engine enables parallelized calculations and querying, which makes any calculation or simulation last under a second, even with millions of data rows.



Our models are human ready


💾 In Excel, numbers are just cell coordinates making formulas a cryptic sausage of numbers and letters. Besides the fact that it's easy to make a mistake, anyone can tamper with formulas.

✈️ Farseer has a spreadsheet with an underlying database, meaning every variable has its name.

Our formulas look like this:

revenue = quantity * price 

Users can write over the formula value, but they can never delete it or tamper with it. Users can also create their mini-models and use formulas, with the integrity of the main financial model always protected.



Our modeling speed is 20x


💾 Making changes to a financial model in Excel, whether adding a variable, a version, or a year is manual and tedious. Creating and maintaining a model that can involve hundreds of accounts and dimensions (markets, product categories, products, product price, product quantity, product revenue...) is a proper nightmare.

✈️ Version and timeline management are inherent to Farseer's interface. Administrators can add or retire timelines and versions with a single click.

But the thing that really brings modeling speed and accuracy to new levels is template-based modeling, or modeling "in bulk". Accounts and dimensions are entered or imported in Farseer once and can be replicated multiple times without adding them one by one. To make modeling even faster and completely error-free, users can define universal formulas for those accounts. Very complex models containing hundreds of accounts and dimensions can be added to Farseer with a few clicks.



Our model is top-down AND bottom-up


💾 A few Excel planning nightmares match the manual struggle of top-down planning when having multiple different dimensions involved. Scenarios like these in Excel translate to weeks spent on the manual tuning of numbers so that management-based revenue goals are appropriately broken down to lower levels such as product categories, SKU's, distribution centers, etc.

✈️ In Farseer, numbers entered as top goals are propagated to the lowest levels of the model hierarchy in a matter of seconds. You can enter the revenue goal and get a first draft plan for quantities in all the markets, product groups, and SKU's. Changes on the lower levels (bottom-up) impact all the other values while always maintaining the top-down goal. This shortens the planning process drastically and makes the plans better aligned across the company.



Our model knows the future


💾 People are bad at predicting time series. Which is basically what planning is. Predicting utility costs is easy, but predicting the quantity of a particular SKU sold at a particular shop is not. In Excel, you can rely on regression and you can not even do that on a bigger scale.

✈️ Farseer's spreadsheet runs on machine learning algorithms that become more accurate as you fill the model with actuals. You can ask Farseer to do a forecast for a single time series or the entire model. The algorithm will honor seasonality and any other specifics from the historical data. Planners can decide what historic data they want to use for the automatic forecast, and eliminate outliers to get a more accurate forecast.

Our model supports super-fast what-if scenarios


💾 What-if scenarios and sensitivity analysis lately gained traction even among the planning skeptics. In Excel, creating a what-if scenario that includes a single driver translates to a few hours of redoing the model. Involving more of them and testing their impact on P&L, balance sheet or cash flow requires a few all-nighters from our modelers.

✈️ Since Farseer is a driver-based planning software, any driver from the model can be pulled to the dashboard. Users can change one or multiple drivers (margin, average salary, discount rates, etc.) and see a visual and numerical impact on the relevant reports such as P&L, product profitability, or cash flow. Scenarios do not impact the plan version, but if you are happy with your scenario you can copy it to the plan version with a single click.



Farseer's room for improvement


Incumbents are incumbents for a reason. During the 37-year tenure, Excel has amassed a lot of great features. A great deal of them we will never support. Here are a few crucial advantages of Excel over Farseer when it comes to financial modeling.


Pivot improvements


We do support pivoting but granted, our UX is not super intuitive at this moment. Excel is still the pivot king. We have a great opportunity to create the next generation pivot which will be contextual and will allow automatic forecasting and propagating numbers back to the underlying cells. These are all elements Farseer already supports.


Search and filtering


We provide a basic search. But we need improvements in the overall accessibility of the search feature as well as adding advanced filtering functions. Excel has some great and handy tools in this domain, such as nested sorting, find and replace, and we need to catch up.


A higher modeling freedom


To get 20x faster modeling compared with Excel, we had to make some concessions in terms of the average user's modeling freedom. In Excel, anyone with the most basic knowledge can get a quick calculation. In the next major version, we will allow greater modeling freedom while still protecting the structure of the master financial model.


To conclude...


Fun times are ahead of us. Companies like Airtable, Coda, and Notion have already proven that spreadsheets are not untouchable and that a clever vertical or even horizontal solution can snatch hundreds of thousands of users from the famous gridded sheets.

We know where we are going, and the winds are favorable. To grow into such a larger-than-life vision, we will need an amazing and motivated team, hard work, and some good fortune. If you want to help us create a powerful planning solution that is extremely simple to use, get in touch.

1,069 views0 comments

Recent Posts

See All