Six ways to remove human error from financial reports


Date: 19 May 2021

A child uses an abacus to solve a problem

Data can be the lifeblood of a business. But it can also feel like endless streams onto a sinking ship. The more data-driven your financial planning is, the more dependent you are on the quality of the data you have at your fingertips.

For CFOs to make the smartest and best-informed business decisions, it's critical to have numbers you can trust in your reports.

The issue is that reports, for the most part, are created by people. Even the most level-headed employee with the very best of intentions can still make a minor mistake. Whether your eyes are on an Excel formula or a Google Ad campaign report, removing human error should be a primary consideration for any CFO interested in getting more signal and less noise out of their data.

Here are six ways CFOs can minimise the inevitable mistakes people make when compiling and analysing reports.

1. Automate

All smart CFOs know this rule: if you're repeating something, you should automate it. It saves time, but it also drastically reduces the chance of human error.

The more times you copy-paste, the more likely your team is to make a mistake. The easiest way to remove human error from vital reports is to remove the human aspect. The more aspects of data collection and standardisation you can automate, the fewer human errors you'll have.

There are a few areas to look for opportunities in which to automate. First, look at data collection. Different sources spit out data in different formats and collating it into a meaningful spreadsheet can cause easy misunderstandings and errors. DataRails is a financial planning and analysis tool that automates repetitive work including consolidation and data aggregation by connecting to external sources easily, thereby making your reports more reliable and agile.

"Businesses typically spend between 10 to 14 days every month on manually gathering data from different sources and bringing it together to understand the current status of the organisation and try to predict future performance," DataRails co-founder and COO Eyal Cohen recently told VentureBeat.

CFOs can also automate report creation multiple reports are being created from similar data sources. If your employees are crafting reports from Excel, it's highly probable that the work they're doing can be streamlined and automated.

2. Alleviate the pressure

Stress can be a major cause of errors. Humans are, of course, human, making plenty of mistakes in the normal course of a day. This can be amplified when pressured or stressed. One scientific journal found that surgeons make 66% more mistakes in the operating room when they're stressed, which the study measured by collecting heart rate data.

If surgeons make potentially life-altering mistakes under pressure, it's clear that corporate employees certainly will as well. In this case, mistakes mean bad business decisions.

To reduce stress in the office, and hence reduce human error, foster an office atmosphere that is understanding, collaborative, and intentional. By reassuring employees that mistakes are normal, you can create a more productive, genuinely stress-free environment in which to input data.

3. Clean your data

Automation is a great help and time-saver, but if your data isn't reliable in the first place, it will do more harm than good. If employees think the data can be trusted, they're more likely to make a mistake if the data is incorrect.

Cleaning data before aggregation is a vital step in gaining valuable, actionable insights. For example, imagine a duplicate row is accidently included in an automated report. An employee might see that doubled data point and think it indicates high significance, worth highlighting.

There are a few easy checks you can implement on any data being collected prior to aggregation. Singapore Management University library analyst Aaron Tay outlines eight key steps to clean data, such as deleting extra spaces, standardising use of caps and removing duplicates.

If you prefer to outsource this kind of work, there are plenty of enterprise solutions who can complete this step. For instance, Xplenty is a data pipeline tool that includes a data cleaning layer in the ETL process.

4. Leave some data streams behind

A common mistake many data-rich companies make is trying to work with all potentially relevant data sources. However, by including more data streams to clean, collect, and analyse, it introduces more opportunities for mistakes and misunderstandings.

Instead of going the "just-in-case" route and collecting any potential sources that might come in useful, be frugal with data collection.

Start with a set of questions that need to be answered, like "Will we surpass our budget? By how much?" and work backwards to determine which data sources are vital to that answer.

More data doesn't necessarily mean more useful information. Data is only valuable so long as it is relevant. By adding more unnecessary sources, you only run the risk of including more human errors in your reports at little gain.

5. Invest in basic Excel training

Every prospective applicant nowadays will include "proficient in Microsoft Excel" on their CV. That doesn't necessarily mean they're trained in how to avoid the most basic mistakes that it's possible to make in Excel.

In line with point two above, it's best practice to avoid creating an atmosphere of shame. Mistakes happen – it's your responsibility to ensure you're minimising them. Instead, invest in training employees in basic Excel competencies.

One number stored as text can cause drama and mistakes - make sure employees know what mistakes to watch out for.

The first step is to ensure employees know they can admit they made a mistake instead of trying to hide the mistake. The second is to encourage employees to take training, like STL, a London-based company that specialises in training employees to minimise human error. Finally, back to point one, try to automate wherever you can.

6. Employ common-sense checks

Even after taking all possible measures, it's likely there will still be errors in your reports. Making mistakes is simply human. That's why the final step in reducing human error is catching it after it happened.

One recent example is when the NHS called one man with no underlying health conditions to get his COVID vaccine very early. The reason was that the data indicated he was 6.2 cm tall, giving him an incredibly high body mass index (BMI) of 28,000. A few simple checks could have prevented this error. For example, instituting a minimum height allowed for data entry, or a cap on BMI possibilities.

There are a few common-sense checks you can implement on any final reports. For example, if you're examining budget reports, ensure that all the numbers fall within the typical range, with conservative estimates. In the example above, even limiting to extreme ranges like at least four feet of height, and a max BMI of 200 would have been enough to catch that.

Final thoughts on reducing human error

Ultimately, the first step in reducing human error is by understanding it's impossible to eliminate them completely. However, by implementing the steps above it's possible to reduce the risk of a small mistake causing major errors down the line.

At every step of the way, from data collection to making decisions on reports, it's completely possible to mitigate risks and ensure you're acting on the best information available.

Copyright 2021. Article made possible by Jeff Broth, business writer and advisor. Jeff has consulted for SMB owners and entrepreneurs for seven years mainly covering finance, stocks and emerging fintech trends.

What does the * mean?

If a link has a * this means it is an affiliate link. To find out more, see our FAQs.