Despite the tremendous strides in automation in recent years, it’s surprising how many asset management firms still have pockets of spreadsheet dependency in their operations. Like many unhealthy habits, it starts innocently enough when someone needs immediate answers, and solving that need through normal IT channels takes too long. Then, before you know it, the organisation is hooked – to the point that spreadsheets become a crucial component of your processing workflow, embedded into the firm’s daily deliverables. 

The effect on efficiency

The problem is not with spreadsheet software as a tool but its overuse and integration into the daily business process. Moving data outside of core systems and into a spreadsheet introduces a manual step that inherently slows down the data process flow. Moreover, it puts the end-user in the middle of the process rather than the technology, thereby reducing efficiency. Spreadsheet use also creates the potential for data manipulation or “fat finger” mistakes. Errors in data will inevitably create downstream issues that require time to correct, resulting in further lost efficiency. 

Fortunately, today’s advanced automation technologies – including but not limited to artificial intelligence and its variants – have the potential to drive dramatic efficiency gains. Moreover, new tools and processes can also confer flexibility and transparency that eliminates the need – or the temptation – to extract and isolate data from the normal process flow. 

Leveraging more modern technology solutions, however, should be combined with rethinking existing processes. Unfortunately, many firms mistakenly believe they can graft artificial intelligence solutions into their existing processes and systems to achieve incremental efficiency. This fix is like putting a turbocharger on your 1982 car. The car will go slightly faster, but you will have the same steering, suspension and interior as before not greatly improving the overall driving experience. Instead, the opportunity to undertake a process reengineering approach that uncovers smarter, leaner processing that fits with new intelligent technologies will produce the most dramatic results. 

Five good reasons to revisit your spreadsheet usage 

Several benefits will occur from increasing automation, leveraging intelligent tools and eliminating offline processes using desktop software. First, however, let’s focus on the top five:

1. Optimising the data processing workflow. The most efficient enterprise data processing approach is an automated workflow moving data along its designated process steps. The data is processed by the appropriate service layer and moved to the following processing component. Automation reduces the “wait state” between manual business process steps when using spreadsheet tools. 
2. Maintaining data integrity. As noted earlier, once data is out of the core processing application into a spreadsheet, the potential for data manipulation can result in an incorrect change of data and inconsistency with the source system. Users may think they can compensate by validating the data as it flows back into the core processing system. However, instead of building programmatic data checks, why not stay within the core processing system where the controls are consistent and data integrity can be maintained?
3. Preserving historical data accuracy. A spreadsheet program is not a database. Therefore, moving data into a spreadsheet constrains the ability to keep historical records of changes. Implementing an automated solution to do its work uninterrupted will ensure that changes are accurately recorded across all dependent systems and enable reliable audit trails. 
4. Optimising the value of the data. The ability to implement artificial intelligence algorithms is key to driving process efficiency. To be effective, AI needs to consume data. However, data manipulated in a spreadsheet is generally not accessible for use in AI programs. Leveraging data directly from a database (and avoiding pulling it offline in a spreadsheet) will enable you to take advantage of AI more effectively and consistently in your operations. 
5. Avoiding “Key Person Risk” with proper change control. At times there is a need to change the calculations or data from a system, usually due to a business change. If an organisation has multiple spreadsheets running production processes, the enduser who controls each separate spreadsheet needs to be the one who makes changes, which introduces “Key Person Risk” if that individual is unavailable or leaves the organisation. Maintaining data within an automated solution makes centralised changes easier and is not dependent on one person’s knowledge.

Take the spreadsheet challenge

The key is to step back and take this challenge for firms concerned that spreadsheets may have become embedded too deeply into their production processing.

• First, analyse the usage of the spreadsheet program in daily processing and assess the potential inefficiencies that have come from its usage. Start by breaking down the process steps within an operation (many organisations use Value Stream Mapping, a lean-process methodology for documenting the steps in a process¹). 
• Next, inventory the spreadsheets used within each process step. Then measure accurately the total time required outside of the core processing system to open, maintain, process, back up and move data in and out of the spreadsheet environment. 
• Finally, compare the overall duration of processing a spreadsheet in the business process to processing within the core systems, accounting for artificial intelligence to increase overall processing efficiency. 

We believe you will find that highly automated processes will consistently outperform a process involving human intervention and data manipulation. Advanced artificial intelligence technologies in conjunction with re-engineering existing business processes make it possible to implement a more efficient end-to-end service model.

¹ “What is Value Stream Mapping?” American Society for Quality (ASQ), 2021, https://asq.org/quality-resources/lean/valuestream-mapping