Should I automate this task?

Author: Benjamin Saladin
​  
​We all have probably been in discussions about automation of repetitive tasks and very likely the outcome was: “This seems too expensive to automate.” But this decision making process is most likely based on experience and gut feeling. Don’t get me wrong, this can be a valid approach if there is no underlying data or models that would help give you a better outcome on your ROI (return on investment).

On the other side, automation is often a valuable approach. Using the dashboard below, you can check out the impact on ROI and the break-even date from which your investment starts to pay off. The tool helps you to evaluate how much time and money you could save by automating repetitive tasks. Give it a try - automating will be worth it, believe me.

​But first a small anecdote from my past experience as a finance manager in a smaller company where I had the opportunity to not only take care of finance, but also IT, HR and in analytical matters in other departments. This was a great opportunity for me to grow my skills not only for personal reasons but also to be able to constantly deliver an adequate level of quality.

When I started this job I had 15 days in a normal month to report sales and financial information to the head quarter, of which the majority of the information was requested by the 5th workday of the month. The workflow looked as following:

1. Download and print out the first sales analysis report (revenue, quantity and average price). In more technical terms, download csv from wholesaler in mail to filestore, import csv with the correct encoding into excel, create sums and averages in the required context. Of course these are very low-hanging fruits to improve upon, but this was the reality I was facing.
2. To report this to the head quarter we had to enter these numbers into our ERP (enterprise resource planning) system. This had to be done manually and average prices to 4 digits behind the comma had to be entered. Error-prone? You bet. These and additional tasks took me at least two working days, 16 hours of my time.

At that time, I had experience with Excel but did not know anything about the tools Power Query,  PowerPivot and the concept of business intelligence. This changed rapidly when started to capture the possibilities automation. The end result was that these 16 hours of reporting tasks were cut down to 2 hours. In retrospect, I would say that it would take a maximum half a day for implementation. At that time, of course, it was more as I was learning to use the tool. And of course, this is why analytics consultants are more costly.

What was the value for the business? In IT investment projects we took 5 years to write the investment off, so this is the timeline we are going to use in our automation dashboard and we assume a rate of CHF 50 for my hourly salary. As I was both the employee doing the task and implementing the automation the cost of automation would not over-proportionally impact the net benefit. Also, for simplicity reasons we disregarded DCF (discounted cash flow-method).

Imagine the impact not only on cost and time savings, but even more the opportunities automation can create for your company to reallocate resources. Moreover, you could grow faster with more efficient and data-driven decision making. The following three aspects are considered most important when it comes to the question why automation makes so much sense in most organisations:
​

• Time to market: imagine a head quarter that can ingest all relevant sales data on a daily basis. For example, consider product campaigns and compare them across countries for optimisation purposes. Imagine these campaigns are even online: A data-driven feedback loop could adapt parameters in the campaign setup to get the best return on marketing investments.
• Save money: Human labour is an expensive good that needs to be allocated with care and based on well-balanced considerations. Hence, we want our employees to give proactive recommendations and creative ideas rather than letting this person do what a machine could do a thousand times better and cheaper.
• Cut off inefficiencies: Ingestion services like Fivetran and the use of APIs (application programming interfaces) will help you to connect different data sources in a centralised EDW (enterprise data warehouse). You will no longer need manual searches and downloads, much less repetitive tasks and hence, be exposed to inefficient labour allocation.

To answer the question in the headline, the short answer is yes. The long answer is, if the task is repetitive, prone to errors, annoying, not too complex to automate and is not based on human considerations: Go for it!