The NASPP Blog

Automation is essential for data management because of its efficiency and accuracy. Any time a person is tasked with manual data entry, there is a significant risk of error. That risk increases exponentially as the data volume increases. These are the three basic steps to building and executing a project plan to incorporate automation into your daily routines.

1. Identify


The first step towards automation is to identify the data sets that require regular manual manipulation as well as the databases or software systems where the information resides. Processes could be either data entry/transfer (e.g., entering employee demographic information) or the processing and interpretation of data (e.g., share usage projection). This is the time to really brainstorm. Every single process you do can go on your list; each process may include several steps that should be broken down into individual opportunities for automation. Take your list and create a spreadsheet with a list of each process, where the data originates, where it needs to be transferred to, and what departments or outside service providers are involved.

Take time at this step to get to know each of the database or software systems that your data touches because this knowledge will carry forward to each piece of automation you build. Drill down each system’s capabilities for formatting output of data and accepting data as an upload. If you are manipulating data for analysis, it may be possible to build data output reports that either reduces the manual processes for analyses or provides the data in a format that requires no further manipulation. When starting your automation project plan, you are only looking to understand the potential from each of your data systems; you don’t need to detail the entire business specification at this point.

2. Prioritize and Build

Once you have identified which data processes are candidates for automation, it’s time to prioritize. A good place to start is to eliminate the processes that require data output formats that your systems can’t accommodate at this time. These items can take a different route as topics to approach with the database architects, whether they are internal or external contributors. Rate the remaining processes on how easy each is to automate as well as how significant the process is. Significance is the more difficult to define because it could mean processes that impact the highest volume of data, represent the highest risk of error, or are components of the most visible data outputs. This helps when determining which pieces of automation to tackle first.

Once you’ve established your project priorities, build your automation. You don’t need to tackle projects one at a time. You may find that some pieces may build on each other because they use the same original data. Additionally, you may be able to oversee more than one automation project simultaneously if the process requires time commitments from separate groups.

3. Bust

The final step for automating a process is to break the automation. Well, at least try your hardest to break it. Take each piece of the automation and think of as many situations as you can where data will feed into or come out of that process, even if it’s not what the process was originally intended for. Don’t limit yourself to using correct data, either. It’s important to know what will happen if incorrect or incomplete data sets are used so that you can recognize these problems if or when they happen in the future. Remember, you may not be the only one using the process and it could be the platform for further automation in the future.

-Rachel