Automate or Not: ResearchOps Edition. When and Where It Makes Sense to Automate Your Job?

One of the main focuses for every ResearchOps is optimizing processes to work more efficiently and build a system that can scale. This includes the challenge of understanding which tasks we need to do manually and when we can turn the autopilot button on.

“To automate or not to automate - that’s the question,” Shakespeare would say.

Whether you’re committed to a full-time ResearchOps role or a Researcher taking care of the operation tasks, don’t miss these tips and tricks to figure out when and where it makes sense to automate your job to add value to your team’s work.

Before deciding if a task can be automated, I first recommend understanding the complete picture of a process and the variables involved, such as the tools in place, the people affected by the assignment, etc.

For me, a useful and easy way to visualize a process and its flows is to create a journey map and compile the actions and steps into a timeline. With this exercise, you can quickly identify the actors, the stages, the tools used, the pain points or delight moments, and the main improvement opportunities. Keep an eye on the repetitive tasks or those requiring manual work and not involving people because they’ll be the best candidates for automation.

I recently did this exercise to improve our B2B recruitment process. First, I listed all the steps and filled in the actions involved:

Then I completed the happy moments and the pain points:

Finally, I ended up adding the touchpoints or tools we use, questions that the stakeholders or I might have, and the opportunities I saw for improvements:

To fill the gaps that I was missing or didn’t have access to all the information, I ran some user interviews with them to add their insights to this journey map.

What was the result? I outlined a few things that needed to be fixed:

• Work on a solution to automate new research requests coming from our researchers and non-researchers to identify overlappings, prioritize the projects, and plan them in advance.

• Rethink the entire process with our Account Managers when it comes to approving users before contacting them for research to reduce the manual tasks, the number of tools involved, and the time and effort invested in this step.

• Look for a solution that can help me automate the incentives management.

Some of the things that I personally suggest automating are those low-impact and admin tasks. Anything that is not adding significant value to your team’s work (but it’s still necessary!) and requires a manual intervention should be computerized:

• Personal notifications: they’re helping to keep us updated with real-time changes. Especially if you’re the only one working as a ReseachOps (or if you’re a Researcher taking care of the Ops), notifications are a necessary evil that will greatly help keep track of things. But be mindful of having a healthy balance between being informed and spammed, as you probably don’t want to create too much noise in your inbox or messaging tool.

• Confirmations and reminders: those emails take a lot of time to be sent, and they’re only sharing information about something, for example, an upcoming user interview. They’re very much needed to inform our users and ensure they have all the details to join a call, but they don’t really add much value to your work. Explore the tools you use to set those emails to be sent automatically and save some time for other things. You can find some examples below.

• Newsletters: do you need to inform your stakeholders about anything related to your research? You can include any information or metrics that can be pulled automatically from other tools, such as the weekly user interviews or how much money you spend in a month on incentives. Plan your team’s communication and the tools you need to use and set this to run during the week and keep everyone updated without you even noticing it.

• Repetitive things: think of those admin tasks you perform every day or week, like pulling reports or sending invoices to the Finance team. Once again, explore your tools and how this can be done automatically without setting any reminders to do them.

• Connect tools without native integrations or APIs: although integrations are becoming more and more developed, you may find that some tools are not connected with each other. You can explore the automating tools to link them together if you need to perform some “behind the scenes” tasks (updating records, moving files from one tool to a shared space, etc.).

Three examples that I put into practice:

1. I created a private Slack channel to receive all types of notifications that I need to have quick visibility. Those include if a user interview is canceled or rescheduled (in case of any last-minute changes), RSVP reminders if I need to catch up with a colleague to kick off a new research piece, or a teammate’s birthday. They come from different tools we use (Calendly, Typeform, Google Calendar, BambooHR), and most are connected via Zapier.

2. To keep our Account Managers informed about the user interviews happening every week, I set an automation from our recruitment panel that automatically sends a Slack message every Monday morning with a summary of the users having a call with our Product team. With this, I keep everyone informed without taking my time to do it.

3. To share our research findings within our company, I created a Slack channel with a workflow where the owner of a research piece shares a summary of their key insights and research report, so everyone has access to it. This is integrated with Notion, automatically creating a new item in our research repository.

Although automation will solve a portion of your job, this doesn't mean it will dominate every part of your operational work. There will always be tasks that even the smartest tools can't handle. Jobs involving human interaction, strategic planning, decision-making, or subject matter expertise won't be replaced by machines anytime soon.

When it comes to the day-to-day tasks from a ResearchOps, in my experience, we can’t automatize:

• Things that require human interaction: and here, the list is long. We can go from supporting someone in the team when planning their research to running moderated user interviews or being the buddy for a new team member. Collaboration requires the skills of actual human beings, and they always will.

• Actions that depend on others’ behavior: while taking good care of your research budget, you don’t want to give away incentives to people that are no-shows to your user interviews, right? Yet again, there are things that are attached to others’ behavior that still require a little monitoring from your side and can’t be fully automated.

• Things that need to be tailored to a person (unless you can automatize this): be careful when automating emails to your users because some of them might need some adjustments, like changing the titles, the recipients, or particular details. Unless you can set placeholders (like {customer_name}) to customize an email, avoid creating impersonalized templates that make you sound like a robot.

• Requires human thinking: anything involving empathy, creativity, decision-making, or analysis can’t be computerized. Although automation can help your team simplify part of a process, core researcher skills are valuable for a reason and can’t be done by machines.