The concept of queues for operational teams

The concept of queues for operational teams

In my experience implementing queue systems across various teams in three different companies, I’ve realized that while the concept might seem repetitive, each implementation journey is unique—especially when you’re starting from scratch.

Queues often become necessary when you’re trying to meet SLAs (Service Level Agreements) and ensure tasks are completed within specific deadlines. When you don’t have the luxury of hiring more agents, a well-structured queue system can be a lifesaver.

What are queues, really?

Think of a queue as a tunnel with several exits. There’s a mechanism directing tasks to the appropriate exit based on a set of rules, and behind each exit, there’s someone ready to handle the task. It’s like a bus stopping at a station: passengers get on, sit, and then get off at their designated stop. The goal is to keep things moving without anyone getting lost or left behind.

In product management, a queue system is all about managing workflows by prioritizing and directing tasks based on what resources you have. It’s about getting right processes, reducing bottlenecks, and making sure tasks get handled by the right people—or machines—so everything runs smoothly and efficiently.

You could compare queues to jira kanban or sprint boards, and even though it is similar, the rules and functions on how tasks get to that board and are picked up are not the same.

Balancing demand and supply

Demand is the work that needs to get done: customer requests, compliance checks, or other operational tasks. Supply, on the other hand, is the team, tools, and any other resources you have to meet that demand. Balancing these two is where the real challenge lies.

For example, in a KYC team at Company XYZ, the demand is the volume of customers needing identity verification. The supply includes the internal tools, third-party suppliers, and the people managing the process. In an accounting team at Company ABCD, demand might be requests from customers who need to start electronic invoicing, and supply is the team of accountants, the tools they use, and the rules that guide their work.

The "remove, outsource, automate" approach

One of the best ways I’ve found to optimize these processes is using a framework called "remove, outsource, automate." I didn’t come up with the name—it actually started as a meme from one of my favorite bosses—but I’ve pushed the concept forward and adapted it into my work.

1. Remove: First, ask if each task is even necessary. Why does it exist? Can you get rid of it or at least deprioritize it? For instance, our accounting team used to process people who hadn’t paid for their subscription. We stopped doing that. In KYC, we stopped prioritizing users who didn’t have a payment when ID verification wasn’t crucial.
2. Outsource: Look at whether you can delegate tasks to a third party—or even within your own team. It’s about deciding if it makes sense to use a human or a service, and understanding the trade-offs.
3. Automate: Finally, see what can truly be automated. It doesn’t have to be high-tech; sometimes, a hard-coded solution from an engineer can speed up verification processes or submissions. The key is aligning the level of automation with your deadlines and the complexity of the task.

Getting a grip on supply and demand

To quantify and manage demand effectively, you need to really dig into the work your team does—every single task, no matter how minor. If you’re dealing with accounting, get to know the nuances of every customer type. For KYC, step into the agent’s shoes and understand every part of the process. Observations like side-by-sides are a start, but they’re not enough. The more hands-on you get, the easier it will be to optimize later on.

If data isn’t readily available, you’ll need to create it. Use tools like Google Scripts to connect data flows from spreadsheets, or link them to platforms like Zoho to visualize how tasks and demand look. But even if you have data, don’t just rely on dashboards like Looker. You’ll need to dig deeper into the schemas and tables your company uses. Learn the logic behind them, read the code, and get a sense of how things really work under the hood.

On the supply side, it’s crucial to understand the resources available to you—internal teams, tools, and third-party providers. Read the contracts with your third-party providers. Check their terms, learn their integrations, and find out where their work transitions from automated to manual processes. Identify failure points and the teams involved. Know their capacity and availability. And don’t forget the internal side: understand the size of your teams, their roles, and even their salaries if it’s relevant. This helps you gauge cost implications and prioritize tasks based on your current goals, whether that’s saving money, speeding things up, or just managing the workload more effectively.

So, when do queues come in?

Queues come into play once you’ve done all the groundwork: mapping out every task, understanding all your resources, and getting a clear picture of what needs to get done to deliver value to the customer. The moment you have this big picture, that’s when you start thinking about queues. It’s not just about handling the volume of tasks; it’s about visualizing the entire process from start to finish.

Map it all out—multiple times, if needed. Don’t worry about getting it perfect on the first go. Create diagrams, draw workflows, and try to see how tasks flow from entry to exit. Think of each task as a passenger on a bus: it enters at one door, sits for the ride (or the process), and exits at its designated stop. For instance, if your KYC task is to "verify an ID document," your entry point might be when the customer uploads their ID for the first time. The exit is when the document is verified. In between, there could be checks by providers, failures, customer follow-ups, and additional steps if something goes wrong.

By mapping out each step and putting numbers to it, you can clearly see where the bottlenecks are and where manual intervention is needed. That’s where queues come into play. Whenever a task needs to be handed off to a human because the automated or outsourced solutions can’t handle it, that’s your queue. It’s like building a funnel where everything flows through in an organized manner, guided by the rules you’ve set.

But it’s not just about creating a line for tasks to follow; it’s about designing a system that prioritizes efficiently, so the most critical tasks get the attention they need first. Imagine your queue as a sorting hat, deciding what goes where and who handles what based on urgency, availability, and the rules you’ve defined. It’s about creating flow and reducing friction, ensuring that nothing sits idle longer than necessary.

Oh, and by the way...

You can’t do this alone. Having a right-hand person on the operational and engineering team makes a world of difference. Those two people become your go-to for keeping things on track and making sure the day-to-day stuff gets done as planned.

They’re your eyes and ears on the ground and on the code (unless you are the one coding it), making sure the strategy you set actually plays out in real life. It’s almost impossible to juggle the big picture and all the little details without someone helping you connect the dots.

By truly understanding both your supply and demand, and thoughtfully implementing queue systems, you can align resources effectively, keep quality high, and optimize your operations—even without constantly adding more resources. So, map it out, set it up, and let your queues do the heavy lifting.