Achieving operational excellence

May 26, 2020

Bernat Fages

The key to achieving optimal performance in any operation is to accept no process is perfect from the start. It always takes time to perfect.

When a new Ops process is established, it is often done in a waterfall kind of manner: objectives are set, guidelines are defined, tooling gets built, staff receives training, and metrics are established.

The assumption many managers make is that once all the different pieces of the process are in place their job is done. This mindset tends to lead to mediocre results because it turns a blind eye on the many opportunities for improvement that appear as you progress.

Those in charge of operational processes who want to get the best out of their teams need to acquire an iterative mindset of constant change to drive improvement.

It's all about iterating

The idea of operating in terms of cycles has existed for a long time. The first acknowledged model in modern times to propose iterative functioning was the OODA loop. OODA was famously created by Colonel John Boyd, and used by USAF fighter pilots, to secure a military advantage in war.

OODA stands for Observe, Orient, Decide and Act. It describes a cyclic mindset where one collects information, processes it, makes a decision and executes it. As war unfolds, certain events happen and more information becomes available, one faces the need to constantly reevaluate their strategy. And doing so with agility is key. The OODA loop facilitates this behaviour.

Since then, the OODA loop has made its way into other fields such as Business. When I worked at Facebook, I had the opportunity to learn about the company's core decision-making framework for building products: UIE, or Understand-Identify-Execute. It has percolated through its culture, internally taking an almost legendary status. The UIE framework builds upon the same principles that constitute OODA, and is also meant to be applied iteratively.

OODA for Operations processes

I believe the principles of the OODA loop apply to any Operations process. These should be founded on the idea of understanding the functioning of the process to find gaps, implement fixes to those and operate the improved process to yet again find other gaps, and so on.

Our framework is based on the idea of Understand, Improve and Operate.


Understand refers to any activity that improves your knowledge and awareness of the process. The goal of this stage is to uncover issues, gaps, or just find ideas on what could be done better.

For instance, knowing what your KPIs are, tracking them, seeing when they're behaving abnormally, finding who in your team is underperforming or which processes are in a worse state – in terms of quality or efficiency – are all things worth doing. The more information you can evaluate, the less likely you are to miss out on what matters most. You must ensure you understand where your most meaningful lever is at any time.

This might imply measuring things like your queues' intake, your team's throughput, AHT or TAT, auditing your team's work, or analyzing how each team member's accuracy is progressing over time. Whatever it is, the end result of this stage should be an actionable insight that will drive improvement.


Improve encapsulates any initiative with the goal of mitigating or alleviating an issue, or perhaps further enhancing an already positive aspect of an operational process.

This could mean including more information in the process to reduce mistakes, remove redundant data to drive efficiency, stand up a QA process to gain awareness on accuracy, flag mistakes to an over-erroneous rep, improving your team's tooling, automating one step of the process or running a re-training for the team or adding a notifications system to improve TAT.

In essence, Improve consists in taking action upon an insight uncovered in the Understand phase.


Operate means to run your improved process so you can gather enough new data to move back to the Understand stage.

Each improvement made needs to be run to generate and collect data that can then be evaluated to assess what's the next priority.

Implementing the loop

Implementing this loop can often require a lot of work from different stakeholders to enable analytical capabilities, measurement, tooling changes, re-trainings... The resulting dependencies and need for coordination can generate a lot of drag thus hurting your ability to iterate with agility.

This is a huge pain point we have suffered ourselves and many other Ops teams I have met throughout my career, which is our product's raison d'être. With Human Lambdas, we give you all the tools you need to Understand, Improve and Operate any manual process so you can iterate faster. As a consequence, you should be able to drive improvements rapidly and at a much lower cost.