R075 - Measuring Productivity, with Toni Tassani

Measuring productivity might seem like a clear answer to improving team efficiency, but it also raises many questions.
To what extent does the clarity provided by data truly drive progress?
When does it become a source of stress or an obstacle to creativity?
How far does measurement help us be more effective, and when might it limit a team's autonomy?
It's a delicate balance. Measuring can provide structure and transparency, but can it also distract teams from what truly matters? How much transparency is constructive, and at what point does it become invasive?
Today, with Toni Tassani, Deputy CTO at Ocado Technology, a global company born in the UK with a base in Barcelona, we will explore how to find the balance between numbers and the true impact of work.

R075 - Measuring Productivity, with Toni Tassani

What is the real world to you?

The real world is what you and I decide it is. It's an agreement between our perceptions. If I think about it alone, I might be biased. But when we dialogue, we create that reality together.

What is productivity?

In classical terms, it's the relationship between inputs and outputs in a system. But in intellectual work, especially in software, it's more complex. It's not just how much code you produce, but the quality, the ability to make changes, and sustainability. Sometimes, it's even said to be an illusion. Nowadays, we prefer to measure the development experience rather than productivity itself.

Is it more important to improve the engineer's context and experience than to focus on how much they produce?

Exactly. Quantity is still interesting, but we don't want that productivity to come with friction or burnout. The important thing when measuring productivity is: Why are you measuring it? We quickly realized that was the most important thing. For us, the important thing is to improve, to try to be able to produce more and in a more efficient and effective way, rather than just measuring it. Measuring is a vehicle for improvement, not an end in itself.

Why did you decide to measure productivity in engineering teams?

It wasn't a personal decision, but a natural evolution within the organization. Being a data-driven company, we thought about how we could use that information to improve. We created a group called Engineering Productivity to focus on how to facilitate engineers' work, not just measuring for the sake of measuring.

Each team has its own way of improving. The important thing is that they decide what they want to measure because you can't focus on everything. The key is in attention and continuous improvement.

What is Ocado Technology?

Ocado started over 20 years ago as an online supermarket without physical stores. The idea was ambitious: to automate warehouses and use artificial intelligence to make direct-to-consumer delivery viable. Now, we offer this technology to global partners, from robots to last-mile logistics, working in countries like Australia, the United States, Sweden, and Korea.

In my role, I work with our engineers to improve engineering practices and help them be more effective. It's not just about tools, but facilitating their day-to-day so they can do their job better.

20 years ago, having such a value proposition was absolutely visionary. Surely it was something many couldn't imagine.

It was, and many judged it as unfeasible. But we managed to prove that with new technologies and ways of doing things, it made sense. Today, with more advanced technologies, others are trying the same.

It must be an incredible challenge to work with so many different markets.

As an engineer, it's exciting. Each market has unique challenges: earthquakes in Japan, narrow streets in Europe, regulations like alcohol restrictions in Scotland, or designs for Saudi Arabia, where the right-to-left format affects even e-commerce. Solving these problems is what motivates us.

Productivity in software is not just about how many lines of code you produce, but the quality, the ability to make changes, and the sustainability of long-term work. More than measuring it, we want to understand and improve it.

Focus on productivity

How do you protect the organizational culture from becoming obsessed with productivity?

The focus is not on productivity itself, but on understanding and improving it. We look at metrics at the system level: interactions with code, deliveries, errors. We draw inspiration from frameworks like Microsoft's SPACE, which includes aspects like well-being and friction, but we never forget that output is not everything.

How do metrics influence the daily work of teams?

Metrics should be tools for making informed decisions, not an obsession. They help improve processes within the team, like deciding on an architecture or a change in tools. It's key that teams have the agency to decide what to measure and how to do it.

Setting individual measures would be counterproductive. Gamifying metrics leads to gaming the system, and when you start cheating, you lose all the fun.

Measuring productivity

What advice would you give to companies that want to start measuring productivity?

Understand why you want to measure it. If it's just to meet external expectations or maximize individual benefits, it's a bad idea. Don't set goals based on productivity or measure individually. Measurement should be a tool for improvement, not an end in itself.

In the pursuit of maximizing productivity, we sometimes encounter unintended effects. How do you avoid those risks?

It's about having a clear purpose. If we measure to improve well-being and effectiveness, we're on the right track. But if we turn metrics into a goal, we risk straying from the real impact we want to achieve.

AI and productivity

How does artificial intelligence affect productivity?

AI is changing work patterns. Some engineers consider it indispensable, while others prefer not to use it. It generates more code, but also more review and maintenance work. We are measuring its impact because the relationship between more AI and more productivity is not so straightforward.

Work patterns will change with artificial intelligence. These tools can generate more code, but they also imply more review work. It's not a linear relationship.

In the pursuit of maximizing productivity, we sometimes encounter an unexpected effect. Measuring and improving doesn't always lead us where we want. A good example is the Jevons paradox. In the 19th century, William Jevons observed that by making steam engines more efficient to consume less coal per unit of work, the total use of coal increased instead of decreasing. We observe this phenomenon today with all resources, from fuel in cars to electronic devices, to water consumption. Obsessing over improving a metric doesn't guarantee the impact and real value we want to create. If, in addition to economic gain, we want to generate a positive impact on people and the planet, productivity should be a means, never an end in itself.

A big hug and see you in the real world.