BY Fast Company 4 MINUTE READ

Apple’s new work policy, which mandates that employees be in the office three days per week, stripped their hybrid model of its flexibility. The announcement led to intense pushback from employees, including the departure of one high-level director who resigned in protest.

If a flexible-work policy can improve workplace satisfaction and employee retention, as well as a company’s ability to hire and retain the best talent, why are some companies still forcing a return to the office? In an open letter to Apple, one disgruntled employee said the decision was driven by the “fear of the future of work, fear of worker autonomy, fear of losing control.”

Bringing employees back to the office is a step backward. “Offices are an artifact of the industrial age,” says Brian Elliott, executive leader of Future Forum. Instead of mandating a return to offices to regain control over employee output, companies should embrace new ways of improving productivity with technology. With the rise of digital tools, such as Slack and Zoom, it has become easier to not only centralize virtual work—removing dependence on shared physical spaces and rigid 9-to-5 schedules—but also track, measure, and improve output.

While some jobs obviously can’t be done from home (construction, transportation, and film production, among many others), most fall into a spectrum of flexibility. According to a study by McKinsey, finance and insurance have the highest potential for remote work, with the ability for 75% of time spent on activities to be done remotely without a loss of productivity.


Customizing work policies across teams and departments can be a murky, time-consuming task. The companies most likely to succeed in creating flexible work policies will be data-driven, constantly gathering data and feeding it back into the system to drive continuous learning.

So how can leaders effectively use data to enable flexibility? Software engineering teams provide an interesting example. A developer’s day consists of many feedback loops. Some loops are big, such as validating that a feature meets requirements; while others are small, such as checking that a code change works in a local development environment. When developers face complexity and friction (such as too many meetings, lack of automation, and coupled architectures), the time it takes to complete a single feedback loop increases.

Effective teams optimize feedback loops so that more can be performed in a shorter period of time, improving the overall throughput of the system. For instance, if a team sees a spike in the length of time it takes to ship a new feature—the team might ask themselves, are code reviews being completed in a timely manner? If not, then the team can set up reminders for code reviewers when work is slipping, or set up a dedicated time for reviews.

The same methodology can be applied to just about any department. Just as engineering teams aim to release software quickly in small batches (elite teams release software multiple times per day), finance departments try for continuous closing processes to avoid end-of-quarter work spikes. Accounting teams can measure the time it takes to close the books every month and then experiment with processes to bring them closer to a continuous close, like setting up change management, reducing the number of data feeds, and using machine learning to spot and flag anomalies.

Every team and department is different, so it’s important to gather data that helps teams answer questions like these (and as a result, home in on the right policies and processes that enable flexible work).


As the cofounder of a company that has studied developer productivity among 300,000 developers, I’ve had a unique vantage point into the work-policy constraints that many companies are facing. For instance, we found that developers code less than an hour per day globally, which may indicate that there is a general industry need to protect more focus time for developers.

There are several learnings that have become apparent from studying software development that can generally be applied to all departments when setting a strategy to empower individuals and teams with data: protect privacy, dimensionalize data, and measure change over time.

Individual data should always be private, unless it is shared by that individual. Spying on employees is self-sabotage; in the pursuit of answering a question about productivity, it destroys psychological safety and creates a culture of fear. Instead of using data as an arbiter of individual productivity, teams should use data as the fuel for positive change.

Even with good intentions, making data visible to improve collaboration can be met with fear from individuals that it will be used to micromanage them. Managers can communicate the benefits of gathering and measuring data by taking a “data for one, data for all” approach. In this approach, individual data is never shared with a manager; rather, everyone sees the same aggregated and anonymized team data in a single dashboard.

In some cases, making data visible can also be a driver of positive change across multiple departments. Product analytics, for instance, has historically been siloed inside of product teams. “By making customer data visible across all teams—including product, marketing, sales, and customer success—everyone can work within the same context,” says James Gross, cofounder of customer growth platform Variance. Revenue teams can instrument metrics like conversion rate between product milestones and the average time to complete each milestone. By setting up team alerts when milestones are achieved and using those alerts to guide customers through a product funnel, teams can optimize the time to convert product-qualified leads.

Context matters, which is why teams should also dimensionalize their data. Microsoft’s SPACE framework is built around the notion that software-engineering productivity cannot be reduced to “one metric that matters,” and instead should be balanced across at least three dimensions. Engineering teams can balance quantitative measurements for performance and efficiency, like code review velocity, lead time, and deployment frequency, with satisfaction measures compiled from surveys and one-on-ones. “Making feedback a regular, cultural practice is key to not only improving the happiness and retention of employees, but also building a high-performing team,” says Andrew Zhou, cofounder of Kona, which provides daily mental health check-ins for remote teams.

In a company update last year, Amazon CEO Andy Jassy shared his perspective on the future of work, and it was refreshing: “We’re going to be in a stage of experimenting, learning, and adjusting for a while as we emerge from this pandemic.” Teams are their own best reference classes, which is why it’s important to measure change over time.

By setting up systems to observe short- and long-term trends, teams can determine if feedback loops are improving and run experiments to measure the effectiveness of flexible-work policies. Engineering teams that invest in better documentation, for instance, can shrink the time for new developers to onboard, which will be a significant factor in how well companies can retain tech talent going forward.

There are many solutions, practices, processes, and tools that can help teams thrive in a flexible work environment. Instead of requiring employees to work from an office, leaders should instead focus on setting up the digital infrastructure and processes that improve visibility and support teams to make their own decisions.


Brett Stevens is the cofounder of, a data platform that helps engineering teams measure and improve performance.