Skip to content
Yeti Distro

How to Improve Execution: Learnings on OKRs

Alfred Lua / Written on 03 June 2020

Hello there,

There is a lot going on around the world right now, so I appreciate you opening my email. I'm assuming you might be looking for something different to read. If you are, I hope you'll find this useful in some ways.

(If now doesn't feel like the time to read something like that, I'll understand if you choose not to.)

Ideas vs execution

As much as I hate process, good ideas with great execution are how you make magic. — Larry Page, Google Cofounder

As marketers and product people, we have many ideas. New marketing campaigns, new features. Some are good, some are bad.

While it is important to have good ideas, I think it is even more important to be able to execute them well. An okay idea executed well is better than a great idea executed poorly.

I have been feeling that our execution at Buffer has gotten worse over the past few years. We used to be the most popular social media management tool but many competitors have caught up and overtaken us. For example, Later, which started four years later and has a smaller team, has built a product I'd recommend over our own product. I have so much respect for the Later team.

An example of a good idea executed poorly is our multi-product strategy. We unbundled our products to let customers mix and match the products they need rather than pay for an expensive all-in-one solution. But the unbundling didn't work for various reasons. One reason is that after two years of unbundling, visitors to still couldn't actually mix and match the products. There was a lot of friction in trying more than one product.

Could we have done better? I think so.

We have a team of smart, hardworking individuals. I believe we need a better system to help us execute well. We recently decided to try Objective and Key Results (OKRs) again. We had tried OKRs a few years back and thought it didn't suit us. Perhaps we didn't understand it well.

I thought if we want to implement it well this time, we need to fully understand it.

After reading Measure What Matters by John Doerr, I realized I had misunderstood many things about OKRs. I'll share my learnings here.

(The quotes below are from the book.)

The additional guide to OKRs

There are many blog posts about OKRs, which usually cover only the basics:

  • An Objective is WHAT is to be achieved, no more and no less. By definition, objectives are significant, concrete, and action oriented, and (ideally) inspirational.
  • Key results benchmark and monitor HOW we get to the objective. Effective KRs are specific and time-bound, aggressive yet realistic. Most of all they are measurable and verifiable.


I actually think it's bad to understand only the basics of OKRs if you want to adopt it. I'll admit I didn't properly understand OKRs when we tried it the last time, and it didn't go well.

Here's what I wish I had known:

1. Start with why

When we tried OKRs previously, we simply dived into setting objectives for each department and then individuals. I think we overlooked the first step, which is understanding the reason for pursuing any objectives.

The top-line objective—to make a wealthy person wealthier—lacks intrinsic motivation for the general manager, much less for the team's East Coast scout or the PR intern slaving away at the copy machine.

Jini Kim founded Nuna, a healthcare data and analytics company, because she wants to improve health care for everyone. Her brother was diagnosed with severe autism at two years old. As immigrants with limited resources, her family would have gone bankrupt without a safety net. She had to sign her family up for Medicaid—at nine years old.

Nuna has been using OKRs to achieve its mission of improving health care for everyone.

Having a strong WHY motivates the team to think about and set more ambitious objectives (WHAT) and key results (HOW).

2. Objectives can be span many months and years

While OKRs are often set quarterly, objectives can span a longer timeframe. When Google started working on Chrome in 2008, they set an annual objective to "develop the next-generation client platform for web applications." The main key result was "Chrome reaches 20 million seven-day active users."

They failed to hit that target. In 2009, they kept the same objective and upped the key result to 50 million seven-day active users. They failed again. In 2010, they upped the key result again to 111 million seven-day active users. And this time, they achieved their goal.

Where an objective can be long-lived, rolled over for a year or longer, key results evolve as the work progresses.

Having a longer-term objective makes it easier for everyone in the company to plan their quarterly OKRs. If employees know that the company is still focusing on that particular objective, they can immediately start planning their OKRs at the start of a new quarter instead of waiting for their managers to tell them a new objective.

3. Not all OKRs should be cascaded down the organization

Cascading OKRs from the top of the company to the entire company can make operations aligned. But if all OKRs are cascaded down the hierarchy, it actually becomes ineffective.

Consider Google. It has more than 100,000 employees. It will take months for the company OKRs to cascade down from managers to reports after numerous meetings and reviews. And because it will take so much effort to create those OKRs, people will become less flexible. They will be reluctant to revise their OKRs even when circumstances have changed.

A fully top-down system will also prevent "frontline" employees from contributing. Imagine I know how customers are using Buffer but I'm unable to influence what we build. (This did happen.)

The solution is to allow employees to set their own OKRs according to the company's OKRs:

Precisely because OKRs are transparent, they can be shared without cascading them in lockstep. If it serves the larger purpose, multiple levels of hierarchy can be skipped over. Rather than laddering down from the CEO to a VP to a director to a manager (and then to the manager's reports), an objective might jump from the CEO straight to a manager, or from a director to an individual contribution. Or the company's leadership might present its OKRs to everyone at once and trust people to say, "Okay, now I see where we're going, and I'll adapt my goals to that."

4. OKRs should help us be focused

The ideal number of quarterly OKRs is between three and five. This is to ensure that the company, team, or individual is focusing on only what counts.

But OKRs should not be a catchall wishlist or the sum of a team's mundane tasks. I made the mistake of crafting my OKRs to include everything I have already planned to do. Instead, I should have used OKRs to narrow down my projects to those that align with the company's OKRs.

[OKRs] are a set of stringently curated goals that merit special attention and will move people forward in the here and now. They link to the larger purpose we're expected to deliver around. "The art of management," Grove wrote, "lies in the capacity to select from the many activities of seemingly comparable significance the one or two or three that provide leverage well beyond the others and concentrate on them."

5. OKRs can be changed mid-cycle

One thing we didn't like about OKRs was that circumstances at Buffer were (and are) always changing and we thought quarterly planning doesn't give us the flexibility to change things during the quarter.

In fact, OKRs are meant to be changed or dropped if they no longer serve our purpose. John Doerr recommends weekly check-ins to track progress and make adjustments.

If the team is failing to meet an OKR, they should devise a plan to turn things around. If the team realizes an OKR is no longer relevant, they can drop it. They should also inform everyone who depends on that OKR and reflect on what has changed that they did not foresee at the start of the quarter.

Giving it another go

A consistent theme in the book is that many of the companies mentioned didn't successfully implement OKRs on their first try. This gives me hope that we might do better this time round. I'm excited to share the results once we have tried OKRs for a few quarters.

There are many things that I have not covered here, which I think are important, too. I highly encourage you to read the book if you wish to properly understand OKRs.

And if you do, I'd love to chat.