Take a Lesson From Google, Intel and Others: Use OKRs Instead of Product Roadmaps

Image: Ricardo Arce on Unsplash

This is Part 16/18 in the series, “How to Build an Innovative New Product or Company,” on the topic of using quarterly, bottoms-up, transparent objectives and key results (OKRs) as an organizing principle.

OKRs are a simple data-driven organization platform created by Andy Grove of Intel and championed by Google, among many other well-organized companies in recent years. John Doerr was the glue between Grove and the rest of us, and Doerr has been OKR’s greatest evangelist. I collected many of the ideas here from his book, Measure What Matters. Doerr credits the existence of OKRs as the system that allowed Intel to quickly rally around a competitive crisis against Motorola in 1980, leading to Intel’s dominance of the CPU industry for the following 20–30 years.

OKRs are not about adding administrative work. You can create yours in only a few hours per quarter, if there is already a clear corporate direction in place. Nor are they about signing up team members for superhuman accomplishments. They are simply about making sure everyone gets the greatest leverage from their efforts.

Progress on objectives and key results should be tracked and shared weekly as green (“complete” or “on track”), yellow (“updating” or “needs attention”) or red (“is this still valid? should we drop it?”).

At the end of each quarter, progress and notes about each objective and key result are documented. Google measures progress against each item on a percentage basis at the end of each quarter: 70%–100% is green, and 0%–30% is red. Notes allow teams to adjust progress beyond the numbers. If the goal were 100 sales calls, but you had 80 calls with great follow-up, shouldn’t that count as a complete success?

Doerr has a saying: “We need teams of missionaries, not teams of mercenaries.” By giving teams a business objective (e.g., “decrease customer onboarding time by 50%”) instead of a product roadmap outcome (e.g., “deploy onboarding tutorial”), you are motivating and empowering everyone in your company to be an imaginative contributor in the pursuit of customer value, instead of deploying them as a blunt object in building what the leadership team dictates.

The connection of OKRs to strategy

In “Four Steps to Develop a Strategy,” I outlined a process for developing a corporate strategy and gave the following example for Southwest Airlines. I’ll re-create Step 4 here:

Southwest Airlines’ strategy (circa 1970s)

User + how we will delight them. What are the two to five unique and pivotal decisions that will define our solution?

• Limited passenger services … such as no first class; no frequent flier miles; open seating/first-come, first-on; passengers cleaning up after themselves to speed up gate turnarounds; no meals.
• Be lowest cost … such as flying one aircraft type, the 737; thus, also not needing to train and certify pilots on other aircraft; no travel agents or third-party travel websites — you can only book with the airline, thus avoiding royalty fees.
• Highly agile and invested grounds and air crews … such as personality-infused pre-flight safety presentations; employee stock ownership; ground and flight crews well compensated; crews allowed to join unions; crews empowered on any safety issue. Let employees be our “eyes and ears” for new ways to continually improve our service to customers.

Step 4 of a strategy has several business objectives for the company. These can become company objectives in annual or quarterly OKRs. Or, more likely, the company’s OKR objectives would be the next level of operational detail behind these, evolving them over time, while the strategy remains more consistent.

For example, “Let employees be our ‘eyes and ears’ for new ways to continually improve our service to customers” could be a corporate OKR objective that hangs around for a long time with evolving key results: hiring the right people, training them, providing the tools for them to perform this role.

In my experience, startups (perhaps all companies) have four objectives, so OKRs should roll up to them, even if they don’t explicitly claim to do so:

1. Serve more customers.
2. Provide the best service we can to those customers, thus driving more usage and value; this also increases the likelihood of renewals/repeat purchases.
3. Conduct prudent cash management.
4. Be a preferred place to work, grow and develop a career/craft.

How to create OKRs

OKRs …

• Should be created by each employee for themselves every quarter. At the same time, the prior quarter’s OKRs should be evaluated. In some cases, quarterly and annual OKRs may exist in parallel. Annual OKRs are released a couple of weeks before the year. In the first couple of weeks of a quarter, teams develop and communicate their quarterly OKRs.
• Should be big ideas that move the business forward and about which customers will care.
• Exist for the company overall, every department and every employee. Team leaders should have OKRs similar to, but not identical to, their department’s objectives.
• Are public among other employees in the company. Post them on an employee’s office door, or make them searchable online. Because your objectives are public, it can help you say “No” to others’ requests for your time that are outside their scope; your priorities are clear for everyone to see. (That is not to say employees shouldn’t help each other achieve their OKRs.)
• Are a mix of top-down (e.g., from your manager) and bottoms-up (e.g., from you or your team members).
• Are divorced from compensation. They should be used to motivate big goals and structure a path to success, not for creating criticism and sandbagging.
• Objectives and key results should all start with a verb: establish, deliver, test, get, develop, publish, prove, prepare, recruit, hire …

OKRs have two components

Objective: What you are trying to achieve

• For example: “Release the first version of our email product.”
• Three to five objectives per quarter. What you say “No” to is as important as what you say you will focus on.
• Most are “committed objectives” you’re expected to achieve. Some can be marked as “aspirational objectives.” The latter are pure stretch goals that are there to push the boundaries of what may be possible. Committed objectives should consume all the team’s resources. Aspirational ones go above and beyond expected abilities and may not get completed. Thus, they would carry over to the next quarter.
• Only list what needs to be given emphasis. Do not include all ongoing aspects of someone’s job. For example, don’t list “conducting weekly 1–1 review sessions” or “managing daily standup meetings.”

Key results: The measurable steps you will take to achieve the objective

• For example: “Deliver the three new features to staging by March 1.″
• Three to five key results per objective.
• Measurable=has a number in it. There must be evidence of completion, such as a link to a product or document created, or the results of a test report.
• Timebound=when applicable, has a target date in the quarter for delivery, or end of quarter is implied.
• Flexible=key results can be edited/added/deleted during the quarter, as new information comes in.
• Managers’ key results often become some of their direct reports’ objectives.
• Accomplishing all the key results should trigger success for this quarter on the objective, though objectives may hang around for multiple quarters, if needed.
• Quantity goals should be paired with quality goals. For example, if having 100 sales calls is one key result, having them lead to 20 in-person meetings could be another.
• Identify dependencies early on: Does one team’s key result require another team to contribute? If so, is completing this in the other team’s OKR?

OKRs can help keep multiple layers of a larger organization aligned to a shared strategy

I have been asked if corporate strategy and product strategy are the same. If a company has multiple products, do they all have individual strategies? How do strategies roll up or down?

As my experience is with companies of fewer than 500 employees (and mostly 30–60), I have to defer to others on what it takes to set a strategy within larger companies. P&G, for example, has a corporate-level strategy, a category-level strategy (e.g., skin care) and a product-level strategy. The individual strategies of which customer to target and the specifics of how to win will vary, but the core “two to five unique and pivotal decisions of our solution” (which P&G refers to as “reinforcing rods”) remain consistent. P&G doesn’t use OKRs, though they use a similar process. Either way, those “reinforcing rods” can help OKRs remain consistently anchored in the corporate strategy, even at lower levels of the organization.

OKRs help keep companies from falling into the waterfall product development process — thus keeping the focus on fast product-feedback loops

Be thoughtful and strategic about setting objectives. They can (and should) persist for many quarters, if needed, to fully realize them. Jeff Bezos, Amazon’s founder, says: “We are stubborn on vision. We are flexible on details.” He could have also said: “We are stubborn on objectives. We are flexible on key results.”

In other words, choose carefully what is most critical to your success, and anchor your objectives to them. Then iterate your key results every quarter, as you make progress and learn. No one can predict how many iterations a new innovation will require before it will work.

Marty Cagan, founder of Silicon Valley Product Group and author of Inspired: How to Create Tech Products Customers Love, gives an example that objectives should be business results, not tactics; the former are only a mechanism by which a company can measure progress. For example, “implement a PayPal mechanism in our product” should not be an objective. It is a potential way to reach an objective that may or may not work. “Reach more customers outside of France” is a business result, of which a PayPal integration may be a component.

At the end of the video of Cagan’s talk, he discusses how using OKRs is the more effective antithesis of a standard product roadmap most (non-startup) companies use today. A product roadmap is where the executive team provides ideas; puts ideas through a business-model prioritization process; creates requirements for each item; builds and tests; and finally, months later (and thus far with no user input or feedback); deploys, markets and releases the features. That process is all about features and is slow. OKRs re-orient teams to be about business outcomes.

For startups, consider setting up OKRs in terms of learning a set of lessons
Tom Chi, a product management rapid prototype evangelist (formerly from Google X), advocates for a culture of learning. Part of that is to test out ideas as quickly as possible and create key learnings from the experiment, instead of debating and choosing from within closed-door meetings. Key learnings are actual test results of the types of decisions that otherwise would be argued about by non-users prior to building something. Why not just build it and find out the answer? In a video, Chi calls these closed-door arguments a fallacy — looking for “the truth of the invention that hasn’t been invented yet.”

Setting up a key result in terms of key learnings can be especially powerful. A key result of “test 20 new prototypes and generate 10 new key learnings by Oct. 1” is great, for example.

Also published on strategist.blog, where I share thoughts on building analytics-based products and businesses (“product strategy”). Any books or other resources mentioned or referenced in this article are listed here.