#okr #management #startups

Outcome-focused, input-driven OKRs

How to crack them?

16 min readJan 31, 2021

How do you lead your company or organization to achieve maximum impact?

Even though hundreds of different management frameworks exist, I will argue they all try to accomplish one thing: help us allocate the scarce resources to where they move the needle the most.

And in this article, I would like to present my take on this challenge, which takes inspiration from several sources:

Objectives & Key Results (OKRs) methodology
Outcomes over output concept
Input culture

Over the last two years, we’ve worked very diligently to implement the OKR framework at the company I work for. It gradually evolved to something unique to us, and what makes it unique is how it attempts to reconcile all three elements listed above.

As a result, we have a framework that helps us consistently find new ways to grow the company, despite the currently unfavorable market environment.

How do you come up with fabulous, outcome-based OKRs?

What helps is realizing what an outcome is.

In my work, I define an outcome as a meaningful and lasting change in stakeholders’ behavior.

Meaningful means that it has a positive impact on the organization and that its effect is aligned with the overall direction that the organization is pursuing.

Stakeholders’ behavior is a broad term that could include almost everyone who deals with the organization. In most cases, though, this includes either clients or employees (team members).

Before we move any further, let’s review the OKR structure that I find particularly useful, although not easy to execute correctly. It usually consists of the following four layers:

Objective (impact)
Key result (outcome)
Initiative (to-do)
Milestone (output)

Let’s do a deep dive into each one individually.

Objective

Your objective is the highest-leverage goal that you can think of achieving given your / your team’s place in the organization. It needs to be aligned with the objectives that in the OKR tree are closer to the ultimate, top-level organization’s objectives.

Think of it as a significant development that will have the maximum impact you can provide in any given period.

The objective is not something you can manage directly — this is the end state that you want to see thanks to multiple changes in human behavior.

One useful mental shortcut for coming up with great objectives is thinking in bets. And no, it has nothing to do with gambling.

It’s about setting a direction that’s maximizing your winning chances when you don’t have all the facts. We don’t necessarily bet against anyone in particular. Most bets are with ourselves or with the broader market.

When you feel your objective is unclear and vague, ask yourself: what’s my bet here.

Let’s imagine your thinking to develop a new product. How do you translate it into an objective?

Ask yourself: what's my bet related to this new product? Is it breaking into a new market segment? Then this should become your objective, not the product development itself.

Key results

Outcome-based key results, my special flavor of key results, are measurable changes in human behavior that ultimately lead to your objective.

Very importantly, we need to have a way of directly influencing these changes. There needs to be an intuitive link between our powers to affect and the necessary change.

This is where deep thinking needs to happen to understand what drives the behaviors that translate into the expected impact.

This is also where it’s easy to fail and produce a shallow key result that will derail the entire effort.

Here’s an example to illustrate the point.

Let’s imagine your objective is to fix your ailing after-sales customer support.

A shallow OKR would say something along the lines of:

Improve NPS score from 10 to 50

It looks promising at first glance. It’s measurable, and it is logically related to the objective itself.

Let’s see what’s wrong with it.

It’s an abstract expression of what you want to see, an aggregate number that could be influenced by many discrete issues.
It is a lagging indicator — it shows you what ultimately results from different actions (or lack of action), but it’s not telling you what to focus on to make a difference.

Looking at this key result, you’re not instantly inspired to take specific steps that will get you closer to your goals. You have not dug deep enough yet.

What should you do instead?

I suggest you start with the “5 Whys” method. It couldn’t be simpler — you continue asking “Why”, and usually by no longer than a fifth iteration, you will get to the bottom of things.

For example, continuing with the hypothetical example above and starting from the same objective of fix your ailing after-sales customer support.

Let’s begin with questioning the objective itself: why do you need to fix your customer support? (the first “why”).

Maybe it’s because customers are complaining about slow customer service.

Why are they complaining? (the second “why”)

Because we take on average more than 3 days to respond to their initial and follow-up requests.

Why is that the case? (the third “why”)

Because their requests get buried in our employees’ internal email threads, and it becomes complicated to see what has been closed vs. what needs further action, and sometimes requests even get lost in spam folders.

Why? (the fourth one)

Because we don’t have a system that would make it easy to manage incoming customer support requests.

Let’s stop here.

Already after the second “why” we’ve arrived at something that becomes an excellent key result material — we could decide to change our average response time from 3 days to less than 24 hours.

Depending on our particular situation, we could go one step further and decide that we want to ensure our average time-to-resolution is less than 48 hours.

What’s important though that this outcome will change behaviors (mostly ours, but also customers’), and if we do it the right way, it will be a lasting change of great benefit to the company.

The third and fourth (and possibly fifth) whys would give you inspiration for your initiatives (more on that below), i.e., specific actions you should undertake to achieve your key result.

The above example also assumes that there would be more than one reason behind your objective, leading to more than one key result, i.e., instead of having the objective which says fix customer support, you would have an objective that says beat top industry benchmarks for response time (or similar), and you would break it down into its key results.

When you follow this process, there’s a fine balancing act that you need to achieve.

On the one hand, you want to ensure that you focus your attention on the underlying behaviors leading to the expected objective.

On the other hand, you need to be careful not to take it one step too far. That would happen if, in your “why” analysis, you reached a point where a logical connection between the challenge at hand and the underlying behaviors becomes too loose.

If that happens, what you risk is that your actions will not make a dent in the objective you want to crack.

Initiatives

Interestingly enough, the initiatives layer is not even a part of the original OKR system, as described in some classic books on the subject.

Yet, over time, I’ve started to find it more and more indispensable.

Initiatives are where you create your systems.

Before you read any further, I suggest you look at this classic blog post from James Clear on how systems are more important than goals, and an annual letter from James Bezos on the importance of controllable inputs.

The only thing we can fully control and manage is our actions.

Our actions are in the inputs that propel the flywheel to spin faster and faster.

We don’t achieve our goals by “working on the goals.” We reach them by consistently, systematically doing the essential work that culminates in great results.

You can’t manage abstract numbers, for example, your revenue.

What you can manage, though, is the number of sales calls, outreach emails, presentations, and meetings, etc. These actions become your leading indicators.

How consistently you can work on them, what improvements you can produce in these areas will be the best predictor of your success with achieving your key results.

That’s why it was so important to make sure that you can influence the key results through your actions in a meaningful way.

When you design your initiatives, don’t confuse them with your general to-do list. Think about your key result as your North Star and find 1–3 actions that you will focus on (almost) obsessively to get you there while rejecting any distractions.

Milestones

Not every key result will need milestones. If your result can be expressed through a single, easily measurable number that shows progress, milestones are optional (although still useful).

However, if your key results are more project-style, you might need milestones to show a key result completion plan.

Think of milestones as outputs — individual pieces of work that you deliver that ultimately add to the key result (outcome).

It’s always best to illustrate with an example.

Let’s assume your objective is to get VC funding for your new SaaS project.

Your key result would be to get 10 investors to meet you and listen to your pitch (and yes, technically, this example has a simple number that can be used to measure progress).

Your initiatives could include attending VC / startup conferences for networking purposes, using referrals for introductions, or doing cold reach-out to VC funds. Whatever gets you closer to your North Star (10 meetings).

Your milestones in this example could include outputs such as:

A list of VC funds you want to target.
A list of key decision-makers with their contact information.
A pitch deck to present to investors.
A demo to present during a startup conference.

You should not rush through your milestones. Think of them as the tools that will help you get the job done. You will need the best possible tools that will help you work smarter, not harder.

The tools you develop through the milestones will also help ensure your key results are sustainable beyond the current quarter.

Additional rules of engagement

When you get more serious about implementing OKRs in your organization, you will soon realize it’s not strict science, and there might be many flavors to the process.

Team members will come up with their doubts, they will challenge the rules, and question what should / should not be a part of a good OKR.

If that happens, you should appreciate it, as ultimately, you will develop a system that works for you, in your unique circumstances.

Based on my experiences, though, there are a few additional tips I would like to share.

Things to keep outside of OKRs

Initially, when people get excited about OKRs, they try to use them to manage every project in the organization. Sometimes, you get the sense that if it’s not an OKR, it’s not important and worthy of attention.

However, this quickly leads into situations in which we try to force the OKR framework on things it has not been designed for, and the results are… awkward at best.

In reality, OKRs work best for problems which 1. are changing our behaviors in meaningful ways, and 2. can be reasonably tackled within a quarter or two.

If your challenge is to build a new office for your company, don’t manage it through OKRs. You will most likely do much better using traditional project management tools and methodologies designed for complex systems with multiple timelines and multiple resources that need to be allocated.

It might be a critical project for you, but the OKR framework is not a good fit here.

If you need to get a cyber-security certificate to be approved as a supplier for your partner, it is essential that you get it done. Still, most likely, it’s a matter of following a clearly-defined checklist — no need to bend this checklist into the OKR framework.

As a rule of thumb, work that you can manage through checklists should not be part of the OKR process even if essential to the business.

Amount of time that goes into planning OKRs vs. time for execution

I will admit it is a big trap. It’s straighforward to go down the rabbit hole of trying to craft perfect OKRs, to the point that you realize mid-quarter that you have not started any work. I’ve been there, done that.

To solve this problem, we can use a helpful framework on decision types that I borrow from Jeff Bezos.

There are two types of decisions.

Type 1 — decisions with major consequences that are not easily reversible. Jeff Bezos calls them “one-way doors.”
Type 2 — decisions that can be easily reversed if the outcome is not to our liking. These are like “two-way doors.”

Most decisions we make in everyday work are Type 2 decisions, and, generally speaking, it is better to be 70% informed and act fast than wait to be 100% informed and take little to no action as a result.

The cost of a mistaken fast action will usually be minor, while the upside can be significant.

The cost of lack of action will almost certainly be high.

The vast majority of your OKRs will be Type 2 decisions.

Instead of deliberating your OKRs endlessly, I would suggest you finalize them no later than in the first week of a new quarter (if you have quarterly OKRs) and then focus on actually working on them.

When the quarter ends, run a retrospective, draw conclusions, and improve.

Focus on leading indicators

A well-designed OKR should give you a good understanding of progress throughout the quarter. If your success or failure becomes apparent only close to the end of the period, it’s problematic.

Even when your ultimate goal is simply sales, it’s better to think deeply and discover the most vital signals that correlate with sales and serve as early (leading) indicators.

Again, let’s use an example.

Imagine your objective is to break into the CRM software market.

It would be counter-productive to have a key result that says something like “10,000 USD in monthly recurring revenue”. That’s a very late indicator of your success that tells you little about how successful you are on your way there.

However, when you realize that 50% of your trial users who add at least 3 customers of their own to the system become paying users of your app, your key results could look like this:

Increase the monthly number of signups for the 30-day free trial from 500 to 1000.
Increase the percentage of customers adding at least 3 customers to the system from 10% to 20%.

Shorten your development cadence (time-to-value)

There’s a particular type of OKRs, which I call “build OKRs.”

Build OKRs are common in software development, and the challenge here is that to show value (create impact by changing behavior in a meaningful way), you have to first develop a piece of software.

A similar challenge will occur if your objective requires a good dose of research, experimentation, or any other time-consuming preparatory stage that might stretch beyond a single OKR cycle for completion.

Simply building something is just an output, not an outcome yet. And it could turn out that what you created is completely irrelevant or useless.

There are a couple of strategies you could consider in this case.

First of all, make sure that you’re not suffering from feature creep. Are you sure you need to develop every feature you designed to deliver value to your customers/users? Can you reduce the scope, cut the build time and create value sooner?
Consider fake-it-until-you-make-it approach. Can you build just the bare-minimum system, stripped down to absolute essentials but useful to the users?
For example, instead of a highly sophisticated algorithm that automatically recommends products to customers, can you initially prepare them manually and send them to customers while working on the algorithm?

Your OKR lives in a context

It took me some time to understand that outcomes are contextual, and what sets the context is your place in the organization.

It’s an easy mistake to always focus on the top-level impacts you want to achieve. However, if you do it, you’ll discover that crafting powerful OKRs becomes more and more difficult as you move to the lower levels of your organization.

Instead, I now believe that one team’s outcomes become outputs for another team, i.e., building blocks that help you achieve the ultimate goals (impacts).

However, I will be the first to admit that it’s easier said than done. Many organizations will struggle here, and the main reason for that is the inability (or lack of effort) to communicate strategy down the organizational hierarchy.

What happens then is that anything that’s not clearly and immediately connected to sales (revenue, profit, etc.) become a contentious exercise.

Communicating the strategy, the organization’s purpose, mission & vision, and how each team relates to that is essential to create a robust, fully aligned tree of objectives and key results. Everyone can then produce outcomes at the appropriate level of relevance.

Deliver & abandon phenomenon

The (usually) quarterly cycle of OKRs has many benefits in creating a lot of urgency and bias for action. However, there are some downsides that you should be aware of.

One of the biggest problems is lack of continuity and lack of retrospection on past quarters’ lasting outcomes.

Teams feel the pressure to complete their OKRs within a quarter and develop a different set for the next quarter. As soon as a new quarter starts and the previous quarter’s final scoring is done, attention shifts elsewhere, and the accomplishments of the last quarter drift away.

It can be a very acute problem and something you should be very aware of. A conscious effort needs to be made to revisit older OKRs and ensure the value they created still exists.

Real-life example

Let me give you one practical example of how I would go about rewriting my OKRs to make them more powerful and more outcome-focused, input-driven.

Let’s start with the overview of the original OKR:

Objective: Develop an Intuitive Platform that radically simplifies Product Information Management (PIM)

Key results:

Develop MVP of PIM that becomes a sole product management platform in Q3
Data library that enables customer-centric, effortless product searchability, comparisons, smart recommendations

It was followed by initiatives such as “Identify stakeholders”, “Finalize roadmap and sprint plan”, and plenty of more technical ones, such as “Finalize the ERD”, etc.

When done, I was happy with what we had. Both key results — I felt back then — pointed to a change in behavior.

However, this is, in fact, not a great OKR. Let’s use our framework described in this article to make it much better.

Starting with the objective, what’s our bet here?

I would argue that we bet that we can sell much more with advanced features enabled by the PIM tool and the structured data it provides. Why don’t we start here, then?

Revised objective 1: Increase sales through advanced new tools enabled by the PIM tool (product comparisons, cross-sells, up-sells).

Can we do better? The revised objective is very us-centric. Wouldn’t it be better if we focused on the benefits to the customer?

Revised objective 2: Make it easier for customers to identify the most suitable products through intelligent comparisons and recommendations.

When your company has clearly defined values, it helps to see how your objectives fit with them. In my case, one of our values (that we strongly live-by) is to “wow customer.” On the other hand, “generate more sales” is not one of them.

I like the last version of the objective — it is customer-focused, it has an impact relevant to my team’s context in the organization, and it is consistent with our values.

How do we measure it through key results? Remember, we want to see meaningful behavior change. That’s our starting point.

Let’s try with two key results that reflect slightly different moments in the customer journey:

KR1: A 5% click-through-rate (CTR) on recommendations based on PIM data
KR2: 10% of all add-to-cart action should come from customers who used at least one of the following tools: filtering, comparison, or recommendation.

Both key results point towards a specific change in behavior (customers using new methods to discover products and buy them), and both are highly meaningful.

These two key results become our North Star, but what action do we need to undertake to get there? By answering this question, we will get to our initiatives.

We can argue that nothing will happen if we don’t have the required product data. Since we might not have enough time to collect data for all products, let’s come up with:

I1: collect complete data for all products in the top-3 most-selling categories.

Equally, nothing will happen if our new data is not available on the website.

Please notice how we’re reversing priorities here: we will focus on taking the shortest possible path from having product data to presenting it to customers.

Instead of waiting for the PIM tool to be ready, we will try to present them with the bare-bones technical setup to get customer feedback as soon as possible.

So, our second initiative could be:

I2: show product filters and recommendations based on new PIM data on the e-commerce website.

These 2 initiatives are the minimum we need to obsess about to see the objective’s impact: we need data for the fast-selling products, and we need to present it on the site.

In this setup, our original objective and its key results get moved to milestones. They are outputs on our way to the real impact.

However, we should not neglect them. The way we’re showing the value (impact) in the shortest possible amount of time is a bit “hackish.”

Good milestones will help ensure we have a proper platform for the long-term.

Your milestones could include:

M1: PIM database structure that supports all product information and combinations.
M2: PIM API connectors to the storefronts that we want to support.
M3: PIM UX wireframes.
M4: PIM editor that can replace built-in product management tools of the e-commerce platform.

Please note that my milestones don’t replicate the initiatives. Instead, they provide “the glue” that connects everything and creates a platform that can be used long-term.

What’s one of the great advantages of this rewritten OKR?

It helps you maximize your chances of winning at your bet while minimizes the chances of making a severe loss.

Instead of investing serious resources into developing a new tool, you seek to immediately validate the value that you believe it will generate. Even if initially you operate in a “fake it until you make it” mode.

You get fast learnings, and you can pivot if required before your sunk cost becomes very significant.

Parting words

As the example above shows, writing good OKRs is not easy — you have to think deeply about the impacts you want to create and behaviors that you want to change. And then you have to find actions and tools that will make it happen.

However, the effort is well worth it, as it provides excellent clarity and understanding of what needs to be done to get to the next level in your growth.