What is an objective in the OKR system? Formulating qualitative goals and tracking them with AI

An objective is the qualitative, overarching goal in the OKR system (Objectives and Key Results). In plain language, it describes what a company, a team or a person wants to achieve within a certain period, usually within a quarter. The objective states the "what" and the direction, while the associated key results make progress measurable with numbers (Balanced Scorecard Institute).

An objective is deliberately not expressed in numbers. It should be understandable, ambitious and actionable. An example from practice: "Significantly improve onboarding for new customers" is an objective. Only the key results, such as "Increase the onboarding completion rate to 85 percent", turn it into a goal you can track (Asana). Objective and key results therefore always belong together: one sets the direction, the other measures the path.

Objective and key result: the two parts of an OKR

The OKR model consists of two components. The objective captures qualitatively what you want to achieve, the key result captures quantitatively how you measure success (Balanced Scorecard Institute). The objective is the headline, the key result is the supporting evidence.

A common guideline is three key results per objective, all measurable and time-bound (businessmap.io). At every level, whether company, team or person, a manageable number is advisable: Asana suggests three to five objectives per level (Asana). More goals do not mean more focus, but less.

If you want to go deeper into the measurement side, you will find the details in the article What is a key result?. The article What are OKRs? gives an overall overview of the method.

What makes a good objective

A good objective is significant, concrete, action-oriented and ambitious (What Matters). It sets the direction but leaves open exactly which path is taken to reach the goal. This kind of wording creates room for the team and prevents a goal from shrinking into a mere task list.

A simple rule helps when writing: start with the goal itself, formulate it briefly and understandably, and focus on the result rather than on individual activities (Asana). An objective such as "Build a high-performance analytics team" works because the goal is clear and achievable for everyone (Profit.co).

Typical weaknesses arise when an objective remains too vague ("get better"), when it already contains a metric (which belongs in the key result), or when it actually only describes a measure. An objective is the goal, not the project plan.

From objective to quarterly rhythm

Objectives only unfold their value in ongoing operations. They are set at the start of a cycle, usually per quarter, and measured against the key results during the quarter. A mix of top-down and bottom-up is common: some goals come from company leadership, others emerge within the teams themselves. A guideline suggests that teams and individuals should create at least 40 percent of their OKRs themselves (businessmap.io).

The recurring effort rarely lies in formulating the objectives, but in following up: gathering interim status, assessing progress, producing quarterly reports. This is exactly where many organizations experience friction, because status has to be collected from spreadsheets, emails and tools.

Where AI supports objectives

At this point, AI takes over the preparation, not the goal setting. Which objective is right and how progress should be assessed remains a leadership decision. An AI employee can handle the laborious part, however: gathering interim status from the connected systems, assigning it to the matching key results and producing an understandable interim report. The article OKR tracking with AI describes how this works without manual reporting effort.

This fits into our field of Strategy Intelligence: AI prepares information, people decide. In this way, mere status becomes a basis for decisions, for example when an objective is clearly getting out of sync and needs to be corrected. The step from the metric to action is explored further in the article From metric to decision.

The benefit can be described soberly: less time for collecting status, more time for the conversation about content. The objective remains what it should be: a clear, ambitious goal that a team stands behind.

Frequently asked questions

How does an objective differ from a key result?

The objective describes qualitatively what you want to achieve and sets the direction. The key result describes quantitatively how you measure success. An objective without key results stays non-binding, key results without an objective lose their context. Both belong together and only form a complete OKR when combined.

How many objectives should a team pursue at the same time?

As a guideline, three to five objectives per level apply, whether company, team or person (Asana). Fewer goals mean more focus. Anyone pursuing too many objectives at once spreads attention and resources too thinly and often ends up fully achieving none of them.

May an objective contain a number?

As a rule, no. The objective is the qualitative part of the OKR and describes the direction in plain language. Numbers and metrics belong in the key results. If you already name a target value in an objective, you mix the two components and lose the room that a good objective is meant to provide.

How does AI help with tracking objectives?

AI prepares the progress: it gathers interim status from the connected systems, assigns it to the key results and produces understandable interim reports. The assessment and the goal setting remain with those responsible. This primarily saves the recurring effort of manual reporting, as described in the article OKR tracking with AI.

Who should define objectives?

A mix of top and bottom makes sense. Some goals come from company leadership and ensure alignment, others emerge within the teams themselves and encourage participation. A common guideline suggests that teams and individuals should formulate at least 40 percent of their OKRs themselves (businessmap.io).