How to Set Team Goals and Run OKRs for a Small Team

Why do most team goals get forgotten by the second month?

Most team goals are written in a hurry, in response to a request from above, and then filed. The pattern is common enough to have a nickname: set and forget. The wording is rarely the real problem. What is missing is a routine for looking at the goals every week and acting on what you see.

The cost is real. Gallup has reported that only about half of employees strongly agree that they know what is expected of them at work, and that just 13% strongly agree their manager helps them set performance goals. Gallup also estimates that managers account for at least 70% of the variance in team engagement scores. Clarity about goals is one of the few things a team lead can change without anyone's permission.

A useful way to think about it is as a cycle with five beats: draft the goals, align them with the people who matter, track them weekly, review them honestly, and reset for the next quarter. The rest of this guide walks through each beat.

What is the difference between OKRs, KPIs and SMART goals?

They are three tools for three jobs, and a team can use all of them at once. Objectives and Key Results (OKRs) were developed by Andy Grove at Intel and popularised by John Doerr. SMART is a wording test usually credited to George Doran's 1981 article. KPIs are simply the indicators a team watches.

The quickest test is one question: if we do nothing special, will this number be fine? If yes, it is a KPI and you watch it. If no, and it matters this quarter, it belongs in a key result.

How do you choose which goals to set?

Start by collecting candidates from five sources: published company goals, what your own manager is accountable for, customer complaints and comments, the problems your team already grumbles about, and must-do deadlines such as audits. Fifteen to twenty-five candidates is normal.

Then score each candidate from 1 to 5 for impact, urgency, the team's control over the result and effort. Control is the factor most people skip. A goal that depends almost entirely on another department is a wish, however important it is.

Sort every candidate into one of five piles: objective this cycle, health metric, initiative only, park for later, or no. Before you polish anything, take the ranked list to your manager and ask which two they would choose if only two could move. That fifteen-minute conversation heads off many surprises later in the quarter.

• Stop at five objectives. For a first cycle, three is better.
• Write down what you parked and when you will look at it again.
• Aim for objectives where a manager's goal meets a problem the team already wants solved.

How do you write objectives and key results that are measurable?

An objective is a short sentence, twelve words or fewer, that describes a better state of affairs: 'Customers get the right answer the first time.' It has no numbers in it. Your team should be able to say it from memory.

A key result proves the objective is happening. It needs four parts: a measure, a baseline, a target and a date. 'First-contact resolution from 62% to 75% by 16 April' has all four. John Doerr's What Matters guidance suggests three to five key results per objective; for a single team, two to four is usually enough.

The most important distinction is between outputs and outcomes. 'Train all agents on the billing module' is an output: you can finish it and nothing may change. 'Reopened tickets from 14% to 8%' is an outcome. Write key results as outcomes and treat the tasks as initiatives, the bets you place to move the number.

Never guess a baseline. If nobody knows today's value, spend two weeks measuring and set the target afterwards. And wherever you measure speed or volume, add a paired quality measure so that the number cannot be improved by damaging what you care about.

Should team goals be stretch targets or promises?

Both, as long as each one is labelled. A committed key result is a promise that other people are planning around; you expect to hit it in full and you escalate early if it slips. A stretch key result is an ambition with roughly an even chance of being fully achieved.

Google's re:Work guide to OKRs describes a sweet spot of 60 to 70% for ambitious goals and notes that consistently scoring 100% suggests the goals were not ambitious enough. Decades of research by Edwin Locke and Gary Latham point the same way: specific, difficult goals tend to produce higher performance than easy or vague ones, provided people are committed to them and get feedback.

A simple calibration helps. Ask each owner for a confidence score from 1 to 10 at the start. A committed result should begin at 8 or higher. A stretch should begin near 5. A 'stretch' that starts at 9 is a sandbag, and a 'promise' that starts at 4 needs a conversation with your manager today.

How do you get the team and other teams aligned without a long cascade?

Alignment comes down to three questions. Up: which higher-level goal does each objective serve, and does my manager agree? Sideways: which teams do we need something from, and which need something from us? Within: does each person know which key results they contribute to?

You can answer all three in about a week. Draft the goals with the team in a 90-minute workshop that starts with fifteen minutes of silent writing, so that every voice is on the wall before discussion begins. The CIPD's guidance on performance management observes that it is not generally better for employees to set their goals entirely on their own; what matters is that people understand and accept them. A workshop where you bring the direction and the team shapes the measures achieves that.

Then check the draft with your manager, make a specific request of each team you depend on, and treat 'no' as an acceptable answer. An honest no in week one costs far less than a reluctant yes that fades by week six. Finally, publish the goals on one page where the team already looks.

What should a weekly OKR check-in look like?

Fifteen minutes, the same slot every week. Before the meeting, each owner updates two numbers for their key result. Progress is arithmetic: how far the measure has moved from baseline to target. Confidence is judgement: from 1 to 10, how sure are you that we will hit the target by the end of the cycle?

In the meeting, look at the board for three minutes and read out only what changed. Spend eight minutes on exceptions, the key results that are at risk or whose confidence has fallen, with three questions each: what changed, what is in the way, and what will we do this week? Take two minutes for any health metric that has gone amber or red, and two for one commitment from each person.

Do not solve problems in the room. Name them, then let the two or three people involved deal with them afterwards. And when someone reports a low confidence score, thank them before you ask anything else. People watch what happens to the first person who tells the truth, and the answer decides whether you ever hear bad news early again.

• 0 to 3 minutes: the board, changes only
• 3 to 11 minutes: exceptions, three questions each
• 11 to 13 minutes: health metrics
• 13 to 15 minutes: one commitment each

How do you score OKRs and review the quarter?

Around the midpoint, replace one check-in with a longer review that asks a different question: is the plan still right? Change initiatives freely if they are not moving their key result. Change a target only when a fact has changed or the baseline proved wrong, and record why.

At the end, score each key result from 0 to 1 by how far it travelled from baseline to target. Compare it with the bar for its type: 1.0 for committed and about 0.7 for stretch is a common choice. Then spend most of your time on why. A miss usually has one of four causes: the wrong goal, the wrong target, the wrong bets or no capacity. Each leads to a different fix.

Finish with a decision for every key result (retire it, move it to your health metrics, continue, rewrite or drop) and choose no more than three changes to how you will run the next cycle.

Should goal scores feed performance reviews?

With care. Google's guidance states that OKRs are not synonymous with performance evaluation, and there is good reason for the caution. In the paper 'Goals Gone Wild', Ordonez, Schweitzer, Galinsky and Bazerman describe the side effects of goals applied carelessly, including narrowed focus, more risk-taking and unethical shortcuts. Wire a single score directly to pay and you invite people to manage the score.

A fairer approach treats goal results as evidence within a wider picture: what the person contributed, how hard the target was, what lay outside their control and how they acted when things became difficult. How goals connect to ratings and pay is governed by your organisation's policy and, in many places, by employment law, so speak to your HR team before changing anything.

If you would like a starting measure of your own habits, the AssessAll assessment Performance Management & Accountability is a relevant one. For personal goals, Jobulary's article on the WOOP method is a helpful companion. And if you want the whole cycle as ready-made tools, Bodhih Training's Team Goals and OKR Cycle Kit includes the workbook, forms, scripts and calendar described in this guide.