---
title: "Objectives and Key Results: The Complete Map of OKR"
summary: "OKR - objectives and key results - is a short-horizon change management method that travelled from Andy Grove's Intel with John Doerr into Google and from there into the world. This article assembles the model whole: the division of labour between a textual objective and numeric key results, at most five objectives per level, weekly confidence tracking, the retrospective at the turn of the quarter, and the two conditions whose breach destroys the model - no money attached to a key result, and no punishment for falling short. It also covers the difference from KPIs, where OKR sits between strategy and everyday work, radical transparency and its limits, and the two areas where the model is known to get hard: sales and individual appraisal. The source is Negotiator episode 58 with Henri Sora, co-author of a Finnish book on the model."
datePublished: 2021-01-17
dateModified: 2021-01-17
originalLang: en
section: tools
sections: ["tools","economy"]
authors: ["Samantha"]
tags: ["OKR","objectives and key results","goal setting","change management","KPI","strategy","self-organisation","incentives","John Doerr","Andy Grove","Henri Sora","Sami Miettinen"]
canonical: https://ai.neuvottelija.com/okr-malli-tavoitteet-ja-avaintulokset-opas/
---
# Objectives and Key Results: The Complete Map of OKR

# Objectives and Key Results: The Complete Map of OKR

> **Summary:**
> OKR is a **short-horizon change management method**, not a measurement system. The objective is text, the key results are numbers, and their job is to tell you whether the objective is being reached.
>
> The model's two conditions are its whole idea: **no money attached to a key result**, and **no punishment for falling short**. Breaking either turns goal-setting into a budget game in which a person's interest is to negotiate the target as low as possible.
>
> And its most important constraint is a count: **at most five objectives per level, preferably fewer.**

**What this article is made from.** The source is Negotiator episode 58, in which Sami Miettinen interviews **Henri Sora**, co-author of the Finnish book *Strategia arkeen OKR-mallilla* ("Strategy into everyday work with the OKR model"; the other author is Juuso Hämäläinen). The article is assembled from the episode's content; where Sora describes the model from a practitioner's standpoint, Miettinen pushes back from negotiation and incentive theory. Points where the answer stays open are marked as open.

---

## Where the model came from

The chain is short and worth knowing, because it explains the model's character.

**Andy Grove** developed the mechanism at Intel. **John Doerr** learned it there, moved into venture capital, and brought it into **Google's early days**, where it was adopted immediately. Google has, on Sora's account, said in several places that OKRs produced **more than tenfold results** relative to what would have been expected — and the model is still part of how the company is run.

That last point is the strongest of the arguments: this is not a startup-phase trick that a grown company discards. It is in use in a very large organisation, which makes it **quite a substantial exercise**.

Doerr's own book **Measure What Matters** tells the stories. Sora names directly what it leaves out, and why their book was written: *afterwards you can be left with the feeling of, but what do I actually do?*

## Two components: text and numbers

The core is a division of labour between two things.

**The objective is text.** A short, clear, mildly inspiring description of where you want to go. Sora's example is three words: **go to the American market.** It should describe clearly enough where to go and what to do.

**The key results are numbers.** Under the same example: *five per cent market share by year end*, *five reference customers*.

And their job is precise:

> Those numbers are what tell you whether we are reaching the objective we are aiming at.

From this follows the terminological precision Sora insists on maintaining: **key results are not metrics.** If you are talking about metrics, you are talking about KPIs. If you are talking about OKR, you are talking about key results. The distinction is not pedantry; it is what stops the model sliding into something else.

## Where OKR sits

Miettinen offers his own time-horizon taxonomy in the episode, and Sora confirms the placement:

| horizon | what |
|---|---|
| 0–1 year | **tactical** — this is where OKR belongs |
| 1–5 years | strategy |
| 5–10 years | vision |
| beyond 10 years | system-level disruption, where the basic business model starts to age out |

Sora's own frame is three boxes in sequence: **strategic thinking** (options, scenarios) → **strategy** (the decision about what to do) → **OKR** (how that decision is brought into the company's everyday work).

> Its purpose is precisely to bring the strategy into the company's everyday work, into the tactical side.

The model can look further out, but above all it is **an instrument for managing the near term**.

## How objectives travel through the organisation

This is where OKR differs both from traditional Taylorist command and from pure self-organisation. Sora calls it a **mechanism that complements self-organisation**: it too needs a strategy, and pointing the direction is typically management's job.

The order is:

1. **Annual objectives** at company level.
2. **The leadership team sets its own objectives for the next quarter** — only the first one, not the ones after it.
3. Objectives are derived **from each level to the next**.

But the direction of travel is not only downward. Miettinen asks about this directly, and Sora confirms: a lower level can raise something back up. *Hold on, that is an important thing* — and then it is lifted to the top level as a company-wide matter.

And the thing that makes this more than administrative cascading:

> Here people do what they negotiate. It is not a matter of a manager arriving and ordering that here are your objectives, now execute. **The mechanism is that we negotiate together how we might achieve this.**

## Two conditions whose breach destroys the model

This is the most important part of the article, because it is where OKR implementations actually fail.

### Condition 1: no money attached to a key result

Sora states it as absolute: **no monetary incentive may be tied to reaching the objective.** The reasoning is mechanical rather than moral.

If a bonus attaches to a key result, the person's incentive is to **negotiate the target as low as possible.** Sora describes the sequence precisely: you say that was a tough target, perhaps we land at half of it; then there is haggling upward and you settle at 80 or 90 per cent — and *people can already be smiling to themselves that we got it low enough, the money is safe.*

That is not goal-setting. That is a **budget game**.

Miettinen connects this to economics: he has interviewed **Bengt Holmström**, who developed principal–agent theory and won a Nobel for it — and whose later thinking is that **monetary incentives work poorly in a holistic environment**. Holmström's recurring example is care work, where a bonus system does not work well.

### Condition 2: no punishment for falling short

The second condition matters at least as much, and only together do the two produce the model's benefit:

> Once these two exist, **people can decide for themselves how far they would dare to think about this objective.** And that can be far beyond what a manager would have dared to ask for.

Sora adds his own experience, which is the model's psychological core: when you have promised it yourself, the moment of delivery can bring a sweat — *but you promised it yourself, so you also have to do it yourself. And that is genuinely binding.*

Miettinen's summary: this is about **team honour rather than sub-optimisation**, which easily follows once money enters.

## The weekly rhythm and the confidence figure

OKR differs from other goal management in its tracking frequency: **weekly**, or if that is impossible, fortnightly. A short session, typically attached to the company's existing management routine.

For each objective, three things are reported: **where we are, what we intend to do, and what the confidence is that we will reach the target.**

And attached to it is a cultural condition that determines the whole mechanism's value:

> If there are problems, **the hand has to go up.** Not the way where you leave it to the wall at the end of the quarter and then explain.

The benefit is earlier reaction. Without this, the weekly rhythm is merely reporting.

## At most five

The focus limit is hard: **at most five objectives per level, preferably fewer.**

Sora's example is telling. He was in a workshop where the leadership arrived with **17 objectives** — each with a good reason. Once they had been condensed to five, the situation was a fine one for everybody:

> Now we know that we are only doing these.

The reasoning is not aesthetic but practical: when there are many odds-and-ends objectives, **nobody knows which of them matters.** And the limit also gives refusal a form — when somebody comes asking for more, the answer is that it is not among this quarter's objectives, shall we look at it at the turn of the quarter.

## The turn of the quarter: retrospective first, forward second

Sora is very precise about this order, and it is the step most people skip.

1. **Retrospective first.** Look back: what happened, and **what can we learn from it** — whether it went well or badly.
2. **Only then forward.** Are the objectives still valid, are they good, do they need changing?

And what must not be done:

> Not the way where you arrive, **update the numbers and off you go again.**

There are two practices for scoring key results: some do it slightly before the quarter ends, so the next set can be made at the turn; others want exact results and score only at the end. Once an organisation has practised this for a while, the process speeds up — people arrive ready: *these were our lessons, I would propose we look at these next.*

The rhythm is not necessarily quarterly. The episode mentions a company using **three periods a year**, because in a quarterly rhythm the handover falls across the holidays and there is no time to score the outcome.

## OKR is not KPI

Sora gives the distinction an analogy that is the article's most usable mnemonic:

**A KPI is an engine's oil pressure gauge.** It shows the situation at this moment. There can be hundreds of them in a company. And **a hundred per cent is a perfect performance.**

**An OKR aims at the future.** It is a direction you want to go in, and it may be a long way off. **If you get close, that is enough.**

Miettinen raises the risk that is the episode's sharpest question: if three of five objectives recur every quarter, do they **go stale and turn into KPIs** — so that measuring change becomes merely measuring bureaucracy?

Sora does not dispute the risk. His answer is terminological discipline: this is why it is worth being pedantic that metrics are KPIs and that in OKR we are talking about key results. **That is the only protection against the slide.**

In practice there are two traps, and Miettinen names both:
1. compensation starts getting tangled into key results, or
2. the company's **permanent metrics** start getting tangled into change metrics.

## Encouraging stretch, and why 70 per cent is enough

Because no money is attached, objectives can be set as **moonshots**. A 70 per cent outcome is then an acceptable result, if it produced meaningful change.

From which follows a limitation worth understanding before trying to use the model for appraisal:

> OKRs **are not a mechanism for evaluating a person's or a team's performance.** Because these are relative and agreed, you cannot say one reached 60 per cent and another 70, and therefore the second was better. They may be quite different situations.

They can be **part** of an appraisal. They cannot be its measure.

## Radical transparency — and its limits

The model's third structural feature is that **all objectives are visible to everyone.**

The benefit, on Sora's account, is twofold. First, you can go and see what a neighbouring team or another country organisation is doing and what matters to them — so that before a meeting you know what is worth discussing. Second, in a large organisation other people's work is often invisible, and looking at their objectives **illuminates it rather well**.

Miettinen draws a parallel with **Ray Dalio's** Bridgewater, where *radical honesty* and *radical transparency* are taken far: decisions are recorded and scored, and believability weightings are built from them. Sora does not adopt the parallel — OKR's transparency is narrower and concerns objectives, not decision-makers' credibility.

A practical detail Sora gives about leadership objectives: if a CEO or CFO can produce their own OKRs, they should — abstract work becomes **visualised** through them. Miettinen's half-serious addition: a sixth objective could be *if you came and looked, get in touch.*

## Tools

Sora is pragmatic about tools: **there are hardly any pure OKR products** — they always arrive from some other angle. He gives two examples representing opposite starting points:

- **Weekdone** (Tallinn) starts from **planning the working week**: five days is a bounded amount of time in which you should be able to promise that these things will be done by Friday — and those link upward to OKR objectives. The side benefit is a question that arises by itself: *if you put something there that does not connect to any OKR objective, is it really important?*
- **Tangible Growth**, run by the book's co-author **Juuso Hämäläinen**, starts from **managing strategy**: first the strategy and its context, then change programmes, the OKRs attached to them, and methods of digital facilitation. Sora names it his own favourite, partly because it speaks Finnish.

And what to remember before choosing a tool: the model works on a **flip chart or a whiteboard**, if the organisation is physical and people can see them. An electronic tool is the whiteboard's equivalent, not the model's precondition.

## How it connects to HR

Because money is not attached, motivation comes from elsewhere: reputation, honour, success and fulfilling a mission.

Doerr's book pairs the model with **CFRs** — conversation mechanisms attached to the same cycle, used to go through how people are doing and what they need.

Sora's most essential point about this is more general, and it applies to the whole adoption:

> When you take this model into use, **it has to be fitted to the company that already exists.**

In practice it affects development discussions, or whatever that mechanism is called in a given company — but **it must not be the sole mechanism of appraisal.** The episode mentions as examples Reaktor's development-conversation model, in which someone else follows a person's development, and Nitor's buddy system, a supporting work pair. The direction of travel is from an annual review toward **continuous tracking and support**.

## Two areas where the model gets hard

In fairness these should be known before adoption. Sora claims neither is solved.

**1. Sales.** Miettinen describes the temptation from his own work: improving sales would land straight in a partner's pocket. But then you are too close to the company's numeric metrics, which would be delivered anyway — and that is not change management.

Sora concedes it plainly: **it is one of the hardest playing fields**, and it would need proper study. His partial answer is that OKR can be used for **qualitative shared objectives** — what kind of customers we want, five of them, from which sectors — while keeping separate the thing money is paid for. On the condition that they point the same way:

> If money is paid for one and the other is a completely different thing, it is obvious which one gets attended to.

**2. Individual appraisal.** Covered above: OKR does not serve as a performance measure, because the objectives are relative and negotiated.

## Scale

Sora says plainly that the model has no size limit: **it works regardless of company size** — the difference is in how the fitting is done.

The extremes from the episode:

- **One person.** The book's translator, **Elisa Heikura**, used the model to build her own strategy and considered it a good exercise: first distil what I want, then define it and move toward it.
- **A couple of hundred people** (Ambientia): a lot of alignment work, checking together that we are heading the same way.
- **Google:** align to this, and that is that. You cannot take feedback from everyone, *or the whole quarter would be gone.*

At the individual level the model also works as a life-management tool, and Miettinen's example illustrates the objective/key-result distinction better than any corporate one: with five kilos of excess weight, the objective is not *lose five kilos* but **I want to improve my general health** — with weight, lifting capacity and whether the stairs leave you winded as the key results.

## The question left open

Miettinen asks in the episode whether OKR could create **optionality** — side projects that do not hit the target directly but from which real value emerges.

Sora's answer turns the question around, and it is honest: perhaps it is possible, but in his experience companies have **an enormous number of ideas** — and the model's benefit comes precisely from being able to say *we are only doing these*. Optionality therefore stays open; the model's demonstrated benefit is in narrowing.

---

## Three things to remember

1. **The objective is text, the key result is a number, and neither is a metric.** Terminological discipline is the only protection against OKR sliding into KPI.
2. **No money attached, no punishment for falling short.** These two conditions are not soft values but the mechanism's operating condition: without them, goal-setting becomes a budget game.
3. **At most five.** The limit is what makes this change management rather than reporting — and it also gives refusal a form.

And how Sora himself sums the whole thing up:

> If we discuss what is important and what should be done, and clarify it — is that not one of the most important things there is?

## Sources and related reading

- [OKR: objectives and key results | Henri Sora | Negotiator 58](https://ai.neuvottelija.com/ep58-okr-objectives-and-key-results-henri-sora/) — the episode this article is based on.
- Henri Sora and Juuso Hämäläinen: *Strategia arkeen OKR-mallilla*.
- John Doerr: *Measure What Matters* — the model's best-known presentation and its case stories.