---
title: "The Most Popular Competency Library Has 38 Competencies. Your 100-Person Company Needs Fewer Than 10."
description: "Enterprise competency libraries run to 38 items or more. A growing company needs fewer than 10. How to choose them, and how to spot bloat."
url: "https://sageo.ai/blog/competency-and-culture/fewer-than-10-competencies"
updated: "2026-10-05T00:00:00.000Z"
author: "Deepti Gupta"
section: "Competency and Culture"
image: "https://sageo.ai/blog/covers/fewer-than-10-competencies.png"
source: "Sageo (sageo.ai)"
---


# The Most Popular Competency Library Has 38 Competencies. Your 100-Person Company Needs Fewer Than 10.

_Enterprise competency libraries run to 38 items or more. A growing company needs fewer than 10. How to choose them, and how to spot bloat._

By Deepti Gupta.

## About This Series

This is the final post in our seven-part Competency and Culture series. Across the first six posts, we looked at why ratings drift, why results alone reward the wrong people, how values become behaviour, how hiring for "fit" narrows a company, why pay and growth need separate conversations, and how calibration keeps ratings honest. Every one of those ideas depends on the same foundation: a competency framework that managers can actually use. This post is about building that foundation, and about the most common way it goes wrong: making it too big.

Here is a quick test. Ask any manager in your company to name, from memory, the competencies they are supposed to rate their team on. Not look them up. Name them.

If they cannot, they are not using them. They are filling in a form. And a framework that lives in a form rather than in a manager's head will not change a single decision, however well it was written.

Most growing companies fail this test for the same reason. At some point, someone builds or buys a competency framework, and in the effort to be thorough, it grows. Twenty competencies. Then twenty-five, each with four levels and several indicators. It looks rigorous. In practice, it is where good intentions go to become copy-pasted comments.

## Why do competency frameworks get so big?

Because every addition feels reasonable on its own. Engineering wants technical depth covered. Sales wants commercial skills in. Leadership wants strategic thinking, communication, and stakeholder management. Someone adds resilience after a hard quarter, and innovation after an offsite. Nobody ever removes anything, because removing a competency feels like saying it does not matter.

The result is a framework designed to describe everything, which in practice guides nothing. We regularly see growing companies with 20 to 30 competencies, each with its own indicators and levels. At that size, three things happen.

1. **Managers stop reading.** Faced with dozens of items per person, managers skim, default to the middle score, and reuse comments from last cycle.
2. **Nothing stands out.** When everything is a competency, nothing is a priority. People cannot tell which behaviours actually decide their promotion.
3. **Calibration becomes impossible.** No calibration session can check the evidence behind 25 ratings per person. So it checks the overall score, and the framework underneath it is never tested at all.

## What happens when you copy a competency library built for large enterprises?

Many frameworks start from a library. One of the most widely used is the [Korn Ferry Leadership Architect](https://www.kornferry.com/capabilities/talent-suite/korn-ferry-assess/leadership-architect), which describes 38 competencies. Its predecessor, the Lominger model, had 67. These libraries are well researched and genuinely useful for what they were built for: large, complex organisations with many job families, dedicated people teams, and the capacity to assess against a long list.

A 100-person company is not that. And as we argued in [You Don't Have a Hiring Problem. You Have a Talent Density Problem](https://sageo.ai/blog/talent-density/headcount-vs-talent-density), what a growing company needs is fewer, sharper standards, not more of them. It has a handful of role families, managers who are still learning to manage, and no one whose full-time job is running assessments. Copying a 38-item library into that company is like giving a startup a large company's expense policy. Every rule is sensible. Together, they stop anything from moving.

The right way to use a library is as a menu, not a meal. Read it, pick the few competencies that genuinely separate great performance from average in your company, and leave the rest.

## Why do Amazon's Leadership Principles work, when most lists don't?

Amazon is the obvious counter-example. It has 16 Leadership Principles, after [adding two in July 2021](https://time.com/6077578/amazon-leadership-principles/), and they are among the best-known company principles in the world. Sixteen is more than we recommend. So why do they work?

Because they are used, not just published. Candidates are assessed against specific principles in interviews (we looked at Amazon's related bar-raiser idea in [The Best Hire You'll Ever Make Is Not on Any Job Board](https://sageo.ai/blog/talent-density/how-to-hire-exceptional-talent)), and interviewers are expected to bring evidence of past behaviour for each. The principles show up in how decisions are written and discussed. They are part of how the company runs, not a page in a handbook.

That is the lesson for a growing company, and it is not "copy Amazon's 16." It is that the number of competencies you can support depends on how deeply you use each one. Amazon has built years of practice, training, and interview design around its list. A 100-person company does not have that yet. Fewer competencies, used every time, beat many competencies used occasionally.

> **What this looks like in practice**
>
> for each competency in your current framework, ask one question: has it changed a single hire, promotion, or exit decision in the last year? If not, it is decoration. Cut it or merge it.

## Which competencies does a 100-person company actually need, and why fewer than 10?

The research on competency frameworks points to a consistent range: a small set of company-wide behaviours that apply to everyone, plus a small set of role competencies for each role family. Beyond that, managers stop using them. Here is the model we use.

| Layer | How many | Who it applies to | Example |
| --- | --- | --- | --- |
| Company behaviours | 3 to 5 | Everyone | Your values, written as behaviour at each level: for example ownership, candour, customer focus |
| Role competencies | 3 to 5 | Each role family | Engineering: system design, delivery quality, technical judgement |
| Leadership add-on | 1 to 2 | People managers only | Coaching, fair and evidence-based rating |

Most growing companies land at six to eight competencies per person, and should stay below ten. People managers carry one or two more, because managing people is a different job, not a bigger version of the same one.

Three rules keep the model honest.

**Company behaviours come from your values.** They are the same values from [our third post](https://sageo.ai/blog/competency-and-culture/values-fail-in-the-review), written as observable behaviour for each level. If a value cannot be written that way, it is not ready to be a competency.

**Role competencies must separate great from average.** "Writes code" is not a competency for an engineer; everyone does it. "Makes design decisions that other teams can build on" is, because it is what distinguishes your best engineers.

**Every competency needs evidence.** If a manager cannot give two examples of behaviour for a competency in a review cycle, either the competency is too vague or it does not matter in that role. Either way, it should not be on the form.

## How many judgements are you asking each manager to make?

This is the number that makes framework bloat visible. Multiply the competencies on your review form by the number of people each manager reviews. A framework with 25 competencies and a manager with 8 direct reports means 200 separate ratings in one cycle, each of which should, in theory, come with evidence.

No manager does that well. Most do it quickly, which means most of those 200 ratings are guesses.

| Framework size | Direct reports | Ratings per manager per cycle |
| --- | --- | --- |
| 25 competencies | 8 | 200 |
| 15 competencies | 8 | 120 |
| 8 competencies | 8 | 64 |

We built a simple audit to make this concrete for your own company.

_[Interactive tool: competency-bloat-audit. Available on the web version of this page.]_

> **What this looks like in practice**
>
> run the audit, then run the test from the start of this post. If your managers cannot name your competencies from memory, and the audit shows more than 100 ratings per manager, your framework is too big to work.

## How does the whole series fit together?

Seven posts, one system. Each one fixes a different point where culture leaks out of the review process.

| Post | The problem | The fix |
| --- | --- | --- |
| [1. Your reviews measure your managers](https://sageo.ai/blog/competency-and-culture/reviews-measure-managers) | Most of a rating reflects the rater | Compare ratings across raters and teams |
| [2. Your best performer might be your most expensive hire](https://sageo.ai/blog/competency-and-culture/cost-of-toxic-high-performers) | Only results get scored | Score what and how separately |
| [3. Your values fail in the review](https://sageo.ai/blog/competency-and-culture/values-fail-in-the-review) | Values are words, not behaviours | Write each value as behaviour, by level |
| [4. "Not a culture fit"](https://sageo.ai/blog/competency-and-culture/culture-fit-vs-culture-add) | Hiring filters for familiarity | Test values as behaviour; hire for what people add |
| [5. Separate the raise and the growth plan](https://sageo.ai/blog/competency-and-culture/separate-pay-and-growth-conversations) | Pay drowns out feedback | Two conversations, at different times |
| [6. Calibration is the honest part](https://sageo.ai/blog/competency-and-culture/calibration-done-right) | Ratings are never checked | Evidence-based calibration, never a quota |
| [7. Fewer than 10 competencies](https://sageo.ai/blog/competency-and-culture/fewer-than-10-competencies) | Frameworks too big to use | 3 to 5 company behaviours, 3 to 5 role competencies |

The framework is where it all starts. A small set of competencies, written as behaviour, is what makes fair ratings possible, gives the behaviour score something to measure, gives interviewers something better than "fit" to test, gives development conversations something concrete to discuss, and gives calibration something it can actually check.

The reframe underneath the whole series is simple. Culture is not what you write on the wall. It is what you measure, what you reward, and what you are willing to act on. A competency framework that managers can hold in their heads is how you make those three things the same.

## Frequently asked questions

### Why do competency frameworks get so big?

Because they are usually copied from libraries built for large enterprises, where dozens of competencies map to dozens of specialised roles. A 100-person company inherits that bulk without the scale to use it, and managers end up making far more judgements than they can make well.

### How many competencies does a growing company actually need?

Fewer than ten. Amazon runs on a compact set of Leadership Principles precisely because a short, memorable list gets used while a long one gets filed. The goal is a framework people can hold in their heads and apply consistently.

### Why does the number of competencies matter for fairness?

Because every competency multiplies the judgements each manager makes per person, and each judgement carries rater bias. Fewer, clearer competencies mean fewer places for drift to creep in, which is what makes calibration and consistency achievable.