Guides
How to Build a Skill Graph
Step-by-step tutorial for creating a practical skill graph you can use for growth planning.
This guide walks you through building a usable skill graph in under 15 minutes. By the end, you will have a structured map of your capabilities, a clear picture of your gaps, and the beginning of a concrete growth plan.
Step 1: Define Your Scope
Before listing skills, decide what part of your career this graph covers. Trying to map everything at once leads to an overwhelming, unfocused graph.
Choose one of these scopes to start:
- Your current role — map the skills you use day-to-day.
- A target role — map the skills required for the job you want next.
- A single domain — map one area deeply, like backend engineering, product design, or data science.
Starting with your current role is the easiest path. You already have evidence for these skills, and the exercise of externalising them often reveals surprising gaps.
Step 2: List Core Skills
Write down the 15–30 skills that directly impact outcomes within your chosen scope. Do not worry about structure yet — just brainstorm.
Tips for a good skill list:
- Include both hard skills (Python, system design, A/B testing) and soft skills (stakeholder management, technical writing, mentoring).
- Use verbs and outcomes to test whether something is really a skill: "Can I get better at this? Can I demonstrate it?" If yes, it belongs.
- Avoid listing tools as skills unless proficiency in the tool itself matters. "Docker" is a tool; "containerisation and orchestration" is a skill. That said, some tools (like Kubernetes or Figma) carry enough depth to be genuine skill nodes.
- If you are unsure, start with your CV or job description and extract every capability mentioned.
Common mistakes:
- Listing too few skills (under 10) — the graph will not be useful for gap analysis.
- Listing too many (over 40) — the graph becomes noisy and hard to maintain.
- Only listing technical skills — soft skills are often the gaps that block promotions.
Step 3: Group Skills by Domain
Organise your skills into 3–6 parent groups. These become the top-level structure of your graph.
Example domains for a full-stack engineer:
| Domain | Example Skills |
|---|---|
| Frontend | React, CSS architecture, accessibility, performance optimisation |
| Backend | API design, database modelling, caching, message queues |
| Infrastructure | CI/CD, containerisation, observability, cloud architecture |
| Delivery | Testing strategy, incident response, code review |
| Communication | Technical writing, stakeholder alignment, mentoring |
Example domains for a product manager:
| Domain | Example Skills |
|---|---|
| Discovery | User research, competitive analysis, opportunity sizing |
| Strategy | Roadmap planning, prioritisation frameworks, market positioning |
| Execution | Sprint planning, cross-team coordination, launch management |
| Data | Experimentation, metrics definition, SQL/analytics |
| Leadership | Stakeholder management, team building, executive communication |
There is no single correct grouping. Choose categories that feel natural for how you think about your work.
Step 4: Add Depth and Evidence
For each skill, assign a depth level. Use a consistent scale across the entire graph:
| Level | What It Means | Example Evidence |
|---|---|---|
| Exposure | You have studied or experimented with it | Completed a course, read a book, built a tutorial project |
| Working | You can deliver with guidance | Contributed to a feature, paired with a senior on a task |
| Proficient | You work independently and make decisions | Owned a feature end-to-end, made architecture choices |
| Expert | You teach, architect, and handle edge cases | Designed a system, mentored others, published content |
Evidence is what makes levels trustworthy. For each skill, attach at least one piece of evidence:
- A shipped feature or project
- A pull request or code review
- A certification or course completion
- A talk, blog post, or internal document
- A production incident you resolved
Without evidence, levels are just self-assessment. With evidence, they become verifiable claims.
Step 5: Identify Skill Gaps
Now look at your graph through the lens of your career goal. If your goal is a specific role, pull up 3–5 job descriptions for that role and highlight the skills they require. Compare against your graph.
Mark gaps in three categories:
- Missing skills — skills the target role requires that are not in your graph at all. These need to be added.
- Depth gaps — skills you have at a lower depth than the target role requires. For example, you have "working" depth in system design but the role needs "proficient."
- Evidence gaps — skills where you believe you are proficient, but you have no concrete evidence to prove it. These are common and often the easiest to fix.
Step 6: Build a 30-Day Plan
Pick 2–3 high-leverage gaps and define weekly actions to close them. Do not try to fix everything at once — focus on the gaps with the highest impact on your next career milestone.
A good 30-day plan looks like this:
| Week | Gap | Action | Evidence to Produce |
|---|---|---|---|
| 1–2 | System design (working → proficient) | Complete a system design exercise, write up the trade-off analysis | Design document with diagrams |
| 2–3 | Observability (exposure → working) | Add structured logging and alerts to an existing service | PR with before/after metrics |
| 3–4 | Technical writing (none → exposure) | Write an internal RFC or post-mortem | Published document |
After 30 days, revisit your graph, update levels and evidence, and pick the next set of gaps.
Maintaining Your Graph
A skill graph is only useful if it stays current. Here is a lightweight maintenance routine:
- Monthly review (15 minutes): Scan your graph. Update any levels that changed. Add evidence from recent work.
- After major milestones: When you ship a big project, change roles, or complete a certification, update the relevant nodes immediately.
- Quarterly planning: Every quarter, reassess your target role and adjust your gap priorities.
The goal is not perfection. It is having a good-enough map that helps you make better career decisions.
Quick Start
The fastest way to get started is with Skill Graph's AI-powered generator. Upload your CV, transcripts, or certificates, and the system extracts skills, assigns initial levels, and creates your first graph automatically. You can refine it from there.
FAQ
What is the ideal number of skills in a first version?
Start with 15–30 top-level skills across 3–6 domains. Too few (under 10) and the graph will not reveal useful patterns. Too many (over 40) and maintaining it becomes a chore. You can always expand later.
Should I include soft skills?
Absolutely. Communication, leadership, mentoring, and stakeholder management are critical for promotion and role transitions. In fact, soft skill gaps are often the hidden blockers that hold people back at senior levels.
What if I do not know my depth level?
Ask yourself: "Could I do this task without help, right now, if someone assigned it to me?" If yes with confidence, you are proficient. If yes but you would need to look things up, you are working level. If you have only studied it but never done it, that is exposure. When in doubt, rate yourself lower and attach evidence to justify upgrading later.
How long does it take to build a useful graph?
About 15 minutes for a first version. You can always refine it over time, but the initial creation should be fast — speed matters more than completeness at the start.