If your capability framework is a 60-page PDF that nobody has opened since the launch deck, you are not alone.
We have walked into more than a hundred organisations across Australia over the last decade and the pattern is almost always the same. A small team of capable L&D or HR professionals builds a beautiful taxonomy. It lands on the intranet. There is a launch lunch. And then, silence.
The framework was not wrong. It was not ambitious enough in the right places, and it was too ambitious in all the wrong ones.
Why most capability frameworks fail
The capability frameworks that die share three attributes.
They are built as documents, not as operating systems. A framework is supposed to be an interface between strategy and people decisions. If a manager cannot use it to decide who to hire, what to coach, and what to pay, it is decoration.
They chase comprehensiveness before usefulness. Teams try to name every capability that exists in the business. You end up with 180 capabilities across 7 tiers and 14 role families. Nobody can hold that in their head, let alone calibrate against it in a one-on-one.
They are divorced from workforce planning. If your capability model does not feed hiring plans, role design, and learning investment, it will get replaced the next time a new CPO or CEO arrives.
A good framework fixes all three.
The 5-part design that works
Here is the shape we use with clients. It is deliberately lean so that you can ship v1 in six weeks and iterate from there.
1. A capability taxonomy with fewer, bigger capabilities
Most frameworks want 100+ capabilities. Aim for 20 to 30. These are the capabilities that, if your people are strong in them, you win, and if they are weak, you lose.
Split them into three groups:
- Foundational (every role needs these): communication, judgement, accountability
- Professional (the craft of your function): software engineering, clinical practice, underwriting
- Strategic (the capabilities that differentiate your business): customer insight, operational design, commercial negotiation
Resist the temptation to add more. If you cannot name a business outcome that a capability drives, it does not earn a place.
2. Proficiency levels that hold up in a calibration conversation
Four levels is almost always enough: Learner, Applies, Leads, Shapes. The trap is writing level descriptors in abstract academic language. Write them in the first person:
Applies: "I can run a standard discovery interview unsupervised and produce useful notes within the same day." Leads: "I coach others through their first discovery rounds, adapt the method for ambiguous problems, and catch bias in the interview design."
If your manager cannot point at a person on their team and say "that's a Leads", your level descriptors are too abstract.
3. Role families, not individual role profiles
Do not write a capability profile for every single role title in the org chart. You will drown.
Cluster roles into 6 to 10 role families (for example: Customer, Product, Operations, Data, People, Commercial). For each family, define the core set of capabilities and the expected proficiency levels by career stage.
Individual role nuances live in the role brief, not the capability framework.
4. A connection to workforce planning
The framework should produce at least three artefacts that leaders actually use:
- A capability heatmap showing strength and gaps by team
- A hiring priority list derived from gaps
- A learning priority list derived from the same gaps
If your framework cannot answer "what are our three biggest capability risks this quarter?" it is not doing its job.
5. A light operating rhythm
Build the framework so it costs 15 minutes of a manager's time, twice a year. Not 3 hours.
- Twice-yearly calibration: managers mark their team against 5 to 7 focal capabilities for their role family. Not all capabilities at once.
- Quarterly capability pulse: leaders review the heatmap for their area and pick one gap to close in the next quarter.
- Annual framework refresh: small owner group reviews the capability list and strips anything that no longer earns its place.
That is the whole operating system.
How to ship v1 in six weeks
Week 1 to 2: Interview 10 to 15 leaders and senior ICs across the business. Ask one question: "What are the capabilities, if our people had more of them, that would change our results?" Cluster the answers.
Week 3: Draft the 20 to 30 capability list and the four-level proficiency scale. Share with a working group.
Week 4: Build out the role family model with two or three senior leaders. Test the fit against real roles.
Week 5: Run a calibration pilot with one team (10 to 20 people). Watch the conversation. Fix what is too abstract.
Week 6: Publish v1 with a forward note that says, explicitly, "this is v1, we are iterating."
The hardest part of this work is not the design. It is the discipline to ship lean, resist the completeness trap, and let the framework earn its own weight through usefulness.
A practical next step
If you want a starting structure, we built a template that mirrors the approach above, role families, proficiency levels, and a lightweight operating model in a Notion + PDF format.
Grab the Capability Framework Template and adapt it to your business. It is free and includes the calibration guide we use with clients.
Welcome to better.