# Pathways and Stackable Credentials
*Connecting credentials into routes that lead to roles and qualifications.*
## About this course
This course is for people who plan how credentials fit together. It suits learning and development leads and corporate academy leads, programme and micro-credential leads, and staff of professional bodies and sector associations. You should already have completed ICB-P01 Designing a Credential: Claim, Criteria, Evidence and Metadata, because this course treats single credentials as building blocks that you already know how to design.
The course sits in the ICoBC Credentialing Skills Framework (ICSF) at skill CRDS "Credential design" (VRITE pillar: Rigor), level L3 "Enable", stage A "Ability". At this level you design credentials and pathways for a programme, decide granularity, stacking rules and alignment targets, and guide others in using templates. At the Ability stage you practise on realistic tasks in a controlled setting. Every module therefore has a worked example and a practice activity.
**Learning outcomes**
After this course you can:
1. Design a pathway that stacks credentials toward a larger credential, role or qualification.
2. Set stacking rules, prerequisites and credit or recognition arrangements.
3. Represent the pathway in metadata so that systems can read it.
4. Check the pathway against learner workload and progression logic.
**Workload.** The course is a blended cohort course with 8 hours of notional learning. This file covers the self-paced part. Live or workshop elements appear as facilitator notes.
**Assessment.** You design a pathway of at least four credentials with stacking rules and a visual map. A person assesses it using the rubric in "Assessment and badge". A short knowledge check of ten questions accompanies it.
**Badge.** "Pathways and Stackable Credentials (Ability, L3)", alignment code ICSF-CRDS-L3-A. Evidence: the pathway map and the rules document.
**Quality focus.** VRITE Value and Rigor. Value asks whether the pathway matters to learners and employers. Rigor asks whether its rules are clear, consistent and defensible.
**AI focus.** You will use AI-generated pathway suggestions as a starting point and check them against frameworks.
| Outcome | Module | How it is assessed |
|---|---|---|
| 1. Design a pathway that stacks credentials toward a larger credential, role or qualification | Why pathways; Stacking models | Module quizzes; pathway map in the final task; knowledge check |
| 2. Set stacking rules, prerequisites and credit or recognition arrangements | Rules and recognition | Module quiz; rules document in the final task; knowledge check |
| 3. Represent the pathway in metadata so that systems can read it | Pathways in metadata | Module quiz; metadata sketch in the final task; knowledge check |
| 4. Check the pathway against learner workload and progression logic | Testing the pathway with learners | Module quiz; workload and test notes in the final task; knowledge check |
Throughout the course, two settings recur. In the first, a corporate academy at a mid-sized employer builds routes toward internal roles. In the second, a university or training provider builds routes toward a formal qualification. All names and numbers in examples are illustrative.
> A pathway is a promise to the learner: if you do these things in this order, you arrive somewhere that matters. This course shows you how to keep that promise.
---
## Why pathways
### What a pathway is
A single credential recognises one claim, for example "can facilitate a retrospective meeting". A pathway connects several credentials so that a learner can see where each one leads. The destination can be a larger credential, a role or a qualification.
A pathway has four parts:
- **Entry points:** where a learner can start, with or without prior credentials.
- **Steps:** the credentials the learner earns along the way.
- **Rules:** the conditions for moving from one step to the next.
- **Exit points:** the larger credential, role or qualification the pathway leads to, plus any useful stopping points on the way.
The word "stackable" describes credentials designed so that smaller ones count toward larger ones. Not every set of related credentials is stackable. Stackability is a design decision, and it must be stated in rules that learners, assessors and systems can read.
The EU Council Recommendation on micro-credentials (2022) describes micro-credentials as records of learning outcomes from a short learning experience, and it encourages designs that allow credentials to be combined and to build toward larger ones. It is a stable reference for this topic. Check the current text and any national transpositions that apply to you.
### Why build a pathway
Pathways create value in different ways for different parties.
**Learners** gain direction. They can plan their effort, see progress, and stop at a meaningful point without losing what they have earned.
**Employers** gain a shared language for skills. An HR team can link credentials to role profiles and succession plans.
**Providers** gain a coherent offer. Short courses become parts of a visible route instead of isolated products.
**Associations and policy bodies** gain a way to describe progression in a field, for example from assistant to practitioner to lead.
Pathways also carry risks. A pathway that exists only to sell more courses lowers trust. A pathway with unclear rules produces disputes. A pathway whose steps do not add up to its destination misleads learners. This is why Value and Rigor are the quality focus of this course. Value asks: who needs this route, and why? Rigor asks: do the rules hold up?
### Start from the destination
Design backwards. Begin with the exit point and ask three questions:
1. What does the destination require? Use a role profile, a competence framework or a qualification standard.
2. Which parts of that requirement can be recognised on their own?
3. In what order does it make sense for most learners to meet them?
For a corporate academy, the destination might be "Team lead" in the company's job architecture. For a university, it might be a postgraduate certificate. In both cases, the destination comes first and the steps follow from it.
A common mistake is to start from the courses that already exist and draw lines between them. This gives you a map of your catalogue, which is useful, but it is not yet a pathway. A pathway answers a learner's question: "How do I get there from here?"
### Value test for a pathway
Before you design in detail, run a quick value test. Write one sentence for each of the following:
- Who is the learner, and where do they start?
- What destination do they want, and who recognises it?
- Why would they take several steps instead of one large programme?
- What does the learner get at each stopping point?
If you cannot answer the third question, you may not need a stackable design. A single credential may be enough.
**Worked example: Two destinations**
*Corporate academy.* Northfield Logistics (illustrative) wants to grow warehouse staff into shift supervisors. The HR lead takes the company's "Shift supervisor" role profile and extracts five areas: safety leadership, scheduling, coaching, reporting and handling incidents. She notices that "Warehouse operative" staff with two years of service are the main entry group. The value sentence reads: "Operatives who want to become shift supervisors can earn recognised steps over about nine months, and each step is useful in the current job."
*University.* A university continuing-education unit (illustrative) wants a route toward a Postgraduate Certificate in Data Literacy. The programme lead lists the learning outcomes of the certificate and finds that three of them can stand alone as short credentials: working with data, communicating with data and ethics of data use. The value sentence reads: "Professionals can earn three short credentials over a year, and together they count toward the certificate at the university."
In both cases, the destination came first.
**Practice activity: Destination and value sentence**
Choose a destination in your own context: a role in your organisation, a larger credential or a qualification. Write down:
1. The destination and who recognises it.
2. Three to six requirements of the destination, taken from a role profile, framework or standard.
3. The learner group you have in mind and where they start.
4. A value sentence with the four parts listed above.
5. A decision: does this destination need a stackable route, or is one credential enough? Give one reason.
Keep your notes. You will build on them in every later module and in the final task.
Facilitator notes
**Live element (45 minutes, optional).** Run a destination clinic in small groups. Each participant presents their value sentence in two minutes. Peers ask one question each: "Who recognises this destination?" or "Why would a learner take steps instead of one programme?" Collect weak answers on a flip chart and return to them in module 5.
**Watch for:** destinations that are products ("our leadership programme") instead of roles or qualifications; value sentences with no named learner.
### Quiz: Why pathways
**Question 1.** Which description best fits a pathway?
[( )] Any list of courses that a provider offers on the same topic.
[(X)] A connected set of entry points, steps, rules and exit points that leads a learner to a destination.
[( )] A single credential with a long title.
[( )] A discount for learners who buy several courses.
[[?]] Think about what the learner can do with the structure.
****
A pathway is defined by its parts: entry points, steps, rules and exit points. A catalogue lists courses without telling the learner how to move through them. A pathway answers the question "How do I get there from here?"
****
**Question 2.** Why does the design process in this course start from the destination?
[( )] Because the destination is always the cheapest part to build.
[( )] Because existing courses should be deleted first.
[(X)] Because the requirements of the role, credential or qualification decide which steps are needed and in what order.
[( )] Because learners never need to see the steps.
[[?]] Consider what happens if you only connect the courses you already have.
****
Designing backwards from the destination ensures that the steps add up to something meaningful. Starting from the existing catalogue produces a map of what you have, not a route to what the learner needs.
****
**Question 3.** A provider cannot say why learners should take three small credentials instead of one large programme. What is the most sensible conclusion?
[( )] Add a fourth credential to make the pathway look bigger.
[( )] Remove all rules so learners can choose freely.
[(X)] Reconsider whether a stackable design is needed, because a single credential may serve the learner better.
[( )] Rename the credentials as "modules".
[[?]] Stackability is a design decision that needs a reason.
****
The value test asks why a stackable design helps the learner. If there is no answer, forcing a pathway lowers trust. A single credential may be the better choice.
****
**Question 4.** Which pair of parties benefits most directly from a shared route from credentials to role profiles?
[( )] Printers and couriers.
[( )] Search engines and advertisers.
[(X)] Employers and learners.
[( )] Only the provider's finance department.
[[?]] Think about who needs a shared language for skills.
****
Employers can link credentials to role profiles and succession plans, and learners can see how their effort connects to a destination. Providers and associations also benefit, but the direct link is between employers and learners.
****
---
## Stacking models
### Granularity first
Before you choose a stacking model, decide how big each credential should be. Granularity is the size of the claim a credential makes. A credential that is too large cannot be earned in a sensible time, and a credential that is too small adds administrative burden without meaning.
Use three checks:
- **Meaning:** does the credential state a claim that someone else can understand and use, such as a manager reading a profile?
- **Evidence:** can you collect evidence for the claim at reasonable cost?
- **Effort:** is the learner effort in a range your learners can plan around? Many providers use units that fit into a few weeks of part-time study, but your context decides.
In ICB-P01 you designed one credential with a claim, criteria, evidence and metadata. Here you treat each such credential as a building block and ask how big the blocks should be and how they fit together.
### Four common stacking models
Most pathways use one of four models, or a combination.
**1. Linear stack.** Credentials are earned in a fixed order: A, then B, then C. Each builds on the previous one. This model suits fields where later knowledge depends on earlier knowledge. Its weakness is rigidity.
**2. Core plus electives.** Learners earn a required core and then choose from a set of options. This model suits roles with a shared base and several specialisations. Its weakness is complexity: you must state how many electives are needed and from which groups.
**3. Free combination.** Learners collect any credentials from a list until they reach a total, for example any four from eight. This model gives choice. Its weakness is that the result may not add up to a coherent whole.
**4. Layered levels.** Credentials sit at levels, for example foundation, practitioner and advanced, and a learner moves up. This model suits fields with a recognised ladder of competence. Its weakness is that levels must be defined with care so that they differ in more than name.
You can mix models. A pathway can use a linear foundation, a core, and electives at the top.
### Choosing alignment targets
Alignment targets tell others what your credentials relate to. Typical targets include:
- a role profile or job architecture in an organisation;
- a competence framework from a sector body;
- a qualification standard or level in a national or regional framework;
- the credit system of a provider or institution.
At L3 you decide the alignment targets for the pathway and record them. For European contexts, the European Learning Model and European Digital Credentials for Learning offer a shared way to describe learning outcomes, credentials and their relation to frameworks. They are in use across institutions in Europe. Check which national or institutional profiles apply to you.
Choose targets that your audience recognises. A target nobody uses adds no value, even if it is technically correct.
### Exit points and stopping points
A good pathway lets learners stop at several points and still hold something useful. Mark these on your design:
- **Intermediate exit:** a credential that has value in the workplace or in further study on its own.
- **Final exit:** the larger credential, role readiness or qualification.
- **Side exits:** other pathways a learner could join using credits or credentials already earned.
Each stopping point should hold a claim you can defend without the rest of the pathway.
**Worked example: Two stacking models**
*Corporate academy.* At Northfield Logistics (illustrative), the academy lead designs the shift supervisor route as core plus electives. Core credentials: "Safety leadership on shift" and "Scheduling a shift". Electives, of which the learner chooses one: "Coaching operatives" or "Handling incidents and escalation". Final credential: "Shift supervisor readiness". The alignment target is the company's "Shift supervisor" role profile. The intermediate exit "Safety leadership on shift" is useful to experienced operatives who act as deputy.
*University.* The continuing-education unit (illustrative) designs the data literacy route as a linear stack with an optional layer. "Working with data" comes first, then "Communicating with data", then "Ethics of data use". The final exit is the Postgraduate Certificate in Data Literacy. The alignment target is the qualification's published learning outcomes. A learner who stops after the second credential holds two recognised credentials but not the certificate.
### AI-generated suggestions as a starting point
An AI assistant can propose a pathway from a destination: a list of credentials, an order and suggested titles. This can save time in the first draft. Treat the output as a hypothesis, not as a design.
Check every suggestion against sources you trust:
- Does each credential map to a real requirement in the role profile, framework or standard you chose?
- Are the proposed sizes realistic for your learners?
- Did the assistant invent framework levels, credit values or codes? Verify each against the official text.
- Are the titles distinct and clear, or generic?
- Do the claims describe what a learner can do, and avoid overpromising?
Keep a short log of what you accepted, changed and rejected, and why. The log is part of your rigour: it shows that a person decided.
**Practice activity: Choose a model and size the blocks**
Using your destination from module 1:
1. List the requirements again and group them into four to six candidate credentials.
2. For each, write the claim in one sentence and estimate the learner effort in hours (label your numbers as estimates).
3. Choose a stacking model: linear, core plus electives, free combination, layered levels, or a mix. Give one reason.
4. Optional: ask an AI assistant to propose a pathway for your destination. Compare it with your draft. Note two things you accept and two you reject, each with a reason that refers to a source.
5. Mark at least one intermediate exit and the final exit, and name your alignment targets.
Facilitator notes
**Live element (60 minutes, optional).** Hold a model market. Each participant sketches their draft on one sheet. Pairs swap sheets and try to name the stacking model without being told. If the pair cannot, the model is unclear.
**AI comparison.** Ask participants to show one AI suggestion they rejected and say which source made them reject it. Praise specific citations of frameworks, not general doubts.
**Watch for:** hidden dependencies between "free" credentials; intermediate exits that carry no claim of their own.
### Quiz: Stacking models
**Question 1.** A role has a shared base of two skills and three possible specialisations. Which model fits best?
[( )] Free combination of any two credentials from a list.
[( )] A single large credential with no steps.
[(X)] Core plus electives.
[( )] Layered levels with a foundation, practitioner and advanced credential.
[[?]] Look for the model with a required part and a choice part.
****
Core plus electives gives every learner the shared base and then lets them choose a specialisation. Layered levels describe a ladder of competence, which is not what the role describes.
****
**Question 2.** Which question helps you test the granularity of a credential?
[( )] Is the title longer than ten words?
[( )] Does the credential have a colourful image?
[(X)] Can someone else understand and use the claim, can you collect evidence for it at reasonable cost, and can learners plan the effort?
[( )] Does the credential cost more than the one before it?
[[?]] The three checks are meaning, evidence and effort.
****
Meaning, evidence and effort are the checks used in this course. Title length, imagery and price do not tell you whether the size of the claim is sensible.
****
**Question 3.** An AI assistant proposes a pathway and states a framework level for each credential. What should you do first?
[( )] Publish the levels, because the assistant sounds confident.
[(X)] Verify each level against the official text of the framework and record what you accepted, changed or rejected.
[( )] Delete all framework references to avoid errors.
[( )] Ask the assistant to repeat the answer until it agrees with itself.
[[?]] AI output is a hypothesis.
****
Framework levels and codes are easy for an assistant to invent or misstate. Checking them against the official source, and logging the decision, keeps a person accountable for the design.
****
**Question 4.** What is an intermediate exit?
[( )] A credential that is withdrawn from the catalogue.
[( )] A rule that stops learners from leaving.
[( )] The final credential of the pathway.
[(X)] A credential in the middle of the pathway that has value on its own, so a learner who stops there still holds something useful.
[[?]] Think of a stopping point that still means something.
****
Intermediate exits protect learners who cannot finish the pathway. Each one needs a claim that stands without the rest of the route.
****
---
## Rules and recognition
### Four kinds of rules
Stacking rules turn a collection of credentials into a pathway. Write them so that a learner, an assessor and a system can all apply them in the same way. Most rule sets contain four kinds of rule.
**Prerequisites.** What a learner must hold or show before starting a credential. A prerequisite can be a credential, a period of work experience or a placement test. Use prerequisites only where later learning depends on earlier learning. Unneeded prerequisites exclude people without good reason.
**Combination rules.** Which credentials, and how many, add up to the larger credential. Examples: "all four core credentials", "two core and one elective from group B", "any three from the list".
**Time rules.** How long credentials stay valid for stacking. A credential earned six years ago may no longer reflect current practice in a fast-moving field. If you set a window, state it and the reason. If you set none, say so.
**Level and quality rules.** The grade, level or outcome needed for a credential to count. For example, "Pass or higher" or "assessed against the same criteria as the target qualification".
Write each rule as a short, testable sentence. A good test is: could two different assessors read the rule and reach the same decision?
### Credit and recognition arrangements
Credits and recognition decide what happens when a learner brings credentials from outside your pathway or wants to move to another one.
**Credit** is a unit of value that a provider or system assigns to learning, often linked to notional hours. If you use credits, state the basis and the system. Many institutions use a formal credit system and have rules about how much credit a qualification can receive from earlier learning. Check your own institution's regulations and, for European higher education, the relevant credit system documents.
**Recognition** is a decision that a credential or prior learning counts toward a step in your pathway. Typical arrangements:
- **Direct recognition:** your own credentials count automatically.
- **Agreed equivalence:** a named external credential counts, based on a comparison of claims, criteria and evidence.
- **Individual assessment:** a learner submits evidence of prior learning, and a person decides.
- **No recognition:** the step must be earned in the pathway. State the reason, for example a regulated requirement.
Recognition decisions should be based on the claim and the evidence, not on the brand of the issuer. Record who decided, on what basis and when.
### Fairness and access
Rules shape who can progress. Check each rule for hidden barriers:
- Does a prerequisite require a paid course when equivalent experience would do?
- Does a time rule disadvantage people who took a career break?
- Is an appeals route available when a recognition decision is refused?
Rigor includes fairness. A strict rule is good only if the reason for it is clear and the decision can be explained.
### The rules document
Collect your rules in one document with a fixed structure. This is part of your assessed evidence. Include:
1. Purpose and destination of the pathway.
2. List of credentials with a one-line claim each.
3. Prerequisites for each credential.
4. Combination rules for the larger credential.
5. Time rules, or a statement that none apply.
6. Level and quality rules.
7. Credit arrangements, if any.
8. Recognition arrangements, including individual assessment and appeals.
9. Owner of the rules and review date.
Keep it short enough that a learner will read it. A rules document that nobody reads protects nobody.
**Worked example: Rules for two pathways**
*Corporate academy.* For the shift supervisor route (illustrative), the academy lead writes:
- Prerequisite for "Shift supervisor readiness": both core credentials and one elective.
- Prerequisite for "Scheduling a shift": none, but six months in a warehouse role is recommended.
- Time rule: credentials count for stacking for three years, because company safety procedures change.
- Recognition: a named external first-aid credential counts toward the "Safety leadership on shift" credential through agreed equivalence. An individual assessment is available for staff with prior supervisory experience, decided by a panel of two supervisors and one HR partner.
- Appeals: a learner can ask HR for a second review within 20 working days.
*University.* For the data literacy route (illustrative), the programme lead writes:
- Prerequisite: none for the first credential, which is open to any applicant with a first degree or equivalent professional experience.
- Combination rule: all three credentials are required for the Postgraduate Certificate.
- Time rule: credentials count for five years, in line with the unit's regulations.
- Credit: each credential carries credits in the institution's credit system, and the total matches the certificate. Numbers are set by the institution.
- Recognition: individual assessment of prior learning is allowed for one of the three credentials. The decision follows the institution's usual recognition procedure.
**Practice activity: Draft the rules document**
Take your pathway draft from module 2.
1. Write a prerequisite for each credential, or write "none" and give a reason.
2. Write the combination rule for the larger credential in one testable sentence.
3. Decide on a time rule and give the reason, or state that none applies.
4. Write one recognition arrangement of each kind that applies: direct, equivalence, individual assessment, none.
5. Add an appeals route.
6. Run the two-assessor test: ask a colleague to read your rules and decide on two imaginary learner cases. Note where they disagreed.
7. Revise the rules until the colleague reaches the same decisions as you.
Facilitator notes
**Live element (60 minutes, optional).** Prepare four learner cases on cards, for example: a learner with an external credential, a learner returning after a career break, a learner who failed one credential, and a learner with ten years of experience and no credentials. Pairs swap rules documents and decide each case using only the written rules. Debrief on ambiguity.
**Watch for:** rules that rely on unwritten practice; recognition by issuer brand; missing appeals routes.
### Quiz: Rules and recognition
**Question 1.** Which rule is a good example of a testable combination rule?
[( )] Learners should complete most of the credentials.
[( )] Learners must show enthusiasm during the pathway.
[(X)] Learners need both core credentials and one elective from group B.
[( )] Learners may complete the pathway when the provider agrees.
[[?]] Could two assessors reach the same decision from the sentence?
****
A testable rule names what is required and how many. Vague wording such as "most" or "when the provider agrees" lets different assessors decide differently.
****
**Question 2.** On what should a recognition decision be based?
[( )] The reputation of the issuing organisation only.
[( )] The price the learner paid for the credential.
[(X)] The claim, criteria and evidence of the credential compared with the step it would replace.
[( )] The length of the credential title.
[[?]] Look at the content of the credential, not its brand.
****
Recognition compares what the credential claims and how it was assessed with the requirements of the step. Brand and price do not show whether the learning matches.
****
**Question 3.** Why should a time rule include a reason?
[( )] Reasons make the document longer and more official.
[( )] Time rules only apply to university credentials.
[(X)] A stated reason, such as changing procedures, lets learners and assessors see why the window exists and review it later.
[( )] A reason makes the rule optional.
[[?]] Think about how someone can challenge or review the rule.
****
A time rule without a reason looks arbitrary and is hard to review. A stated reason supports fairness and makes it easier to update the rule when conditions change.
****
**Question 4.** A learner is refused recognition for prior learning. What should the rules provide?
[( )] Nothing, because the decision is final.
[( )] A new course to buy.
[( )] A note saying the learner may try again in ten years.
[(X)] An appeals or review route, with the decision, the basis and the date recorded.
[[?]] Fairness includes the right to ask for a second look.
****
An appeals route and a record of who decided and why make the process fair and auditable. Without them, recognition decisions cannot be explained.
****
---
## Pathways in metadata
### Why systems need to read the pathway
A pathway map helps people. Systems need the same information in structured form. Wallets, learner records, skills platforms and registries can only show a route if the credentials describe how they connect.
Metadata lets a system answer questions such as:
- Which credentials are part of this pathway?
- What does this credential lead to?
- What must a learner hold before starting it?
- Which framework or qualification does it align to?
In ICB-P01 you described a single credential with a claim, criteria, evidence and metadata. For a pathway you add relationships between credentials. Each credential keeps its own record, and the relationships connect the records.
### What to record
At minimum, record these items for each credential in the pathway:
- **Identifier:** a stable, unique identifier for the credential.
- **Title and claim:** the name and a short statement of what the earner can do.
- **Part of:** the pathway or larger credential this credential contributes to.
- **Requires:** credentials or conditions needed before earning it.
- **Leads to:** the credentials, roles or qualifications it counts toward.
- **Alignment:** links to frameworks, standards or role profiles, with the level where one applies.
- **Credit or notional effort:** hours or credits if used, with the system named.
- **Validity for stacking:** the time rule from your rules document.
For the larger credential, record the rule that combines the parts, for example "all of A and B, and one of C or D".
### Using shared models
You do not need to invent your own structure. Shared data models exist for this purpose.
The European Learning Model (ELM) is a data model for describing learning opportunities, qualifications, learning outcomes, assessment and related information. European Digital Credentials for Learning use it to issue credentials that wallets and systems can exchange. Institutions across Europe use them. The ELM also served as the data model for education credentials in the DC4EU wallet pilot. Check the current specifications and your national or institutional profiles.
Other credential standards from standards bodies also describe achievements and relationships between them. Choose the model your issuing platform and your audience use. Vendor-neutral means describing the data, not tying the design to one product.
If your platform does not support pathway relationships fully, keep the rules in a published document and link to it from each credential. A human-readable fallback is better than missing information.
### A simple structured sketch
The following sketch is an illustration of the information to record, written as plain key and value pairs. It is not a file format of any standard. Map the fields to the model you use.
```
credential_id: nf-safety-leadership-001
title: Safety leadership on shift
claim: Leads a shift safety briefing and responds to hazards
part_of: nf-shift-supervisor-readiness-001
requires: none
leads_to: nf-shift-supervisor-readiness-001
alignment: Northfield role profile "Shift supervisor", area 1
effort: 20 hours (illustrative)
valid_for_stack: 3 years
```
Keep identifiers stable. If you rename a credential, keep its identifier. Changing identifiers breaks links in learner records.
### Keeping map and metadata in step
The visual map, the rules document and the metadata describe the same pathway. They must agree. When you change one, check the others. A simple habit helps: keep one table that lists every credential, its prerequisites, its place in the pathway and its identifier, and generate or check the other items from it.
**Worked example: Metadata for a pathway**
*Corporate academy.* The academy lead at Northfield Logistics (illustrative) builds a table with five rows, one per credential. Each row has an identifier, a claim, "part of", "requires" and the alignment to the role profile. For the larger credential she records the combination rule: "both core credentials and one of the two electives". She publishes the rules document at a stable address inside the company intranet and links to it from each credential. The wallet used by the company shows the credential but not the full route, so the link serves as a fallback.
*University.* The programme lead at the continuing-education unit (illustrative) records the three short credentials and the certificate using the institution's chosen learning model. Each short credential states that it is part of the certificate. The certificate records the rule "all three required" and its alignment to the published qualification outcomes. The registrar checks that the identifiers match those in the student record system.
**Practice activity: Write the metadata table**
For the pathway you have designed so far:
1. Create a table with one row per credential and one row for the larger credential.
2. Fill in identifier, title, claim, part of, requires, leads to, alignment and effort.
3. Record the combination rule on the larger credential's row.
4. Choose a data model your platform or audience uses, and note which of your fields map to which parts of it. Where there is no match, write "link to rules document".
5. Check consistency: does every "requires" in the table match a prerequisite in your rules document and an arrow on your map?
6. Optional: ask an AI assistant to convert your table into a structured sketch. Check each field against your table and against the official model documentation.
Facilitator notes
**Live element (45 minutes, optional).** Invite a colleague from the platform or IT team to join a short "can the system read this?" session. Participants show their tables and the colleague says what the platform can store and what must stay in a document. Record gaps as requests.
**Watch for:** identifiers that change with titles; "requires" fields that contradict the rules; AI-generated structures that use field names from no standard.
### Quiz: Pathways in metadata
**Question 1.** Why does a pathway need relationships in metadata as well as a visual map?
[( )] Because maps cannot be printed.
[(X)] Because systems such as wallets and learner records can only show a route if the credentials describe how they connect.
[( )] Because metadata replaces the rules document.
[( )] Because learners prefer reading data fields to maps.
[[?]] Think about who or what reads each format.
****
A map serves people. Systems need structured relationships such as "part of", "requires" and "leads to" to display and check pathways. The rules document and map remain necessary for people.
****
**Question 2.** Which field tells a system what must be held before starting a credential?
[( )] Alignment.
[( )] Effort.
[(X)] Requires.
[( )] Title.
[[?]] Look for the prerequisite relationship.
****
"Requires" records the credentials or conditions needed before earning the credential. Alignment links to frameworks, effort records time or credits, and title names the credential.
****
**Question 3.** A learner platform cannot store pathway relationships. What is the most sensible approach?
[( )] Leave the pathway undocumented.
[( )] Rename the credentials so the route is obvious.
[(X)] Publish the rules and map at a stable address and link to them from each credential.
[( )] Remove the prerequisites.
[[?]] A human-readable fallback is better than missing information.
****
When structured relationships are not possible, a stable published document linked from each credential keeps the pathway visible. Removing prerequisites would change the design to fit the tool.
****
**Question 4.** An AI assistant produces a structured record with field names you do not recognise. What should you do?
[( )] Use the record, because it looks tidy.
[(X)] Compare each field with the official documentation of the data model you use and with your own table, and correct or remove what does not match.
[( )] Ask the assistant to add more fields.
[( )] Copy the field names into your rules document.
[[?]] AI output may use names that belong to no standard.
****
Assistants can invent field names or misplace fields. Checking against the official model and your own table keeps the metadata accurate and readable by systems.
****
---
## Testing the pathway with learners
### Why test
A pathway can look sound on paper and still fail in practice. Testing with real or realistic learners shows where the design breaks. Rigor includes evidence that the design works. Value includes learners saying that the route makes sense.
Test four things:
1. **Workload:** can learners complete each step and the whole pathway in the time you expect, alongside work or study?
2. **Progression logic:** does each step prepare learners for the next one, and is the order sensible?
3. **Rules in use:** can learners and assessors apply the rules and reach the same decisions?
4. **Value:** do learners and employers see the intermediate and final exits as useful?
### Checking workload
Add up the notional effort of every credential. Compare it with the time your learners actually have.
- Convert effort into weekly hours over the planned duration. For example, 80 hours over 20 weeks is 4 hours a week (illustrative).
- Consider peak loads: if two credentials run at the same time, add them together.
- Count assessment effort, not only learning time.
- Look at who your learners are. Shift workers, carers and people in full-time jobs have different limits.
Ask a small group to track real time for one credential. Compare the tracked time with your estimate. If the difference is large, revise the estimate, the size of the credential or the schedule.
### Checking progression logic
Take each arrow in your map and ask: what does the learner need from the previous step to succeed in the next? Look for:
- **Gaps:** a later credential assumes knowledge no earlier credential teaches.
- **Overlaps:** two credentials teach the same thing, wasting learner time.
- **Dead ends:** a credential that leads nowhere and has no standalone value.
- **Bottlenecks:** one credential that every learner must pass but that is offered rarely.
A simple method is a walkthrough with a persona. Write a short learner profile, then follow that person step by step through the pathway, using only the rules as written.
### Using learner and employer feedback
Collect feedback in a light, structured way:
- A short walkthrough interview with three to five learners from different starting points.
- A review with one or two employers, line managers or programme partners who know the destination.
- A review of the map and rules by someone who did not design them.
Ask concrete questions: "Where would you start?", "Which step would you skip, and why?", "What would you tell a colleague this pathway gives them?" Record answers and decide on changes. Keep a change log with the reason for each change. Do not claim that completing the pathway proves competency in a job. State what the credentials claim and how they are assessed.
### AI in testing
An AI assistant can help you draft learner personas, generate edge cases for your rules, or simulate questions a learner might ask. This is useful for finding weak points early. Simulated learners are not real learners. Use AI output to prepare real tests, not to replace them. Check any claims it makes about frameworks or regulations against official sources.
**Worked example: Testing two pathways**
*Corporate academy.* The academy lead at Northfield Logistics (illustrative) runs a walkthrough with four operatives from different shifts. She adds up effort: five credentials at about 20 hours each, 100 hours in total, planned over 10 months. That is about 2.5 hours a week. Two operatives work night shifts and say they cannot study during the week of rota changes. She adds a rule that a learner may pause for up to eight weeks without losing progress. A supervisor notes that "Scheduling a shift" assumes knowledge of the company's rota tool, which is taught only in an elective. She moves a short introduction into the core credential to close the gap.
*University.* The programme lead at the continuing-education unit (illustrative) walks a part-time student persona through the three credentials. Each takes about 60 hours (illustrative), and the plan runs over 12 months, about 3.5 hours a week. A learner with a full-time job says two credentials overlap in the same teaching weeks. The lead changes the schedule so that no learner needs to study two credentials in parallel. A review with an employer partner shows that the second credential is the one they value most as a standalone item, so the lead marks it as a named intermediate exit.
**Practice activity: Test your pathway**
1. Add up the effort of all credentials in your pathway and convert it to weekly hours over your planned duration. Compare it with a realistic weekly limit for your learners.
2. Write two short learner personas with different starting points. Walk each persona through the pathway using only your written rules. Note every point where the rules did not give a clear answer.
3. Check each arrow on your map for gaps, overlaps, dead ends and bottlenecks.
4. Ask one colleague or learner to review the map and rules. Record three comments.
5. Write a change log with at least three changes and a reason for each.
6. Optional: use an AI assistant to generate two edge cases for your rules. Test them, and note whether they revealed a real weakness.
Facilitator notes
**Live element (90 minutes, workshop).** Run a pathway walkthrough lab. Each participant presents their map in five minutes. The group plays two personas: a learner with limited time and an employer asking "What can this person do after step two?" Capture issues on cards sorted into workload, logic, rules and value. Close with each participant naming the one change they will make first.
**Assessment preparation.** Remind participants of the final task, the rubric and the evidence list. Offer a short Q and A on the rubric.
**Watch for:** workload estimates with no learner input; claims that completing the pathway proves job competence; testing with only enthusiastic volunteers.
### Quiz: Testing the pathway with learners
**Question 1.** A pathway has 100 hours of notional effort planned over 20 weeks. What is the average weekly load?
[( )] 2 hours.
[( )] 10 hours.
[( )] 20 hours.
[(X)] 5 hours.
[[?]] Divide the total hours by the number of weeks.
****
100 hours divided by 20 weeks is 5 hours a week. Remember to add peaks where two credentials run together and to count assessment effort.
****
**Question 2.** A later credential assumes knowledge that no earlier credential teaches. What is this problem called in this course?
[( )] An overlap.
[(X)] A gap.
[( )] A bottleneck.
[( )] A dead end.
[[?]] Something is missing between two steps.
****
A gap means the progression logic is broken: learners arrive at a step without something they need. Overlaps are duplicated content, bottlenecks are rarely offered steps, and dead ends are credentials that lead nowhere.
****
**Question 3.** What is the best way to find out whether two assessors would apply your rules in the same way?
[( )] Ask the rule author whether the rules seem clear.
[( )] Count the words in the rules document.
[(X)] Give both assessors the same learner cases and the written rules, and compare their decisions.
[( )] Publish the rules and wait for complaints.
[[?]] Test the rules in use.
****
Comparing decisions on the same cases shows ambiguity directly. Author opinion, word counts and waiting for complaints do not give early evidence.
****
**Question 4.** How should you use AI-generated learner personas and edge cases?
[( )] As a full replacement for real learner testing.
[( )] As proof that the pathway works.
[(X)] To prepare real tests and find weak points early, while checking any claims about frameworks or rules against official sources.
[( )] Only after the pathway is published.
[[?]] Simulated learners are not real learners.
****
AI can help draft personas and edge cases. Real learners and employers are still needed to test workload, logic and value, and framework claims need checking against official sources.
****
---
## Assessment and badge
### The final task
**Task: Design a stackable pathway.**
Design a pathway of at least four credentials that leads to a larger credential, role or qualification in your own context (corporate academy, HR department, university, college, training provider, professional body or sector association). Use the work from the practice activities in this course.
Submit two items.
**1. Pathway map (visual).** One page, showing:
- entry points;
- at least four credentials with titles;
- the stacking model, with arrows or groupings that show order, choice and combinations;
- at least one intermediate exit and the final exit;
- alignment targets (role profile, framework or qualification);
- effort per credential, labelled as an estimate where it is one.
**2. Rules document.** Two to four pages, following the structure in module 3, and including:
- purpose and destination, with the value sentence;
- credential list with one-line claims;
- prerequisites, combination rules, time rules, level and quality rules;
- credit and recognition arrangements, including individual assessment and appeals;
- a metadata table following module 4, with identifiers, "part of", "requires", "leads to" and alignment;
- a workload and progression check following module 5, with a change log of at least three changes;
- if you used an AI assistant: a short log stating what you accepted, changed and rejected, and which sources you used to check it.
Do not claim that completing the pathway proves competency in a job. State what each credential claims and how it is assessed.
A person assesses your submission using the rubric below. The file does not grade the task.
Facilitator notes
**Assessor guidance.** Assess against the rubric and the evidence list. Give written feedback for each criterion. A participant who does not reach the pass rule may revise and resubmit once, using the feedback. Check that the metadata table, the rules document and the map agree with each other. Where a participant used AI-generated suggestions, check the log and spot-check at least two claims against the cited framework.
**Live element (optional, 60 minutes).** Hold a peer review round before submission. Pairs use the rubric to review each other's drafts and give two strengths and two improvements each.
### Rubric
Each criterion is rated at one of three levels: Not yet (0), Meets (1) or Strong (2).
| Criterion | Not yet (0) | Meets (1) | Strong (2) |
|---|---|---|---|
| 1. Pathway design (Outcome 1) | Fewer than four credentials, or the credentials do not lead to a stated destination. | At least four credentials lead to a stated destination, with a named stacking model and at least one intermediate exit. | As "Meets", and the design starts from the destination's requirements, the stacking model is justified, and each exit holds a claim that stands alone. |
| 2. Value | No stated learner, destination or reason for stacking. | A value sentence names the learner, the destination, who recognises it and why steps help. | As "Meets", and the value is supported by evidence from a role profile, framework, employer or learner input. |
| 3. Stacking rules and prerequisites (Outcome 2) | Rules are missing, vague or contradict each other. | Prerequisites, combination rules, time rules and level rules are written as testable sentences. | As "Meets", and a colleague reaches the same decisions on test cases, with reasons given for each rule. |
| 4. Credit and recognition (Outcome 2) | No recognition arrangement, or recognition depends on issuer brand. | At least two recognition arrangements are stated, based on claim and evidence, with an appeals route. | As "Meets", and decisions are recorded with who, basis and date, and barriers to fairness and access are addressed. |
| 5. Metadata (Outcome 3) | Metadata table missing or inconsistent with the rules. | A table lists identifiers, "part of", "requires", "leads to" and alignment for every credential, and matches the map and rules. | As "Meets", and fields are mapped to a named data model with a fallback where the platform cannot store relationships. |
| 6. Workload and progression check (Outcome 4) | No workload calculation or progression check. | Weekly workload is calculated and compared with a learner limit, and gaps, overlaps, dead ends and bottlenecks are checked. | As "Meets", and learner or employer input is recorded and leads to at least three reasoned changes in a change log. |
| 7. Map clarity | The map is hard to read, lacks labels or relies on colour alone. | The map is readable, labelled, and shows entry points, steps and exits. | As "Meets", and the map is accessible without colour, uses consistent labels and matches the rules exactly. |
| 8. Responsible AI use and claims | AI output is used without checking, or the submission claims proof of job competence. | Any AI use is logged, and claims are checked against official sources. No claim of proven job competence. | As "Meets", and the log shows specific corrections made after checking against frameworks. |
### Pass rule
- Total of at least 10 points out of 16, with no criterion rated "Not yet" in criteria 1, 3 and 5.
- A score of "Not yet" in any other criterion needs a written action in the feedback.
- A participant who does not pass may revise and resubmit once.
- Both evidence items must be submitted.
### Badge
On passing, you receive the badge "Pathways and Stackable Credentials (Ability, L3)", alignment code ICSF-CRDS-L3-A. The badge states that you practised designing a pathway of stackable credentials in a controlled setting against a rubric. It does not state that you are competent in any particular job.
**Badge evidence:** pathway map; rules document.
### Final knowledge check
Answer all ten questions. The check covers all four learning outcomes.
**Question 1.** (Outcome 1) Which design step comes first when building a pathway?
[( )] Draw arrows between the courses you already have.
[( )] Choose the badge images.
[(X)] Define the destination and its requirements from a role profile, framework or qualification.
[( )] Set the price of each credential.
[[?]] Design backwards.
****
Starting from the destination ensures the steps add up to something meaningful. Existing courses and visual choices come after.
****
**Question 2.** (Outcome 1) A role has a shared base and three specialisations. Which model fits best?
[( )] Linear stack.
[(X)] Core plus electives.
[( )] Free combination of any two credentials.
[( )] A single credential.
[[?]] Look for a required part and a choice part.
****
Core plus electives gives all learners the base and lets them choose a specialisation.
****
**Question 3.** (Outcome 1) Why mark an intermediate exit on a pathway?
[( )] To make the pathway look longer.
[(X)] So that learners who stop at that point still hold a credential with its own claim and value.
[( )] To prevent learners from finishing.
[( )] To replace the final credential.
[[?]] Think of the learner who cannot finish.
****
Intermediate exits protect learners and give employers a meaningful claim at each stopping point.
****
**Question 4.** (Outcome 2) Which sentence is a testable combination rule?
[( )] Most core credentials should be completed.
[(X)] Learners need both core credentials and one elective from group B.
[( )] Learners may proceed when the provider is satisfied.
[( )] Learners should show commitment.
[[?]] Two assessors must reach the same decision.
****
The testable rule states exactly what is required. The others are vague.
****
**Question 5.** (Outcome 2) On what should a recognition decision for an external credential rest?
[( )] The issuer's reputation alone.
[(X)] A comparison of the external credential's claim, criteria and evidence with the step it would replace.
[( )] The price the learner paid.
[( )] The learner's job title.
[[?]] Compare content, not brand.
****
Recognition rests on whether the learning matches the step. Reputation, price and job title do not show that.
****
**Question 6.** (Outcome 2) What should the rules provide when recognition is refused?
[( )] No further route.
[(X)] An appeals or review route, with the decision, basis and date recorded.
[( )] An invitation to buy a course.
[( )] A reduction in the learner's credits.
[[?]] Fairness includes a second look.
****
An appeals route and a record make the decision explainable and fair.
****
**Question 7.** (Outcome 3) Which field tells a system which credential must be earned before another?
[( )] Alignment.
[(X)] Requires.
[( )] Effort.
[( )] Title.
[[?]] The prerequisite relationship.
****
"Requires" records prerequisites. Alignment links to frameworks, effort records hours or credits and title names the credential.
****
**Question 8.** (Outcome 3) A platform cannot store pathway relationships. What is the best approach?
[( )] Leave the relationships out.
[(X)] Publish the rules and map at a stable address and link to them from each credential.
[( )] Remove prerequisites from the design.
[( )] Change identifiers each time you update the map.
[[?]] Keep the information visible.
****
A linked, published document keeps the pathway readable until the platform can store relationships. Stable identifiers should be kept.
****
**Question 9.** (Outcome 4) A pathway has 60 hours of effort over 12 weeks. What is the average weekly load?
[( )] 3 hours.
[(X)] 5 hours.
[( )] 6 hours.
[( )] 12 hours.
[[?]] Divide hours by weeks.
****
60 hours divided by 12 weeks is 5 hours a week. Add peaks and assessment effort when you compare with the learner limit.
****
**Question 10.** (Outcome 4) You test your pathway with a persona and find that a later credential assumes knowledge no earlier step teaches. What is the right response?
[( )] Ignore it, because learners will cope.
[( )] Remove the later credential.
[(X)] Close the gap by adding the missing content to an earlier step or adding a prerequisite, and record the change with its reason.
[( )] Rename the later credential.
[[?]] Fix the progression logic and keep a log.
****
A gap breaks progression logic. Adding content or a prerequisite closes it, and the change log shows the reasoning.
****
---
## Sources and further reading
Status as of November 2026; check before publishing.
- Council of the European Union. *Council Recommendation of 16 June 2022 on a European approach to micro-credentials for lifelong learning and employability.* Official Journal of the European Union. Use the official text on EUR-Lex and check for Commission implementation reporting and national transpositions.
- European Commission. European Learning Model and European Digital Credentials for Learning: official documentation and specifications, published by the Commission. Check the current version and any profiles that apply to your institution or country.
- European Commission. DC4EU wallet pilot: public information on the use of the European Learning Model as the data model for education credentials. Check current project documentation.
- ICoBC. *ICoBC Credentialing Skills Framework (ICSF)* and the *VRITE* quality framework, published by the framework owner, the International Council on Badges and Credentials. Use the current published version.
- ICoBC. *ICB-P01 Designing a Credential: Claim, Criteria, Evidence and Metadata.* Prerequisite course for this course.
Time-sensitive points in this course: the state of national transpositions of the Council Recommendation, the current version of the European Learning Model specifications and the status of the DC4EU pilot. Status as of November 2026; check before publishing.