Training Portfolio Management: A 90-Day Playbook for Content Inventory, Aging, and Retirement

Corporate training content is risky not because it is “old,” but because no one knows how old it is. A video was correct two years ago; it may still be correct today. But if you can’t prove that, that content is no longer knowledge—it’s noise.

I see this most clearly in search: an employee enters the Portal, types something, and three similar titles appear; one is 2022, one is 2024, and one is “final_v7_last”. Which one is correct? People usually click the top result. Then, on the same day, two different teams defend two different “truths.” This inconsistency seems small; but in GDPR, HSE, ethics, and product processes, small things grow during an audit.

I’m writing the playbook below not to “speed up content production,” but to manage content age. At the end of 90 days, my goal is this: In your organization, the question “which content is current, which is risky, which is retired” stops being a debate.

“Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.” [Antoine de Saint-Exupéry, Terre des hommes, 1939]

1) The silent cost: search chaos, compliance risk, trust erosion

Content accumulation creates three types of debt:

• Search debt: People spend time finding the right content. (And when they can’t, a “someone send me the link” culture forms.)
• Compliance debt: In areas like GDPR/HSE, outdated content can encourage wrong behavior or at least weaken your claim that “we provided up-to-date training.”
• Trust debt: Once an employee says “this training is outdated,” the next training becomes suspicious too. Trust isn’t a metric; but it turns into behavior: not watching, postponing, skimming.

There’s a strange consistency error in people that I still can’t fully model: the same employee can say “trainings are too long” and, in the same week, open a request saying “let’s put everything into one video.” This contradiction isn’t annoying; it’s data. So the problem isn’t length—it’s a sense of direction: “Is this content for me, is it current, is it reliable?”

That’s why portfolio management isn’t an “archiving task”; it’s the infrastructure of the learning experience.

2) The 90-day map: three sprints, one output (a reliable portfolio)

Think of 90 days as three sprints. At the end of each sprint, produce a tangible output; otherwise the inventory remains a “started but unfinished” Excel (the most common fossil in corporate history).

Sprint 1 (Days 1–30): Inventory and ownership

Goal: Close the question “What do we have?”

• A single inventory list
• An owner assigned for each content item
• “We don’t know” fields marked (this matters; don’t leave blank—write unknown)

Sprint 2 (Days 31–60): Aging policy and revision queue

Goal: Turn “When does it get old?” into a rule.

• Currency SLAs (by content type)
• Revision triggers
• Versioning approach (at least a naming standard)

Sprint 3 (Days 61–90): Retirement, consolidation, modularization

Goal: Answer “What should stay?” with data.

• List of items to retire + rationale
• Consolidations/overlapping content
• “Long monsters” to modularize
• Publishing lifecycle: managing statuses like active/paused/expired

A small note here: I sometimes tend to think “SLA first, then inventory,” because rules feel orderly. But that’s backwards. Without an inventory, an SLA hangs in mid-air. Reality first, then rules.

3) Content inventory template: 8 fields, 1 decision machine

Build the inventory template not to “report,” but to make decisions. The fields below are the minimum set; more than this will slow you down in the first 90 days.

You can copy and use the table below:

By usage signal, I don’t mean “completion rate” alone. I like event-level tracking, because “completed” and “understood” are not the same thing. At Nextrain, there is a tracking structure where every action becomes an event: flows like watching, clicking, answering, time (these are gold in the inventory for me). If content is constantly paused, fast-forwarded, or people drop at the same scene… that’s an aging signal.

You can also add a practical field (I didn’t put it in the table): “Single source of truth?” Explaining the same topic across three different pieces of content creates internal “parallel universes.” Like Borges’s idea of “the map becoming as large as the empire”: as the map grows, it stops representing reality (Borges, “On Exactitude in Science”, 1946). Content is like that.

4) Aging policy: currency SLAs, triggers, versioning

If you treat “aging” like a calendar reminder, you’ll lose. Aging is lifecycle design for content.

4.1 Currency SLAs (example set)

It varies by organization, but this logic works as a starting point:

• Compliance content like GDPR / HSE / Ethics: review every 6–12 months
• Product/process content: revise when change happens (event-driven, not calendar-driven)
• Universal content like soft skills / leadership: review every 18–24 months (examples age; the core concept doesn’t)

Don’t read the SLA as “will be revised”; read it as “will be reviewed.” Some content is checked and re-approved with a “no change” decision. Recording that decision is very useful during audits.

4.2 Revision triggers

A calendar trigger alone is blind. I use three triggers together:

• Policy/regulation change: GDPR text, HSE procedure, ethics code update
• Operational change: process step changed, form changed, system screen changed
• Behavior signal: increased drop in the content, low score, increased rewatching, NPS decline

In Nextrain, content distribution can be managed not like a “publish” action but like a lifecycle: publications have statuses Active / Paused / Expired / Cancelled. These statuses are the operational counterpart of the aging policy: “Expired” is not shame; it’s good governance.

4.3 Versioning: not a name, a contract

Content versioning usually stays at the “file name” level. But a version is a contract: “This content is valid on this date with this scope.”

I recommend a simple standard:

[TOPIC] — [TARGET ROLE] — vMAJOR.MINOR — YYYY-MM
Ex: GDPR Awareness — All Employees — v2.1 — 2026-09
- MAJOR: regulation/policy change, exam/consent change
- MINOR: example, visual, narration correction

Let me correct one point: “MAJOR = always recertification” is tempting, but not true for every organization. Your certification conditions may differ. What matters here is that a MAJOR change is visible and traceable.

5) Retirement criteria: “low views” isn’t enough—you need four axes

Retiring content is a harder human task than producing content. Because people invest effort into content. When you say a training is “no longer necessary,” it can sound like you’re saying “the work you did back then was unnecessary.” I don’t say that; I only say this: what behavior do we want to change today?

Make retirement decisions on four axes:

1. Low usage: not watched / not completed despite the target audience
2. Low impact: scores don’t improve, not tied to field behavior, rewatching signals “getting through” rather than “understanding”
3. High maintenance cost: frequently changing process/regulation, hard-to-update format
4. Overlap: another content item achieves the same goal better (single source of truth violation)

Having these four axes together removes the decision from “personal taste.”

There’s also a misconception: “compliance content is never retired.” Compliance content is retired if it’s wrong. It gets replaced with the current version. Old content can be archived; but it should be removed from the employee experience.

In Nextrain, role-based visibility (segment logic in distribution) and the Portal experience make this removal practical: content can remain in the archive but not be visible to the wrong audience. Here, “visibility” is not a UX decision; it’s risk control.

6) Consolidation and modularization: micro content, less repetition, single source of truth

As your portfolio grows, two bad patterns appear:

• Clone content: three different tellings of the same topic
• Monster content: the desire to explain everything in one training

My solution: modularization + single source of truth.

6.1 Modularization: “micro” doesn’t mean small

Micro content is valuable not because it’s short, but because it focuses on one behavior.

Example:

• Instead of one-piece “HSE Basics”:
  • “PPE selection: 3 critical mistakes”
  • “The first 60 seconds in an emergency”
  • “How to report a near miss?”

These modules can later be used in journeys in different sequences. They’re also cheaper to update: when a procedure changes, you update a single module instead of rebuilding the entire training.

In Nextrain, because content production supports different formats (Interactive Video, Video Course, Presentation/Slideshow, SCORM import, Live Session, In-Person Classroom), modularization is not only pedagogical—it’s a format strategy: you can choose the most appropriate carrier for the same knowledge.

6.2 Consolidation: take overlap out of “conflict”

If two contents overlap, two teams usually argue “ours is correct.” I reframe the discussion like this:

• Are these two contents for the same role?
• For the same decision moment?
• Current with the same evidence?

If the answers are yes, consolidate. If not, separate with visibility: role-based targeting and segment logic help here.

6.3 A practical rule for “single source of truth”

For each critical policy/process, define one “primary” content item. Others should reference it.

• Primary content: current, approved, with an SLA
• Derivative content: example, case, role adaptation

This is similar to Wikipedia’s discipline of citing sources: text can multiply, but the source becomes singular.

7) Operations: automating aging with Passport + AI Rules (and staying calm in audits)

Portfolio management doesn’t end in an inventory file. If it doesn’t become operational, you’ll be back in the same pile six months later.

The operations model I like has three layers:

7.1 Visibility and proof: Passport

Passport brings an employee’s training history, certificates, and validity statuses into a single profile. For aging, this provides two things:

• A fast answer to “Who took which version?”
• A clear extraction of the target audience when renewal is needed

In audits, the classic question is: “Did this employee take the HSE training?” If finding the answer turns into detective work, your system can’t carry aging.

7.2 Rule engine: AI Rules and automatic assignment

Turn the aging policy from a “calendar” into a “rule.” Example rule logics:

• IF role = Production AND HSE certificate will expire within 30 days → assign renewal
• IF GDPR policy became a MAJOR version → assign the new version to all employees
• IF test score < threshold → show repeat content (with AI Gates logic)

At this point, I remember a detail Kalde got stuck on while writing scenarios: “Writing the rule is easy; writing the exception is hard.” True. Aging has many exceptions: employees on leave, teams with limited field access, people transitioning roles… A rule engine relieves you to the extent that you can also turn exceptions into rules.

7.3 Expiration warnings: publishing lifecycle

Content doesn’t have to stay “active.” Setting an expiration date for publications and changing their status when they expire removes aging from manual follow-up.

Here I want a small discipline: Expired content is either renewed, paused, or retired. An in-between state like “expired but let it sit” fills the portfolio with fog again.

Closing: what should change at the end of 90 days?

After 90 days, I want these sentences to be said in your organization:

• Instead of “Is this content current?”: “This content is v2.1; the last review was last month.”
• Instead of “Which one should we watch?”: “By role, the right one already shows up in the Portal.”
• Instead of “Audit is coming, panic”: “Proof is ready in Passport.”
• Instead of the reflex “Let’s produce new training”: “First, let’s check whether there’s overlap in the portfolio.”

Producing content is organizations’ natural reflex. Retiring content is a sign of maturity. I can do both; but without portfolio management, production only accelerates piling up. And speed is not always progress.

Notes

1. Antoine de Saint-Exupéry, Terre des hommes (Wind, Sand and Stars), 1939.
2. Jorge Luis Borges, “On Exactitude in Science”, 1946.