Your Catalog Is a Garden, Not a Warehouse

Run this on your learning platform and report back.

Sort every course by last updated. Now look at how many were last touched more than two years ago. Then look at how many of those are about your products, your processes, your policies, or your systems — the things that have definitely changed since.

That number is not a content inventory. It's a list of things your platform is currently telling employees that are, to some degree, no longer true.

We talk about content libraries as if they were warehouses: things you fill, things that sit there, things that retain their value while stored. They're not. They're gardens. Everything in them is alive, most of it is dying, and if nobody is weeding, you are not running a library — you are running a compost heap with a search box.

What rot actually costs

The instinct is to treat stale content as harmless clutter. It isn't, and the damage runs in three directions.

It destroys trust, retail. An employee searches for something, finds a module about a system you decommissioned in 2023, and closes the tab. That is not a neutral event. She has just learned that the platform cannot be relied on, and she will apply that lesson to every future moment when she might have searched. Trust in a knowledge system is destroyed one bad result at a time and rebuilt — if at all — very slowly.

It teaches the wrong thing confidently. Worse than a dead link is a live module with an outdated procedure, an old threshold, or a superseded policy. It doesn't announce that it's stale. It just instructs, fluently, in error — and the employee who follows it exactly is the one who gets blamed.

It hides what's missing. A company with four thousand courses feels well-provisioned. The feeling is doing damage: it disguises the absence of the content people are actually searching for and failing to find. The bloat is a fog, and nobody audits a fog.

The three disciplines, which cost almost nothing

Everything I've seen work in course management at scale comes down to three habits. None requires budget. All require someone to decide they matter.

1. Every course gets an owner and a review date — at birth.

Not a department. A name. The owner is the person who receives the review reminder, and who has authority to say "this is still correct" or "this is dead."

Content without an owner is content nobody will ever update, which means it will rot on schedule. The review cadence should track volatility: product and process content quarterly, compliance content on the regulatory cycle, evergreen skills content annually. And the rule that gives this teeth: unowned content is presumptively dead. If nobody will put their name to it, it comes down.

2. A monthly pruning hour, protected on the calendar.

One hour. Three questions. What had zero usage in the last six months? What has passed its review date without action? What duplicates something better?

Then delete. Not "archive pending review" — the graveyard where content goes to remain findable. Delete, or archive somewhere genuinely out of the search index.

The emotional obstacle here is real and worth naming: every course was commissioned by someone, often at cost, sometimes by someone senior. Removing it feels like waste. But the sunk cost is sunk either way; the only live question is whether it's now doing harm. Most of it is.

Teams that skip this hour for a year don't have a backlog. They have an archaeology project.

3. Metadata as a publishing gate.

At scale, findability is metadata. Consistent naming, mandatory skill tags, accurate duration, an audience label. And the rule that makes it stick, borrowed from a team who put it better than I can: a course that isn't properly tagged isn't published, because it may as well not exist.

This feels bureaucratic for about three weeks and then quietly becomes the reason your search box works.

The signal you're already ignoring

There's a fourth habit, and it's the one I'd install first if I could only pick one: read the failed searches.

Every month, pull the list of searches that returned nothing useful. It is the single most honest document your organization produces. It's written by your workforce, in the moment of genuine need, unprompted, and it costs nothing to obtain.

Every entry is a person who wanted to know something, came to you, and left with nothing.

At every company I've watched do this, the same thing happens: the list is almost entirely internal. Product names. Process steps. Client types. Systems. Policy questions. The things no library sells, because no library knows about your company. That list is your commissioning plan, and it updates itself weekly, for free, forever.

Meanwhile, the content you did buy sits in the catalog, un-searched, being counted as an asset.

What "garden" actually implies

The metaphor does more work than it first appears, so let me push it.

Gardens have seasons. Content that mattered enormously during last year's product launch is dead weight now. That's not failure — it's the natural lifecycle, and a garden that never removes last season's growth is not sentimental, it's neglected.

Gardens have a front and a back. The front is what people see on the home screen — curated, changing, seasonal, opinionated. The back is the long tail: the licensed library, available by search, useful to the curious, never promoted. The mistake most companies make is showing everyone the back.

Gardens need someone whose job this is. Nobody in your organization is currently paid to remove content. Everybody is paid to add it — the SME who built a module, the L&D manager who licensed a library, the vendor who sells titles by the thousand. The entire incentive structure points one way, which is why every untended catalog grows monotonically toward uselessness. Someone has to be given the weeding as an actual, named responsibility, or it will simply never occur.

And a well-tended small garden beats a vast neglected one. Fifty current, relevant, owned pieces of content will outperform five thousand unowned ones on every metric that matters — search success, voluntary return, trust, and the thing all of those add up to, which is whether anybody comes back.

The pruning conversation, scripted

The genuine obstacle to weeding isn't intellectual. It's a conversation nobody wants to have: telling a stakeholder that the module their team commissioned is coming down.

So here's the script, because having one removes most of the friction.

Don't frame it as a judgment. "This didn't perform" invites a defence of the content, its author, and the budget that funded it. Frame it as a lifecycle: "Everything on the platform carries a review date. This one has reached its date, and it has had eleven views in twelve months. Do you want to own the refresh, or shall we retire it?"

That framing does three useful things. It makes removal the default rather than the attack. It offers the stakeholder a genuine choice — refresh it, and it stays. And it moves the burden from you (defending a deletion) to them (defending a keep), which is where it belongs, because they're the ones who know whether it still matters.

Report the aggregate, not the incidents. Once a quarter, tell leadership: this many items reviewed, this many refreshed, this many retired, catalog freshness now at this percentage. Pruning framed as an ongoing operational metric is uncontroversial. Pruning framed as a series of individual removals is a series of individual arguments.

And never delete something quietly that someone senior cares about. Not because they're right, but because one ambush will cost you the mandate to prune anything for a year. Tell them first. They will almost always say yes.

The weeding isn't hard. The politics of weeding is hard, and the politics dissolve the moment removal becomes a scheduled process rather than an editorial opinion.

The uncomfortable version

Most organizations don't have a content problem. They have a maintenance problem that they have never once framed as a job.

They bought the warehouse. They filled the warehouse. They report annually on how full the warehouse is. And they cannot understand why nobody wants to shop there.

An hour a month. An owner per item. A review date. And the humility to read what people searched for and didn't find.

That's the whole practice. It's cheaper than the next content licence, and it's the only thing that will make the last one worth anything.