Most L&D teams know they have too much content. What they tend to underestimate is how much that surplus is actually costing them. The conversation is usually framed around learner experience: too much content creates choice paralysis, makes discovery harder, and reduces engagement. Those things are true. But the financial case is just as compelling, and often more persuasive at a budget conversation.

Where the costs accumulate

Content costs operate at several levels. At the most visible level, there is licence spend. Enterprises typically subscribe to multiple content libraries, and those libraries overlap significantly. The same topic, covered by different vendors, sitting in different systems, all being paid for simultaneously. Without deduplication, that overlap is invisible and the cost builds year on year.

Below that sits production cost. Internally produced content takes significant time to create, review, maintain and update. Content that is no longer relevant to the business but has not been retired still carries a maintenance burden. Every piece of content that should be removed but is not adds overhead.

There is also the cost of curation. Somebody has to organise the library, create learning pathways, and surface the right content to the right people. When the estate is large and poorly structured, that task is enormous. Imperial Brands reduced pathway curation time by 83% after implementing Filtered Intelligence. That is a proxy for how much manual effort was being absorbed by an estate that had grown without adequate governance.

What produces the overlap

Overlap is the proportion of content where more than one provider contributed assets for a given format + skill segment. If you look across all formats, the average overlap is at least 20%. In other words, we estimate 20% of the total learning content load for any organisation overlaps. But his proportion rises if any of the following conditions apply:

• There are more than three content libraries in place
• We include internally-developed content in the comparison
• We include freely available content, curated from the web

That’s why Filtered clients tend to realise savings of 30-50%, rather than only 20%.

Our own analysis across content estates with 10,000+ employees and multiple libraries found an average overlap of at least 20% across all content formats, rising to 31% when looking at courses alone.

That overlap alone represents an estimated £3.7bn in wasted licence spend across large organisations globally, and once the full cost of content (production, curation, maintenance) is factored in, the figure rises to £9.1bn.

Content that cannot be found is content that does not deliver value, regardless of its quality. In large estates, content is often siloed across systems with different metadata standards and search functionality. The same asset may be tagged differently in different platforms, making it effectively invisible in one context while duplicated in another.

Poor discoverability means learners either give up searching and go without, or find and complete content that is tangentially related to their need rather than directly relevant. In both cases, the value of the content is not realised, and the investment is wasted.

What content rationalisation actually involves

Rationalisation is not deletion. It is the process of understanding what you have, identifying what is working, and making deliberate decisions about what to keep, update, retire or replace.

Done properly, it requires accurate data on every asset in your estate: quality scores, usage data, skills mapping, freshness and duplication flags. Without that data, rationalisation is a manual exercise based on educated guesses, and it rarely produces meaningful results.

Signal Quality, part of Filtered Intelligence, scores every asset in your content estate continuously and automatically. It flags duplication, drives ongoing optimisation, and keeps your library current as your business evolves. That means rationalisation is not a one-time project but a continuous process, which is the only approach that works at enterprise scale.

The business case

AstraZeneca used Filtered to identify a 41% saving on external learning catalogue spend which they reinvested in more meaningful upskilling initiatives. GSK generated a cost saving through a data-led realignment of its leadership and business skills content. In both cases, the saving came not from cutting programmes but from understanding the estate clearly enough to make better decisions about it.

The hidden cost of too much learning content is not just an L&D problem. It is a business problem. And like most business problems, it is best solved with better data.