Your backlog is not a planning problem. It is a structural one.
Training demand grows with every software release, every reorganisation, every new compliance requirement. Production capacity in L&D rarely grows with it. The result is always the same: a waiting list, stale training, too much alignment and too little time for impact.
More people solve that only as long as you can hire people. And even then it just moves the bottleneck.
The three causes behind the backlog
Before the levers, look at the causes. They are the same in nearly every L&D team.
First: the knowledge sits in people's heads, in documents and in review loops, not in usable training formats. Between the expert and the finished course there are several translation steps, and each one costs time.
Second: L&D is the operational bottleneck although the expertise sits in the business units. Every request runs through the same two or three people.
Third: production itself is too heavy. Alignment, script, recording, editing and translation cost weeks per training.
Work only on cause three and you get faster. Work on cause two and you become scalable. Those are not the same thing.
Lever 1: take production out of the team
The most effective lever is not speed, it is distribution. As long as only L&D can produce, capacity is capped by team size.
At DPD Germany roughly 240 active authors from the business units produce content. In two and a half years more than 1,800 videos were created. L&D no longer builds every video there. L&D sets standards, reviews and approves.
The precondition is a tool that works without instructional design knowledge and without a camera. If the business unit needs training in order to build training, the lever is gone.
Lever 2: the interview as the entry point instead of the script
Writing a script is the most expensive way to get knowledge out of someone's head. The expert explains, someone in L&D takes notes, rewrites, sends it back, corrects. Three rounds later it is approved.
An AI-supported interview shortens that path: the expert talks, the system probes where something stays unclear, and produces the training format from it. A conversation becomes training rather than a text file that still needs producing.
The side effect is qualitative. In conversation, exceptions and edge cases come up that a written script almost always leaves out.
Lever 3: use the documents you already have
Most companies already hold hundreds of slide decks, work instructions and PDFs. They are just not training.
Document-to-Training turns them into a course in a few steps. That is the fastest way to reduce an existing backlog, because the input already exists and nobody waits for a slot with the business unit.
A practical tip from implementation: start with the content that is requested most often and updated least often. That is where effort is lowest and impact is visible immediately.
Lever 4: decouple updates
Scaling rarely fails at initial production. It fails at maintenance.
If you run 200 courses and each needs one revision a year, that is a permanent production stream absorbing all your new capacity. This is exactly where SCORM remote update applies: change the content, the change takes effect, no re-import into the LMS.
Do the sum for your own portfolio. At 200 courses and half an hour of LMS handwork per update, that is 100 hours consisting purely of import and versioning.
Lever 5: turn multilingual rollout from a project into an export
As long as every language version is a separate undertaking with voice talent, alignment and approval, international rollout does not scale. With 140 languages for subtitles and translation and 50+ languages for voiceover, the language version becomes an export step.
That changes the planning. You no longer decide which sites get a training this year, you roll out everywhere at once.
What you must not lose along the way
Scaling without governance creates a new problem: 240 authors can also mean 240 quality levels.
Three things hold it together. Clear approval paths where L&D remains the final instance. Templates and standards that prescribe structure rather than invite creativity. And reporting that makes impact visible: completion rates, certification records, compliance ratios.
L&D loses no control in this. The role shifts from production to steering. That is usually also the role the team was hired for.
The realistic entry point
Do not take the whole backlog. Take one process with high repeat demand, for example onboarding for a key role or the mandatory briefings of one site. Define two or three subject-matter experts as authors, set an approval path, produce for six weeks.
After that you have a number that holds for your company rather than a number from a blog article. And you know whether the bottleneck really was production or alignment.
