Executive summary
Most organizations already hold the evidence a training plan needs: who has completed and passed which learning, which required training is due, what people have asked for, and what they said about the courses they took. What is usually missing is the step between that evidence and the plan: a recorded decision, with a reason, about each need.
Salalem connects that work in five steps. Evidence raises needs. A person makes a decision on each one: train, already covered, knowledge fix, not a training problem, or defer. The response is tied to the decision, and a review of later evidence shows whether the need is still open. Results are reported as requirements fulfilled and assessments passed. The competency framework, where there is one, remains the organization's own.
What completion records leave open
Learning systems keep good records of activity: assignments, completions, assessment results, certificates. The records are seldom the hard part. The hard part is turning them into a decision. When a leader asks what training the organization needs next, the answer is often assembled by hand from exports, email requests and memory.
A completion record says who finished. What to do next is still a decision, and somebody has to make it and write it down.
Evidence sits in different places
Completions live in the learning system, obligations in a compliance tracker, requests in email, and course feedback in survey exports. Reading them together is manual work, so it tends to happen once a year.
Not every need is a training need
Some needs are already answered by an existing program, some by a knowledge article, and some belong to a process owner outside L&D. Treating each one as a course to assign spends budget and leaves the cause in place.
Decisions leave no trace
When a need is deferred or declined, the reason usually stays in a meeting. A year later nobody can say why a program exists, or why a request was set aside.
Salalem treats needs analysis as ongoing work with a record. Evidence raises a need, a person decides what to do about it, the response is linked to that decision, and later evidence shows whether the need is still open.
The model at a glance
The model has five steps and repeats. It starts from records the organization already has, and every step leaves something a reviewer can open later.
Steps 3–4 · Decide, then respond
A person records each decision. Learning for a train decision is chosen, built or connected.
Figure 1. Evidence, needs, decisions, responses, review. Each step is recorded, so a later reviewer can trace a program back to the need and the reason behind it.
Two choices shape the model. First, a need is not assumed to be a course: of the five decisions, only train leads to new learning. Second, the evidence is the records the organization already keeps: completions, passes, obligations, requests and counts of people.
Evidence, and the needs it raises
A need starts from something the organization can point to. Each reading on a need names its source and the group of people it covers, so a reviewer can see what supports it.
What counts as evidence
Completed and passed learning against the requirements a role carries. Required training and its deadlines, including certifications about to expire. Requests for courses, and searches in the knowledge library that found nothing. Course evaluations. And, when an administrator switches them on, an anonymous learner questionnaire and a short intake that asks managers and sponsors for their team's top needs. An assessment connected to a requirement counts once it is passed.
A framework helps, and is optional
Where the organization has a competency framework, each role's requirements read as met, not met, or not yet connected to any learning. The framework is the organization's own: Salalem imports it and works against it. Several sources, such as obligations, expiring certifications, requests and evaluations, raise needs with no framework at all.
Missing evidence stays visible
A source that has not answered, or a group too small to report on, is shown as unassessed. It is never shown as a low figure. Priority comes from named signals, among them how much of a requirement is unmet, how many people it affects and how close its deadline is. The reviewer's own business priority is recorded beside it, separately.
Five decisions, each with a reason
Every need gets one of five decisions, and each one needs a reason. Nothing is pre-selected. When asked, Salalem can draft a rationale and suggest a decision. The suggestion is labelled as one, and it is recorded only if the administrator chooses it.
| Decision | When it fits | What follows |
|---|---|---|
| Train | New or extended learning is the right answer | A learning response for the people the need covers |
| Already covered | An existing program already answers the need | That program is linked to the need as its coverage |
| Knowledge fix | A reference or a policy update answers the need | A knowledge library entry is linked to the need |
| Not a training problem | The cause lies in a process, a tool or a policy | The need is handed to a named owner, and the action is tracked |
| Defer | The need is real, and can wait | The reason is kept, so the need is part of the next review |
The decision history stays with the need. When the same need comes up in a later round, the reviewer sees what was decided before and why.
Responses: choose, build or connect
A train decision needs learning. Salalem can propose a staged plan from the unmet requirements, with a reason for each item, and a person fulfills each proposal in one of three ways. A proposal changes nothing until someone acts on it.
A person attaches an existing catalog activity to the program. Reusing what the organization already owns avoids a second course on the same subject.
A person can ask Salalem to draft a course for the requirement, build one around approved and cited sources, or request it from the Salalem content team. Every draft is reviewed and edited before it is published, and a program cannot be published while one of its courses is still waiting for content.
A person links a workshop, a mentoring arrangement or an external course into the same program, so it is assigned and tracked with everything else. Where mentoring fits, Salalem can suggest a colleague from the organization chart.
The other four decisions have responses of their own: a linked program, a knowledge entry, a named owner, or a recorded reason to wait. None of them needs a role-based program to exist first.
Review whether the need is still open
A published program is available to the people it covers. Whether the need is met shows up later, in the evidence, and the review reads that evidence against the need that started it.
Requirements fulfilled, assessments passed
Results are reported in the terms the records support. A branch manager's checklist might read 18 of 24 requirements fulfilled, 4 not met, and 2 not yet connected to any learning. A requirement is fulfilled when the person has completed, or passed, an activity connected to it, and an assessment result is pass or fail.
Each round keeps its own record
When a cycle closes, its report is frozen: the needs, decisions and counts it saw stay as they were, and later evidence does not rewrite them. The next round compares against the last one, showing which needs are new and which way the unmet share moved. The same need carries its decision history forward.
Governance and trust
A regulated organization has to be able to show who decided what, on which evidence, and when. The model is built so that people make the decisions and every step leaves a record. It works from completion, pass and request records, and Salalem does not rate anyone's proficiency.
-
A person decides every need
Salalem can draft a rationale or a course when asked. Each draft is labelled as a draft, and nothing is recorded, published or assigned until a person approves it.
-
Sources are shown on external material
When a course is built around external material, only approved sources are used, and each one is cited.
-
Each organization's data stays separate
Evidence, decisions and drafts are read and written inside one organization's workspace, and never across clients.
-
Reports stay as they were closed
A closed cycle's report cannot be rewritten by later evidence. It can be exported as a PDF or a spreadsheet, and its needs chapter can go into a compliance audit binder with the decisions and their open or finished hand-offs.
Hosting, data protection and support
What the reports show
Every figure the model reports is a completion, a pass, an obligation, a request, or a count of people. These are the questions it is built to answer.
In each review
- Requirements fulfilled — per role and per person: met, not met, or not yet connected to any learning.
- Assessments passed — an assessment connected to a requirement counts once it is passed; the result is pass or fail.
- Decisions and hand-offs — what was decided for each need, with its reason, and which hand-offs are finished or still open.
- Change since the last round — which needs are new, and which way the unmet share moved.
- What was measured — each report states what it measured and what it did not.
Start with one learning question
The model runs on records most organizations already have. A first review can start from required training and completions, and take in more sources as they become available.
Evidence, needs, decisions, responses, review
See it on a need of your own
Book a walkthrough and bring one question: a role, an obligation, or a request that keeps coming back. We will show how Salalem gathers the evidence, records the decision, and links the response.
Deciding training needs on evidence. A Salalem white paper, revised September 2026. Features described may change; ask us what is live today.