Employers tend to think of the Workplace Skills Plan as a deadline problem. Get it in by 30 April and the mandatory grant follows. In practice a submission can arrive on time and still not produce a payment, and the reasons are almost always in the data rather than the date.
Here is what actually gets checked, and where submissions come apart.
Is submitting on time enough to get paid?
The deadline is real. Practitioners who file across multiple SETAs report that late submissions are generally not accepted without formal SETA approval, and that such approval is rarely granted. So lateness is fatal.
But an on-time submission still has to satisfy a set of conditions before the grant is released. Meeting the deadline gets you assessed. It does not get you paid.
Not sure where your own submission stands? Send us a payroll summary and we will tell you what you have been entitled to and what has actually been claimed.
Request a call backWhy does a correct submission still get queried?
This is the wedge, and it is the reason competent employers still get queried.
Your SETA is not the only organisation holding data about your workforce. SARS already knows what you declared as your leviable amount, because you have been paying 1% of it every month. Your WSP states a headcount and an occupational breakdown. If those two pictures disagree, the submission raises a question, and questions delay payments.
The disagreement is usually innocent. A construction firm counts site staff differently from the way payroll does. A food manufacturer's headcount swings with the season. A retailer's staff turnover means the number on 31 March bears little relation to the average across the year.
None of that is dishonest. All of it looks like a data error to somebody assessing a file they did not compile.
Build the plan from the payroll export rather than from last year's plan. A submission that reconciles to what SARS already holds answers the question before it is asked.
What has to be registered before you submit?
Two registrations sit upstream of the submission and both are commonly assumed rather than checked.
The first is your SDL registration with SARS. The second is the appointment of your Skills Development Facilitator, which the SETA must have on record. MICT SETA's guidance is explicit that the SDF must be registered and accepted before commencing the function, and must be authorised by the company on a letterhead.
An employer who appoints an SDF in March and submits in April has done these in the wrong order. The submission may go in and then sit, because the person who lodged it was not on record as entitled to.
What evidence does the ATR actually need?
The Annual Training Report is the half that fails most often, because it is retrospective. The plan is a statement of intent and nobody can contradict it yet. The report is a claim about what happened, and it has to be supported.
What that means in practice is attendance registers, certificates, invoices and provider details, gathered while the training happened rather than reconstructed afterwards. An employer reconstructing a year of training in April is not really reporting, they are remembering, and the gaps show.
There is a specific and expensive version of this. Statutory and safety training that a company already delivers every year, forklift tickets, first aid, working at heights, SHEQ induction, is very often paid for, delivered properly and never captured as skills spend at all. The work was done. The credit was not claimed.
Consultation that has to be demonstrable
The plan is meant to be developed with the workforce, through a training committee or worker representatives, depending on the size of the organisation. Where that consultation is real but undocumented, the submission has a gap that is difficult to fill later. A dated set of minutes is a small piece of admin that protects the whole file.
Does paying for a course guarantee a grant?
Paying for a course does not automatically make the spend claimable. Guidance aimed at employers is unusually blunt on this point, warning that employers should not assume that paying for an accredited course guarantees a mandatory grant or discretionary funding.
The mandatory grant follows the submission being compliant. It does not follow the training being expensive. Employers who conflate the two are surprised twice, once when the grant is smaller than expected and once when the B-BBEE scorecard does not move either.
A checklist worth running in February
- SDL registration with SARS is active and the leviable amount is what you think it is
- The SDF appointment is registered with the SETA and the authorisation letter exists
- Headcount and occupational breakdown reconcile to the payroll export, not to last year's plan
- Demographic breakdown by race, gender and disability is complete
- Every training event from the past year has a register, a certificate or an invoice behind it
- Statutory and safety training is captured as training, not buried in operations
- Training committee or worker consultation is minuted and dated
- You know which SETA you are actually registered with, and it is the right one
February rather than April, because every item on that list takes longer to fix than to check, and the ones involving other people take longest.
The two documents do opposite jobs, and mixing them up is fatal
They are submitted together, on the same deadline, through the same portal, signed off by the same people. That is where the similarity ends, and employers who treat them as one document routinely fail one half of it.
| Workplace Skills Plan | Annual Training Report | |
|---|---|---|
| Direction | Forward, the year ahead | Backward, the year just finished |
| What it claims | The training you intend to deliver | The training you actually delivered |
| What supports it | Skills audit, training needs analysis, costed plan | Attendance registers, certificates, provider invoices |
| How it fails | Headcount that does not reconcile to payroll | Evidence that was never collected |
The plan is a statement of intent, and nobody can contradict it yet. The report is a factual claim about the past, and it either has evidence behind it or it does not. That asymmetry is why the ATR is the half that fails.
The skills audit that should come first, and usually does not
A plan built from last year's plan is a copy. A plan built from a skills audit is evidence. There is a method for running one that produces the fields a submission needs rather than a capability report you cannot file.
The audit maps the gap between the skills the business has and the skills it needs, role by role. A training needs analysis turns that gap into specific interventions, each with a target group, an NQF level and a cost. Done properly it produces the WSP almost as a by-product, and it produces it in a form that survives questioning, because every line traces to an identified gap rather than to a preference.
It also settles a question employers otherwise guess at, which is whether a given intervention should be accredited and credit-bearing or a skills programme without credits. That choice affects what the training is worth on the grant and on the scorecard, and it is much cheaper to make before the training is bought than after.
Doing the audit in February gives you a plan in March. Skipping it means writing the plan in April from memory, which is the single most reliable predictor of a queried submission.
What the plan actually has to contain
The detail is more specific than most employers expect, and vagueness reads as incompleteness.
- The planned training interventions, each mapped to an identified skills gap
- Learner numbers broken down by race, gender and disability, which is the detail that drives both the SETA reporting and your B-BBEE position
- Occupational categories, and where relevant the scarce or critical skills the intervention addresses
- Each intervention flagged as accredited or not
The demographic breakdown is the item most often supplied loosely. It is also the item that determines whether the same spend earns scorecard points, which we cover in the piece on the twenty points.
Who has to sign it, and why that catches people
A submission is signed off by two parties. The registered Skills Development Facilitator, and an authorised representative of the company.
Both halves cause problems. An SDF who is not registered on your SETA's system for the current cycle cannot sign off the submission at all, which is a different failure from being late and produces a different, more confusing rejection. And an authorised representative means somebody with actual authority to bind the company, evidenced on a letterhead, not simply whoever had portal access.
Check both in February. Registration lapses quietly between cycles and nobody is notified.
What happens after you press submit
Submission is not the end of the process, and treating it as the end is why some employers believe they submitted and were never paid.
The file is assessed. If it is approved, the mandatory grant is paid on the SETA's own schedule, which is not immediate and varies between SETAs. If it is queried, the query arrives through the portal or by email to the registered SDF, which is a further reason the registration needs to be current and monitored.
Approval also unlocks something separate. It establishes the compliance history that discretionary grant applications are gated behind, so a clean submission is worth more than the twenty percent it returns directly. That is set out in the piece on discretionary grants.
What are the most common reasons a submission fails?
- No evidence behind the ATR. The training happened. Nothing was kept.
- Headcount that does not reconcile. The plan and the payroll describe two different workforces.
- SDF registration lapsed or never completed. Fixable in advance, fatal in April.
- Statutory training never captured. Real spend, real delivery, no claim.
- Consultation undocumented. It happened, there are no minutes.
- Wrong SETA. Rarer, but it invalidates everything above it.
What to do if last year's submission was rejected
Find out precisely why before you change anything. A rejection for a reconciliation problem and a rejection for a missing SDF registration are fixed in completely different places, and employers who guess tend to rebuild the part that was already fine.
The previous year cannot be recovered. The mandatory grant is forfeited once the window closes, as we set out in the piece on what it costs to skip a submission. What can change is that this year's file is built to survive the same assessment, and that the same query does not arrive twice.
If you are not certain which SETA your submission should even be going to, that is worth settling first, and our sector guide sets out how the mapping works. If you would rather have someone look at the actual file, ask for a call back and we will tell you what we see.
Sources used: MICT SETA on SDF registration, SARS on the skills development levy.