Ask any course seller who teaches something assessable where the grading actually happens, and the answer is usually a mail client. The student attaches a document, the instructor replies with comments, and both sides lose track of which version was marked, what was said last time, and whether the course was ever really finished.
That is not a discipline problem. It is a missing product feature. This guide explains the model that replaces the email thread — declare, freeze, queue, verdict — and how it works in Alva Courses, the Shopify app that builds, sells and delivers courses natively inside your store.
Why auto-marked quizzes stop short
A multiple-choice question can check whether a fact was retained. It cannot judge a business plan, a case study, a translation, a recorded technique or a portfolio. For professional and creative courses the assessable work is exactly the part a machine cannot mark, so the marking quietly moves into email.
Once it moves there, four things go missing at once. There is no queue, so submissions are found by scrolling. There is no record, so last month's feedback is not beside this month's. There is no freeze, so the student can keep editing the work you are reading. And there is no verdict the course itself understands, so nothing changes for the student when you finally decide they passed.
A per-question comment is not an evaluation
Both exist, and they answer different questions. A per-question review comment corrects one answer inside one quiz attempt. An evaluation is a verdict on the whole body of work: everything the student completed, every attempt they made and every reflection they wrote, judged once and recorded once.
Quizzes in Alva Courses already support written questions. A Short Answer or Long Answer question collects freeform text, you can store a Model Answer that only you see during review, and lesson completion can be held back until those answers have been reviewed. That is real marking, and for many courses it is enough.
What it cannot do is certify. A comment on question four does not say whether the student has met the standard of the course, and it is not the thing an employer, a regulator or a professional body is asking about. That question needs one answer, given once, against the whole submission — and that is what evaluation is for.
The submit-for-evaluation model
Four things happen in order. The student declares the work finished, the work freezes so it cannot change underneath you, the submission joins a queue, and an instructor reads it and returns one holistic verdict. Each step exists to remove an ambiguity that email grading leaves wide open.
1. The student declares it finished
Submission is an explicit act, not an inference from progress. The student presses submit for evaluation and confirms it, and they can only do so once every lesson is complete. That turns "I think I'm done" into a timestamped claim, which is the thing an email thread never establishes cleanly.
2. The work freezes
While a submission is pending, lessons cannot be marked complete and quiz answers cannot be changed. The student is warned about this before they confirm, in as many words. The point is simple: the work you are grading has to be the work that was submitted, and a moving target cannot be assessed fairly.
3. It joins a queue
Pending submissions collect in one Evaluations screen with a count of how many are waiting, each marked as pending evaluation. A queue is not a nicety — it is what lets you promise a turnaround time, split marking across a team, and see at a glance whether you are behind.
4. One verdict comes back
You return a single decision: pass, or revise and resubmit. Written feedback travels with it, a grade can travel with it too, and the course reacts accordingly. The student is not left to interpret a paragraph of comments and guess whether they finished.
What the instructor actually sees
Opening a submission gives you the whole picture on one screen: which lessons the student completed, every submitted quiz attempt with their answers and your model answers beside them, their current reflection entries with any feedback already given, and every prior evaluation of the same enrolment.
The prior-evaluation history is the part that changes how grading feels. On a second or third submission you are not marking in a vacuum: last cycle's feedback, outcome and grade are on the same page, so you can tell the student what has improved instead of repeating yourself. It also makes a second marker possible, because everything a colleague needs to be consistent with you is in front of them.
Pass, or revise and resubmit
The verdict is binary and the feedback is not. Passing closes the review, completes the course and releases the certificate. Revise and resubmit reopens the course so the student can improve the work and submit again, which starts a fresh cycle while keeping the previous one on the record.
Written feedback is a full text field rather than a comment box, and a single checkbox controls whether the student sees it. Keep an assessor's private note to yourself, or share the whole critique — the choice is per evaluation, so a difficult verdict can be delivered carefully rather than automatically.
On the student's side the states are unambiguous. While the review is open they see that the work has been submitted and is with their instructor. When it closes they see the outcome, the feedback you chose to share and, where you use one, the grade.
Grading the work: feedback, a grade, or both
An evaluation can carry written feedback alone, or feedback plus a letter grade drawn from the course's published scale. Enter a percentage and the letter is derived from that scale for you, so the grade you award and the scale you published can never quietly disagree with each other.
A grade also changes what happens next. The completion email has separate versions for higher and lower grades, so a distinction and a bare pass do not receive identical congratulations, and the grade itself can be printed on the certificate. That is covered in certificates with grades for online courses.
When not to use evaluation
Evaluation costs instructor time, and that is the point — it should be reserved for work worth reading. A hundred-student self-paced course with auto-marked quizzes does not need it. A twelve-student certification, a coached programme or an accredited CPD course is where a human verdict earns its price.
- Use it where a person, not a score, decides whether the standard was met: portfolios, case studies, recorded performances, written assignments, client work.
- Use it where the certificate is the product — accreditation, licensing, professional bodies, anything a third party will scrutinise.
- Skip it for quick-win and reference courses, where an auto-marked quiz already answers the only question being asked.
- Skip it at volumes you cannot service. A queue you never clear is worse than no queue, because the student is now waiting on you.
If you do turn it on, price the marking. An evaluated course is a different product from a self-paced one, and the time you spend reading is the reason it can carry a higher price and a real deadline. Pair it with a course deadline that actually expires so the marking arrives in batches instead of arriving forever.
Getting started
Switch on the requirement for one course, enrol yourself, and walk the loop once as a student before any customer does. For the exact steps, read Turn on course evaluation and Review and grade coursework. If your courses use written reflection as well as quizzes, reflection journals are the other half of the evidence an evaluator reads.
Frequently asked questions
Quiz review marks one freeform answer inside one attempt, with your own model answer beside it for reference. Course evaluation is a single verdict on the whole enrolment — every lesson, every attempt and every reflection — returned once as feedback, an outcome and, if you use one, a grade.
No. Submitting for evaluation freezes the enrolment, so lessons cannot be marked complete and quiz answers cannot be changed while the review is open. The student is told this before they confirm, and the work you read is the work they submitted.
Choose the revise-and-resubmit outcome. The course reopens, the student can improve the work and submit again, and the next submission is a new evaluation cycle. The earlier cycle stays on the record with its feedback, so you can see how the work changed.
No. On a course that requires evaluation, finishing the lessons is not the same as finishing the course. Completion and the certificate are both withheld until an instructor returns a passing verdict, which is what makes the certificate mean something.
No. If no grade scale is configured, evaluators still record written feedback and a pass-or-revise outcome. Add a scale when you want the verdict to carry a letter, and Alva Courses will only accept letters that exist on the scale you published.
Turn on evaluation notifications and you receive an email each time a student submits an evaluation-enabled course. Every waiting submission also sits in the Evaluations queue in the app, marked as pending, with a count of how many are waiting.