Grading More Than Multiple Choice

A multiple-choice quiz can tell you whether a student memorized a fact. It cannot tell you whether they can actually write a working SQL query, debug a real networking issue, design a game mechanic, or explain a blockchain concept in their own words well enough to teach it to someone else. For any course that claims to teach a real, applicable skill, as opposed to pure factual recall, a graded, human-reviewed assignment isn’t a nice extra. It’s the only assessment format honest about what the course is actually trying to produce: someone who can do the thing, not just recognize the right answer on a list.

We looked at the real Assignments screen on a working Learnomy install, and the numbers on it tell a story about exactly how central this feature actually is to a serious course catalog, not a peripheral add-on nobody uses.

There’s a temptation, when building a course platform, to treat quiz assessment as the whole story and assignments as an afterthought, a checkbox feature added because a features-comparison page expects it, without much investment in making the actual grading workflow usable at scale. The real data pulled directly from this install argues against that temptation being acceptable: 101 real assignments, spanning courses on digital marketing, game development, and blockchain development, several with double-digit pending counts right now. That’s not a feature sitting unused in a corner of the product, it’s clearly load-bearing across a genuinely diverse catalog, which means the quality of the actual grading workflow matters as much as the quality of the quiz engine covered elsewhere in this series.

What’s actually sitting in a real assignments queue

Here’s the genuine Assignments admin screen, mid-catalog, with real courses and real submission counts:

The subtitle states the mechanism plainly: “Create assignments attached to any Lesson. Students submit text from the Lesson page; you grade inline below.” That’s a simple, direct model, an assignment isn’t a separate, disconnected entity floating somewhere in the admin; it’s tied to a specific lesson, submitted from that lesson’s own page, and graded from a screen that shows exactly which lesson and course it belongs to.

Now look at the real numbers. “Document Your SEO Mastery Architecture,” attached to the lesson “How Google Ranks Pages in 2026,” has 21 submissions, 15 pending. “Unreal Engine 5 Code Review Exercise” has 21 submissions, 14 pending. Further down the same list (visible on the full screen, beyond what’s shown here), “Web3.js & DApp Development Practical Assignment” has 47 submissions, 22 pending. The full list runs to 101 total assignments across this one catalog, spanning topics from digital marketing to game development to blockchain, this isn’t a lightly-used feature bolted on for a demo, it’s an active, heavily-submitted assessment mechanism across a real, diverse course catalog.

The Pending column is the actual product

Look at what column sits front and center, styled distinctly from the rest: Pending, in an orange badge, immediately next to the raw Submissions count. That’s not decorative, it’s the same design principle seen in the analytics dashboard’s completion percentages and the gradebook’s pass rates elsewhere in this series: surface the number that represents unfinished human work, not just the aggregate activity count.

Think about what “15 pending” on the SEO Mastery assignment actually means in practice: 15 real students have submitted real work and are waiting on a real person to read it and respond. Every one of those students experiences that wait personally, did I do this right, when will I know, is anyone actually reading what I wrote. A Pending column that’s easy to see, sorted or filterable, is what keeps that number from silently growing into a backlog a student notices and an instructor doesn’t, until a support ticket or a bad review forces the issue.

Grade, Edit, Delete, three actions, one clear purpose each

Each assignment row carries exactly three actions: Grade, Edit, Delete. The Grade button is styled distinctly and placed first, the primary, most-used action on this screen, taking an admin or instructor directly into that specific assignment’s submission queue rather than requiring a detour through a generic “manage submissions” screen first. That ordering reflects the actual workflow: an instructor visiting this page is here to grade, not to reconfigure the assignment’s settings, and the interface gets out of the way of that primary task.

Why “students submit text from the Lesson page” is a deliberately simple design

It would be tempting to build an elaborate, standalone assignment-submission portal, separate from the actual lesson content. Learnomy’s approach is the opposite: submission happens directly from the lesson page itself, in the same place the student is already reading the material and watching the video the assignment is based on. That’s a meaningful usability decision, a student who just finished watching a lesson on Google Ads auction mechanics doesn’t have to navigate away to some separate “assignments” area of the site, hunt for the right one, and re-establish context about what they’re being asked to do. The prompt for “Create a Google Ads & PPC Workflow” sits right where the lesson “How Google Ads Auctions Actually Work” left off.

This also connects directly to the notes-and-bookmarks system covered elsewhere in this series, both features share the same underlying design instinct: keep the student’s actual work anchored to the specific lesson it relates to, rather than scattering it across separate, disconnected areas of the platform that require the student to remember which lesson something was originally about.

What this actually assesses that a quiz structurally can’t

The exam and quiz article elsewhere in this series covers Learnomy’s real assessment engine in depth, timed exams, six question types, anti-cheat tracking, and our comparison of the best LMS platforms with a quiz builder covers that same territory across the wider market. Assignments exist specifically to cover the territory quizzes structurally cannot: any skill whose demonstration is a genuine artifact rather than a selectable answer. “Document Your Social Media Marketing Architecture” isn’t a question with a right answer among four choices, it’s a piece of real work a student produces, that only a knowledgeable human can actually evaluate for quality, completeness, and correct reasoning.

Look again at the actual assignment titles in the real data: “Debug and Fix Multiplayer Networking Issues,” “Design a Blockchain Fundamentals Solution,” “Build a Small Hardhat Project,” “Debug and Fix DeFi Issues.” Every one of these names a skill that’s fundamentally about doing, not recognizing, debugging requires actually finding a real bug in real code, building requires actually producing a working thing, designing requires actually generating an original solution to a stated problem. No quiz question format, however cleverly constructed, tests any of these the way a real, human-graded submission does.

The submission cap, and why “Max: 100” matters

Every assignment row shows a Max column, 100 in every visible row here. That’s the maximum score an assignment can be graded out of, and its consistency across this catalog (100 across wildly different subjects, from SEO to blockchain) suggests a standard scale the instructors on this install have settled into, rather than each assignment using an arbitrary, inconsistent point scale that would make cross-course comparison meaningless. A consistent max score is what makes the gradebook’s weighted-grade calculation, covered in the analytics article elsewhere in this series, actually produce a meaningful class average, averaging scores out of wildly different maximums across different assignments would produce a number that doesn’t mean anything real.

How this feeds into the weighted gradebook

The Gradebook feature covered in the analytics article elsewhere in this series describes itself as producing “each Course’s weighted grade from its Quizzes, assignments and standards packages.” That word “assignments,” sitting right alongside quizzes in that description, confirms the two systems aren’t separate, disconnected assessment tracks, they feed the same class average. A course that mixes multiple-choice quizzes for factual checks with human-graded assignments for applied skill produces one combined, weighted class-average grade, not two separate numbers an instructor has to mentally combine themselves.

This matters because it means the choice between a quiz and an assignment for any given piece of assessment isn’t a choice between “counts toward the grade” and “doesn’t count”, it’s purely a choice about which format actually tests the specific thing you’re trying to verify. An instructor building a networking course can put factual recall (“what port does HTTPS use”) in a quiz and applied troubleshooting (“debug this multiplayer networking issue”) in an assignment, confident both contribute correctly to one coherent final grade, weighted according to whatever balance the instructor sets between them.

What a careless version of “assignments” looks like

It’s worth naming the weaker alternative some platforms ship, because “supports assignments” is another phrase, like “SCORM support” and “AI-assisted authoring” covered elsewhere in this series, that hides a wide range of actual implementation quality.

A careless version treats an assignment as a generic file-upload field with no real grading workflow attached, a student uploads something, and it sits in a folder somewhere with no queue, no pending count, and no inline grading interface, meaning an instructor has to manually track down and open every submission separately with no aggregate view of what’s outstanding. A careless version also frequently fails to connect the assignment’s score to the course’s overall grade calculation at all, leaving it as a disconnected, informational-only artifact that a student completed but that has no bearing on their actual standing in the course.

What’s shown in the real screenshot above avoids both failure modes: a dedicated Pending count visible at a glance across every assignment in the catalog, a direct Grade action that opens the specific submission queue, and, per the Gradebook’s own description, a real, weighted contribution to the course’s overall grade alongside quizzes and standards packages. The difference isn’t a matter of taste; it’s the difference between an assignment feature an instructor can actually run a real grading workflow through at scale, and one that only looks functional in a single-submission demo.

That spread in quality is exactly why we gave the best LMS platforms with assignments their own dedicated roundup, judged on the grading workflow rather than the “supports assignments” checkbox.

Three real situations where this is the actual assessment method

The technical skills course where “did you build it” is the only honest test. For anything involving real code, real design work, or real technical troubleshooting, the game development, blockchain, and networking assignments visible in the real data are exactly this, a working submission is the only assessment format that actually verifies competence. A multiple-choice question about Unreal Engine’s Blueprint system tests whether you memorized a fact about Blueprints; a code review exercise tests whether you can actually work with them.

The marketing or business course assessing applied strategy. “Design a Content Marketing Solution,” “Design an Email Marketing Solution”, these assignments ask a student to apply a framework to a specific scenario, producing a document or plan that demonstrates whether they actually internalized the strategic thinking the course taught, not just whether they can define the terms.

The instructor who needs to build real rapport with serious students. Grading assignments personally, with real written feedback, is one of the few remaining points of direct, individual contact between an instructor and a student in an otherwise asynchronous course. For an instructor building a reputation and a following, connecting to the certificate-sharing growth loop covered elsewhere in this series, the quality of that personal feedback loop is part of what makes students trust and recommend the course to others.

Three grading workflows worth naming explicitly

Given how central this feature is, it’s worth spelling out three concrete grading routines an instructor or admin could adopt, rather than leaving “manage the queue well” as a vague suggestion.

Batch grading by course, once a week. Rather than reacting to submissions as they trickle in throughout the week, an instructor blocks out one dedicated session, say, every Monday morning, and works through every pending item across all their courses in one sitting. This trades a small amount of response-time delay (a student submitting Tuesday waits until the following Monday) for a large reduction in context-switching cost, since the instructor stays in “grading mode” for the whole session rather than fragmenting their attention across the week.

Same-day grading for time-sensitive courses. For a course tied to a live cohort or a hard deadline, a bootcamp module gating the next live session, for instance, same-day grading turnaround might be a genuine requirement rather than a nice-to-have, since a delayed grade could block a student’s progress into content they’re expected to be ready for on a fixed schedule.

Triage by pending count, then chronological within each assignment. For an admin managing grading across an entire multi-instructor catalog rather than their own courses specifically, sorting attention by which assignments have the largest pending backlogs, the SEO Mastery assignment’s 15 pending, the Unreal Engine assignment’s 14, and clearing those first, then working chronologically within each one (oldest submission first), is a reasonable prioritization that keeps any single assignment’s backlog from growing indefinitely while an instructor focuses elsewhere.

Common questions before you build your first assignment

Can students attach files, or only submit text? The subtitle states submissions come as text from the lesson page, if file attachments matter for your specific assignment type (a design file, a code repository link, a document), confirm directly whether the submission format supports that or expects a link pasted into a text response instead.

What happens to a submission if I edit the assignment’s instructions after students have already submitted? This is worth testing directly before relying on it, editing an assignment’s prompt after 15 students have already submitted against the original wording could create confusion about which version of the instructions each submission was actually responding to, so a deliberate policy (only edit before the first submission arrives, or clearly version your assignment prompts) is worth adopting proactively.

Can more than one instructor grade the same assignment queue? In a multi-instructor course, covered elsewhere in this series, whether co-instructors share access to the same grading queue or only the primary instructor can grade is worth confirming directly against your specific permission setup, since grading access likely follows the same instructor-permission model that governs course editing.

Is there a way to see all pending assignments across every course in one place, not course by course? The real screen shown here already aggregates all assignments across all courses into one list, with the Pending column visible for every row, meaning an admin doesn’t need to check course by course, the whole catalog’s outstanding grading work is visible on this one screen.

Do students get notified when their assignment is graded? This wasn’t directly confirmed in what’s described, but given the notification infrastructure covered elsewhere in this series (the module catalog explicitly lists “Notifications” as covering “grading events”), a graded assignment is a reasonable candidate for triggering the same in-app bell and email notification system used for other course events, worth confirming this specific trigger is active if timely feedback delivery matters for your course.

Can an assignment be reused across multiple courses the way lessons and quizzes can be attached elsewhere? This wasn’t directly confirmed in the real screen examined here, since each assignment row shows a specific single Lesson and Course pairing. If you’re building similar assignment prompts across several related courses, a generic “debug this codebase” style task that fits several different technical courses, check directly whether the platform supports reusing one assignment definition or expects each course to have its own independently created copy.

What to actually do with a real pending queue like this one

Looking at the genuine numbers on this install, 15 pending on one assignment, 14 on another, 22 on a third, here’s the practical response:

• Treat the Pending column as a standing priority list, sorted by whichever assignment has waited longest, not graded in whatever order feels convenient. A student who submitted three weeks ago and one who submitted yesterday should not wait the same amount of time for a response.
• Set an internal grading SLA and hold yourself to it. Whether that’s 48 hours, one week, or something else, the specific number matters less than having one at all, a stated expectation is what turns “I’ll get to it eventually” into an actual commitment students can plan around.
• Watch for assignments with unusually high pending-to-submission ratios. An assignment sitting at 22 pending out of 47 total submissions has a bigger backlog than one sitting at 4 pending out of 8, both proportionally and in absolute terms, and deserves attention first if grading capacity is limited.
• Use consistent max-point scales across assignments in the same course, using a uniform point scale across assignments, so the weighted gradebook calculation covered elsewhere in this series produces a class average that actually means something comparable across different assignments.

The real value of a genuinely functional assignment system isn’t the grading interface itself, it’s what it makes possible for the course content around it. A course that can only assess understanding through multiple-choice quizzes is quietly limited to teaching only what can be tested that way. A course with real, human-graded assignments can teach, and honestly verify, anything a student can actually produce evidence of doing, which is most of what makes a skill worth calling a skill in the first place.

The scale problem hiding behind “101 assignments”

It’s worth sitting with what 101 assignments across one catalog actually implies operationally, because the number is easy to skim past. The real screen shows this list paginated at 20 per page, spanning 6 pages, meaning an instructor or admin managing this catalog’s grading load can’t just eyeball a single screen and know what’s outstanding; they need the Pending column, and ideally some way to sort or filter by it, to make that backlog actually manageable rather than requiring a mental tally across six separate page loads.

This is exactly the kind of “big-site readiness” question worth asking about any admin screen that lists a row per content item: does this hold up at 20 items, and does it still hold up at 200? A list screen that works fine with five test assignments in a demo but has no sorting, no filtering, and no way to prioritize by urgency quietly becomes unusable the moment a real catalog crosses a few dozen items. The presence of a clear, visually distinct Pending count on every row, rather than requiring a click into each assignment to discover how many submissions are waiting, is the specific design choice that keeps a large, active catalog navigable rather than overwhelming.

What this doesn’t solve: grading consistency across instructors

It’s worth being honest about a real limitation, rather than presenting every feature in this series as flawless. Nothing in what’s described here addresses a genuine challenge that shows up the moment more than one instructor is grading assignments in the same course or catalog: two different graders can reasonably score the same quality of work differently, absent a shared rubric both are grading against. A 100-point max score, consistent across assignments as shown in the real data, standardizes the scale everyone grades on, but it doesn’t by itself standardize the judgment applied to fill in a specific number on that scale.

This isn’t a flaw unique to Learnomy, it’s a structural challenge with any human-graded assessment system, on any platform. The practical mitigation, which the platform enables even if it doesn’t enforce it, is for a course’s instructors to agree on and write down explicit grading criteria before opening the queue, what does a 100-point response actually contain versus an 80-point one, and refer back to that shared standard consistently, rather than relying on each grader’s individual, unwritten sense of quality. A site owner running a multi-instructor academy, connecting back to the commission and instructor-management systems covered elsewhere in this series, has a real incentive to standardize this explicitly, since inconsistent grading across instructors is exactly the kind of thing that erodes student trust in a marketplace’s overall quality.