This post is not about which course to buy. That decision is made — you enrolled, or your school handed you a licence, or a firm paid for it. Money is spent and the schedule starts Monday.
The live question is narrower and more urgent: does the course you own actually give you reps on the whole exam, or on the part of it that looks like the old one? And if there's a gap, how do you fill it without turning a 50-hour study week into a 60-hour one, which is the way every supplement plan dies.
One thing this post deliberately won't do is tell you what Barbri, Themis or Kaplan ship this cycle. Provider offerings change between exam windows, we sell a competing product, and a claim about someone else's 2027 course would be stale before your exam. So instead: here's how to audit your own portal in twenty minutes, and what to do with what you find.
The gap has a number, and the number is 51%
(Still deciding which course to supplement? The four-course comparison sets out what each includes for NextGen and what it costs today.)
Every NextGen session is the same shape: 40 multiple-choice questions, 2 integrated question sets, and 1 performance task, on one 180-minute clock. The scoring weight across the exam breaks down roughly as:
| Component | Share of the score |
|---|---|
| Standalone multiple-choice | ~49% |
| Integrated question sets | ~21% |
| Performance tasks | ~30% |
So applied and written work carries about 51% — the details are in our scoring guide. That single fact is what makes a course audit worth doing, because the failure mode isn't a course with bad content. It's a course with excellent doctrine and a practice library shaped like the legacy MBE, where every rep you take trains the 49% and none trains the rest.
The MBE-era instinct — grind multiple choice, treat writing as a formality you'll cram in June — is miscalibrated by half the exam. A course that inherited that instinct will pass it to you.
The twenty-minute audit of the course you already own
Open your portal and look for five specific things. This is deliberately different from the six questions worth asking before you buy — that checklist is for choosing. This one is for auditing something you already have, where the answer changes what you do tomorrow rather than what you purchase.
- Are integrated question sets and performance tasks playable, or are they PDFs? Find one of each and try to sit it. A PDF you read is not a rep. An integrated question set is a scenario that develops across items with a shared fact pattern; a performance task is 60 minutes against provided resources. If they only exist as printouts and model answers, you own the reading, not the practice.
- Count how many of each you actually get. Not "includes performance tasks" — the number. Then divide by the weeks you have left. Two performance tasks across a ten-week course is not preparation for a format worth ~30%.
- Do the MCQ sets include real select-two items? Six options, choose two, partial credit, at the real 1.8-minute pace. Search your question bank's filters for the format. If every question is 4-option single-answer, the bank is legacy-shaped whatever the marketing says — and select-two behaves differently enough that discovering it in the exam room is a bad plan.
- Is the doctrine flagged to the current content scope? NCBE distinguishes topics you must know cold from topics needing general familiarity. If your outlines teach every subject at uniform legacy depth, you'll spend February memorising things the exam decided you don't have to — see starred topics for what that actually changes.
- What is the written-feedback loop, concretely? Human grading, rubric self-grading with model answers, or nothing? Either of the first two works. "Nothing" is the gap that matters most, because written work is the half of the exam you cannot self-assess by counting right answers.
Write down which of the five come back thin. That list — not a brand — is what you're shopping for, if anything.
Substitute, don't stack
Here is the rule that decides whether a supplement plan survives contact with a real study week.
A full course is built to fill your day. Its schedule assumes it owns your available hours, and by design you're mildly behind on it. So the instinct — keep the course exactly as it is and add ninety minutes of extra practice — fails, reliably, around week two. Not because the plan was wrong, but because there were never ninety spare minutes.
Substitute instead. On the day your course assigns a block of legacy-format multiple choice on Evidence, do that block in NextGen format on the same subject, at the same length, in the same slot. Same subject, same hour, same place in the schedule; different item format. You are not adding a second course. You are changing what the practice hour practises.
Three rules make the swap safe:
- Swap format, never subject. The course's subject sequencing is its most valuable asset — it interleaves and spirals deliberately. Keep the calendar; change the items.
- Swap practice hours, not teaching hours. If a lecture is the reason you understand hearsay, keep the lecture. What's substitutable is the drilling that follows it.
- Protect one written rep a week from the start. One question set or one performance task, timed, self-graded against a rubric. Written work is where the ~51% lives and where cramming works least — a task attempted in week two teaches you something a task attempted in the final fortnight cannot.
What not to abandon
Supplementing goes wrong in the other direction too. Some candidates buy a question bank, discover it's more engaging than lectures, and quietly stop doing the course they paid thousands for.
Keep these:
- The calendar. External structure is most of what a full course sells, and it's the thing self-directed study most often lacks. Don't trade a schedule for a shopping list.
- Lectures on your weak doctrine. Being talked through a subject you don't have yet is a legitimate way to acquire it. Being talked through one you already know is the expensive rerun.
- Graded written work. If your course includes human grading, use every unit of it. A grader who scores against a real rubric is telling you something no analytics dashboard can, and you already own it.
- The essay/task model answers. Even where the practice is only a PDF, model answers plus an issue rubric make self-grading possible. That's the cheapest upgrade available to you and it costs nothing.
When the honest answer is "buy nothing"
We sell a question bank, so treat this section as the one where we argue against ourselves.
Don't buy a supplement if:
- Your audit came back clean. Playable question sets and tasks in real volume, select-two items in the bank, scope-flagged outlines, a working feedback loop. Then you have what you need and a second tool is a distraction with a price.
- You have under three weeks left. At that point new material is a net negative. Review your starred rules, finish the course's own tasks, and protect your sleep.
- You haven't used what you own. If the course's own question sets are sitting unopened, a second library will also sit unopened. Do the audit honestly: is the gap the product, or the hours?
- The money is genuinely tight. NCBE's own free materials are the calibration standard everyone should be checking against anyway, and they cost nothing. Start there, and see what prep actually costs before adding a line to the budget.
If the audit did come back thin on formats — and for a lot of courses built quickly for a brand-new exam it will — then what you need is reps in the two formats carrying half the score, not another set of outlines. That's the specific gap our $99 lifetime bank exists to fill, alongside a course rather than instead of one: all 30 question sets, all 15 performance tasks, select-two MCQs at exam pace, and no expiry if February becomes the plan.
Either way, do the audit before you spend. Twenty minutes in the portal you already paid for is the highest-return study block of your week — and if you want a calibrated read on where you actually stand first, ten free questions at the real pace will tell you more than another testimonial.
