Ask a school why it lost a student and you will usually hear about motivation, or money, or a timetable clash. Ask the student and you often get something much more specific: it was too easy and they felt they were wasting an evening, or it was too fast and they spent twelve weeks quietly humiliated. Both are placement failures, both were decided in a ten-minute interaction before the course began, and both are cheap to prevent compared with the cost of replacing the student.
This cluster treats placement, level movement and progress reporting as one system, because they are. A framework defines what the levels mean, placement puts a student into one, progression moves them, and the report is how a family finds out any of it happened. Get the first right and the rest are maintenance. Get it wrong and no amount of good teaching in the wrong room will fix it.
Most small schools invent their own levels — Beginner, Elementary, Intermediate, and the perilous Upper Intermediate — and discover that the words mean different things to different teachers, which means a student moved from one class to another can arrive somewhere entirely unexpected. A framework fixes this by defining what someone at a level can actually do, in terms specific enough to disagree about: can hold a five-minute conversation about a familiar topic, can follow the main points of a recorded announcement, can write a short connected text.
CEFR is the most widely used of these for languages, and its great advantage is not precision but portability: a student who says they are B1 means something to a school in another country, to an employer, and to the next teacher. That is worth a great deal even where the boundaries are fuzzy, and the fuzziness is real — a learner is rarely the same level in speaking and writing, and a framework that hides that does you no favours.
SprintUp models this directly. A framework carries a name, a version, a subject and a list of available levels, and it declares what kind of progression it describes — proficiency, which is the CEFR shape, or grade, stage, band, or something custom you define. Frameworks can be ones the system provides or ones your organisation writes. Separately, a school keeps its own levels as simple groupings for organising classes. The distinction matters: the framework is what a level means, and the school level is where a student sits this term.
Schools that struggle with placement usually treat the test as the answer: the student scores, the score maps to a level, the student is placed, and if it turns out badly everyone blames the test. Schools that place well treat the test as one input into a decision made by a person, alongside two others — a short conversation, and what the student says they want the language for.
The conversation matters because written tests systematically misread certain learners. Someone who studied grammar formally for years can score well and be unable to speak; someone who has lived in the language can speak fluently and score poorly on a gap-fill. Placing either on the written score alone puts them in a room where they will be miserable for a different reason. Five minutes of talking tells you more about which class a person belongs in than forty minutes of multiple choice, and it costs the school almost nothing.
And the goal matters because placement is partly about fit rather than ability. A learner preparing for an exam in June and a learner who wants to talk to their grandchildren may test identically and want completely different classes. Asking is not a formality; it is how you avoid putting a person in a technically correct level they will leave.
A good placement test does not ask everyone the same forty questions. It asks a few, and where it goes next depends on how they went — which is how you get a useful reading in ten minutes instead of an hour, and how you avoid making a beginner work through twenty questions designed to separate B2 from C1.
In SprintUp this is built rather than bought, and it is worth being precise about that because the distinction matters when you are evaluating. There is no ready-made product called a placement test. What exists is the machinery to make one: a syllabus whose units carry progression rules, where each rule is a set of score ranges, and each range can carry a target and a message. A target either activates a particular unit for that student or routes them to a different course entirely. A range with no target is a deliberate ending — the assessment stops there for a student who lands in it. The rules are authored in the Curriculum Designer, and the rest of the platform reads them rather than writing them.
One behaviour worth knowing before you design: an adaptive assessment's units report on their own rather than being averaged together, which is correct — averaging a beginner's easy section with the hard one they never reached would produce a number that describes nothing. The first article covers what to put in the test, what to leave out, and how to keep it short enough that people finish it.
Some placements will be wrong, and the school's quality shows in what happens next. The failure mode is inertia: everyone can see by week three that a student is in the wrong room, nobody wants the conversation, and the student solves it themselves by leaving. Moving early is almost always right and is easiest in the first fortnight, before the student has friends in the group and before the move implies a judgement.
What makes it survivable is how it is framed and who raises it. Presented as an administrative correction — we placed you, we got it slightly off, here is a better fit — it is neutral. Presented as a consequence of the student's performance it is a demotion, whatever words are used. The second article covers both directions, including the upward move that schools forget to offer and that is one of the cheapest retention wins available.
Whatever the school knows about a student's level is invisible to the person paying unless somebody writes it down in language they can use. "72%" answers no question a parent has. What they want to know is whether their child is where they should be, what has actually improved, what is still hard, and whether to worry — and a report that answers those four in plain sentences does more for renewal than any discount.
The third article covers the shape of that report, the cadence that prevents unpleasant surprises, and the discipline of never letting a family hear bad news for the first time in a formal report. It also covers what not to include, which is most of what a platform can generate.
Build a placement test people finish, move the students you misplaced without losing them, and write a report a parent can act on. Each article ends with a checklist.