Create placement test — place learners at the right level

A placement test needs a clear idea of what distinguishes the end levels. For languages, the CEFR (Common European Framework of Reference) with levels A1 to C2 is established; for IT topics, three to five levels such as beginner, intermediate, advanced usually suffice. Define concrete competences per level — what can someone do at B1, what at B2? This definition is the foundation; without it, the test becomes arbitrary.

Translate the competences into measurable tasks. For a B1 language level, test understanding everyday texts, simple conjugations, vocabulary in travel and work. Mix question formats: multiple choice for vocabulary, gap texts for grammar, short listening or reading comprehension tasks for application. Avoid pure translation tasks or open essays — these need human assessment and break the auto-evaluation. Plan 8 to 15 items per level, depending on differentiation needs. This brings tests of 30 to 60 items in total — a processing time of 20 to 30 minutes is realistic.

A classic linear test sends everyone through all items, from easiest to hardest. That is easy to implement but frustrating for participants: beginners fail at C1 items, advanced learners get bored with A1 trivia. Adaptive tests react to the answers and adjust the difficulty. Anyone answering four B1 items in a row correctly is upgraded to B2; anyone answering incorrectly three times in a row goes down one level.

For adaptive logic, use conditional logic in the builder: per difficulty band a block of 5 to 8 items, at the end a branch depending on score. Anyone passing the block goes into the next higher band; anyone not passing goes into the lower one. The calculation engine counts the correct answers and triggers the next branch. This makes the test significantly shorter (typically 15 to 20 minutes instead of 30 to 60) and more accurate in the middle of the level — so where most applicants lie. Keep the logic comprehensible: three to four bands suffice, anything more becomes complex and hard to maintain.

The real aha moment comes at the end of the test. Show the result clearly and positively: the determined level, a short description of the competences at this level and concrete examples of what the person can already do. Avoid school-like schemes like "60 percent correct" or "4 of 30 wrong" — most learners are sensitive in an assessment context and react more positively to a competence-oriented description.

Visualize the level on a scale so the learner sees where they stand. For languages, a bar with the CEFR levels and a marker is worthwhile; for IT topics a radar diagram with the tested sub-competences. Briefly justify the result: "You answered 9 of 12 B1 items correctly, but only 3 of 12 B2 items. This indicates solid B1 knowledge." This transparency creates trust in the result and reduces discussions with trainers. Additionally offer a PDF confirmation for download — many use the result as proof for applications or course registrations.

A placement test without a clear recommendation gives away its greatest value. Link the determined level directly with concrete course offers or learning paths. Anyone who lands at B1 sees a recommendation for the B1 advanced course with start date, duration, price and registration CTA. Anyone who lands at B2 sees the B2 course accordingly. This personalization significantly increases conversion from test to course enrollment.

Use conditional logic to show the recommendation appropriate to the level. Add two to three alternative paths — such as a full-time and an evening course, or an in-person and an online course. This way everyone finds the format that fits their own life situation. Hand over the result automatically to the learning management system via webhook so that the learner lands directly in the correct course on first registration. Additionally send an email with result and recommendation; this enables later follow-up and reduces drop-off between test and registration.