Scored quiz, scorecard, or outcome quiz: the three types of online quizzes score differently, fail differently, and are not interchangeable. A decision table, the failure mode of each, and the four steps that settle the format before you write a single question.
Most quiz projects go wrong in the first ten minutes, before anyone writes a question. Someone says "let's build a quiz," three people picture three different things, and the format ends up decided by whichever template happened to be open. Then the arguments start: why does the result page show a number, why does nobody share it, why are the leads useless.
There are three types of online quizzes worth building. They score differently, they fail differently, and they are not interchangeable.
Ask this before anything else: what is the respondent supposed to learn about themselves?
- There is a correct answer and they either know it or they don't → scored quiz
- There is no correct answer, but there is a "more" and a "less" → scorecard
- There is no correct answer and no scale, only kinds → outcome quiz
That is the whole taxonomy. Everything below is consequences.
Scored quizzes: there is a right answer
Each question carries points, correct answers earn them, and the result is a number against a maximum — often with a pass mark. The respondent finds out whether they know the thing.
This is the right form for training verification, certification pre-checks, first-round hiring screens, and any lead magnet where your authority is the hook. "Test your X knowledge" works because the reader suspects they'd fail and wants to check privately.
Design notes that matter more than they sound:
- One unambiguous correct answer per question. If two options are defensible, you're no longer measuring knowledge, you're measuring whether they guessed your intent.
- Explain on the result page, not just the score. A bare "7/10" teaches nothing. Per-question explanations are what makes someone forward it to a colleague.
- Keep it 10–15 questions. Long enough to feel real, short enough to finish.
If the same quiz gets taken repeatedly — cohort after cohort of trainees, or candidates who talk to each other — pull the questions from a bank and randomize. A fixed twelve-question set has a shelf life measured in weeks once the answers start circulating.
The failure mode is emotional. A low score is a judgment, and people who feel judged don't book calls. If your commercial goal is a conversation rather than a filter, the next two forms are safer.
Scorecards: how much, across dimensions
A scorecard drops the notion of correct. Questions map to dimensions, each dimension carries its own subtotal, and the result is a profile: strong here, thin there, one number on top if you want it.
This is the workhorse for readiness assessments, audits, B2B lead magnets and coaching intake. It's also the form that survives a sales conversation, because "your Process score is 4/10 and everything else is 8" is an agenda, not a verdict. A ready-to-clone example: How Strong Is Your Leadership? Leadership Scorecard — five dimensions, subtotals, a result page built to be discussed rather than admired.
Two rules keep scorecards readable:
- Four to six dimensions. Below four, a single score would have done. Above six, nobody reads the report and the radar chart turns into a spiderweb.
- Name dimensions as things you can act on. "Process" and "Pipeline" beat "Factor 2." If a dimension doesn't imply a next step, cut it or merge it.
The failure mode is dimension sprawl: twelve dimensions, forty questions, a completion rate that collapses in the middle. If you're seeing drop-off mid-quiz, check the length before you touch the copy — the funnel benchmarks piece covers where to look and what counts as a normal stage-to-stage rate.
Outcome quizzes: which kind
No score at all. Answers tally toward types, the highest tally wins, and the respondent gets an identity: an archetype, a style, a persona. What's Your Career Archetype? is the shape — six questions, one of several named results, built to be screenshotted.
This is your top-of-funnel form. It's the only one of the three people share voluntarily, because the result is flattering-adjacent and says something about them. It's also the cleanest segmentation you'll ever get: the type is the tag.
The rule that separates a commercial outcome quiz from a horoscope: every result must imply a different next action. Four types that all end at the same "book a call" button is a personality test with a CTA stapled to it. Four types that route to four different offers, sequences, or resources is a segmentation engine. If you can't write a distinct next step for each type, you have too many types.
Three to five types is the working range. Two reads as a coin flip, and past five you're splitting an audience you don't have the follow-up content to serve.
The failure mode is charm without consequence — high completion, high shares, nothing downstream.
The decision table
| Scored quiz | Scorecard | Outcome quiz | |
|---|---|---|---|
| Result says | "You got 7/10" | "Strong on X, thin on Y" | "You're a Builder" |
| Scoring model | Points for correct answers | Per-dimension subtotals | Tally toward types |
| Typical length | 10–15 questions | 15–25 questions | 6–10 questions |
| Funnel position | Middle — qualifies | Middle to bottom — diagnoses | Top — attracts |
| What it segments by | Competence level | Weakest dimension | Type |
| Shareability | Low | Very low | High |
| Natural CTA | "Here's the course" | "Here's what to fix first" | "Here's the path for your type" |
| Fails when | The score reads as judgment | Too many dimensions | Types imply the same next step |
Read the "fails when" row twice. Most bad quizzes are a good quiz in the wrong format.
What the format changes downstream
Picking the form isn't a content decision, it's a plumbing decision. Three things change the moment you commit:
Lead scoring. A scored quiz gives you one number to threshold on. A scorecard gives you a weakest-dimension field, which is usually a better buying signal than the total — someone scoring 3/10 on the exact thing you sell is a hotter lead than someone at 8/10 overall. An outcome quiz gives you a categorical tag and no ordering at all, so your scoring has to come from elsewhere.
Result-page CTA. One CTA for a scored quiz; one CTA per dimension-gap for a scorecard; one CTA per type for an outcome quiz. Building an outcome quiz and then writing a single generic CTA throws away the entire reason you chose the format.
Where the email gate goes. A scored quiz can gate the score and get away with it — curiosity is at its peak in the last five seconds. A scorecard usually shouldn't gate the whole report, because the report is the value and the sales conversation needs it read. Outcome quizzes should almost never gate: the share is worth more than the address. And if leads aren't appearing at all, that's typically one setting rather than a format problem — see why your quiz has no leads.
Pick in four steps
- Write the sentence the respondent should say out loud when they see the result. If it starts with a number, you want a scored quiz or a scorecard; if it starts with "I'm a…", you want an outcome quiz.
- Write the next action for each possible result. Can't differentiate them? Wrong format, or too many buckets.
- Count your dimensions or types. Four to six for a scorecard, three to five for an outcome quiz, one for a scored quiz.
- Only now write questions.
Ten minutes on those four steps saves the rewrite where you discover, at question thirty, that what you actually needed was a scorecard.
Photo by Raul Petri on Unsplash