How Figure Us Out figures you out

Figure Us Out starts with about 85 questions and usually takes around 15 to 20 minutes.

But there is quite a bit more going on underneath those questions.

I wanted the assessment to get a detailed picture of someone’s personality without making them sit through a 60 to 80 minute personality test. So instead of giving everyone the same giant list of questions, I built an adaptive assessment that tries to figure out what it already knows, what it still needs to know, and where the next question will actually be useful.

The result is an assessment that can cover a much larger personality space while asking most people only a fraction of the questions that could potentially be asked.

About 350 questions worth of coverage in around 85 questions

If Figure Us Out used a traditional fixed questionnaire and asked every question that could potentially be useful, the assessment would be about 350 questions long.

That would probably take somewhere around 60 to 80 minutes.

I didn’t want to make anyone do that.

So a pretty large part of building Figure Us Out became figuring out how to get the information I needed without asking every person every possible question.

For most people:

• Full question space: about 350 potential questions
• Questions actually answered: around 85
• Estimated time for the full fixed question set: 60 to 80 minutes
• Typical Figure Us Out assessment: 15 to 20 minutes
• Reduction in questions: about 75%
• Reduction in time: about 75%

The goal wasn’t just to make the test shorter. It was to stop asking questions when they weren’t adding much useful information.

Basically, each question needs to earn its place.

The assessment adapts as you take it

The assessment is evaluating several parts of your personality at the same time.

Early questions help establish the broader patterns. As you answer more questions, some parts of your profile may become pretty clear while others are still close.

If the system already has enough information about one area, there isn’t much reason to keep asking you questions about it.

Instead, it can spend more of your time on the parts of your profile where another answer could actually change or clarify the result.

Underneath that are several parts of the routing and scoring system:

• Multidimensional routing: Several personality dimensions are being evaluated at the same time.
• Adaptive item selection: The questions you see later can depend on what your earlier answers have already established.
• Accumulated evidence: Important conclusions aren’t based on one answer. Evidence builds across related questions.
• Confidence thresholds: The system tracks how much evidence it has for different parts of your profile and whether there is enough to resolve them.
• Stopping criteria: Once an area is sufficiently resolved, the assessment can stop spending questions there.
• Disambiguation questions: If two possibilities are still close, the system can prioritize questions that are especially useful for separating them.
• Verification questions: Important conclusions can be checked from more than one angle instead of depending too heavily on a single question.
• Contradiction handling: If an answer doesn’t fit the larger pattern, the system doesn’t automatically throw it out or let that one answer override everything else. It can keep that part of the profile open longer and gather more information.
• Deterministic scoring: Your answers are evaluated using the same underlying scoring and routing rules every time. AI does not decide what personality type you are.

That also means two people can finish in about the same amount of time without necessarily answering the exact same questions.

If one part of your personality becomes pretty obvious early, there isn’t much value in asking another 15 questions to keep proving the same thing.

If another part is still close, that is where those questions are more useful.

What we’re actually measuring

Figure Us Out isn’t trying to reduce your entire personality to four letters or one number.

The assessment looks at multiple personality patterns and underlying traits, including:

• MBTI-style personality preferences
• Enneagram core type
• Enneagram wing
• The relative strength of different personality dimensions
• Evidence across related questions that helps settle your personality types
• Patterns that help distinguish between personality profiles that can look similar on the surface
• Areas where your results are especially strong
• Areas where you fall closer to the middle
• Combinations of traits that can change how a broader personality pattern actually shows up

That last part matters quite a bit.

Two people can both be ENFPs, for example, and still be very different people. One person’s preferences may be extremely strong while another person’s are much closer to the middle. Their Enneagram patterns can also be completely different.

So the type is useful, but it isn’t the whole profile.

The assessment uses that evidence to settle your MBTI-style preferences, Enneagram core type, and wing. Your report then explores that combination of types.

A little more technically

The general idea is related to computerized adaptive testing, but Figure Us Out isn’t simply moving up and down a single scale of difficulty.

Personality doesn’t really work that way.

The assessment is trying to resolve several dimensions of a personality profile at the same time.

As you move through it, the system accumulates evidence across questions, tracks which parts of the profile are becoming clear, identifies where meaningful uncertainty remains, and routes additional questions toward the places where another answer has the most value.

The architecture is custom, multidimensional, rules-based, and deterministic.

It combines:

• Adaptive item selection
• Multidimensional scoring
• Accumulated evidence
• Confidence thresholds
• Profile disambiguation
• Cross-question verification
• Contradiction handling
• Dynamic routing
• Stopping rules

So the assessment isn’t really trying to ask fewer questions.

It’s trying to ask fewer unnecessary questions.

How the scoring works

Your questionnaire answers are scored first.

AI does not read your answers and decide what personality you seem like.

The scoring engine evaluates your responses across the different dimensions being measured. Those results are then used to determine your MBTI-style personality type, Enneagram core type, and wing.

A few important things about that process:

• The assessment is scored deterministically.
• The same completed response pattern produces the same scored result.
• Your personality profile is determined before any AI-written interpretation is created.
• Evidence accumulates across individual dimensions before the system settles your types.
• Your MBTI-style type, Enneagram core type, and wing are considered together when creating your full guide.
• Couples comparisons use each person’s settled types, name, and pronouns. Raw questionnaire answers and individual questionnaire scores are not passed to the report writer.

The basic flow looks like this:

Your answers → deterministic scoring → personality profile → interpretation

The scoring determines the profile.

The interpretation explains it.

Why I built it this way

One of the things that has always bothered me about personality tests is the tradeoff between depth and time.

A short test is easy to take, but it can miss a lot.

A really thorough test can gather more information, but most normal people don’t want to answer hundreds of questions about themselves.

And honestly, they probably shouldn’t have to.

If someone’s answers have already made one part of their personality pretty clear, asking them 10 more slightly different versions of the same question doesn’t necessarily make the assessment better.

Sometimes it just makes it longer.

The adaptive system was my attempt to get around that problem.

Spend questions where they are useful. Stop spending them where they aren’t.

How I tested it

This part took a lot of work.

During development, I ran many thousands of synthetic assessment simulations while adjusting the scoring system, question weights, routing behavior, confidence thresholds, and profile boundaries.

I didn’t just want to test easy profiles where every answer pointed cleanly in the same direction.

Those are the easy cases.

The more useful tests were the weird ones.

So the development testing included things like:

• Clean, obvious personality profiles
• Profiles close to classification boundaries
• Similar-looking profiles that needed to be distinguished from each other
• Unusual combinations of traits
• Inconsistent response patterns
• Noisy answers
• Strong profiles mixed with misleading answers
• Dimensions sitting very close to confidence thresholds
• Cases designed to see whether the system stopped too early
• Cases designed to see whether it kept asking questions after an area was already clear
• Cases where one or two unusual answers conflicted with a much larger pattern
• Routing edge cases specifically created to expose weaknesses in the system

I was basically trying to break it.

Did it resolve something too early?

Was one question carrying too much weight?

Could a couple of strange answers pull an otherwise clear profile in the wrong direction?

Was it still asking questions after an area was already resolved?

Could two similar profiles actually be separated?

What happened when someone’s personality didn’t fit neatly into the most obvious pattern?

Every time I found something I didn’t like, I adjusted the system and ran it again.

The final validation run

Once I was happy with the assessment and scoring system, I ran a final validation using more than 1,000 synthetic assessment simulations across the personality profiles the system was designed to identify.

That final run produced greater than 98.5% agreement with the intended profile.

Most of the remaining differences were very close results, such as a single MBTI preference landing on the other side of a scoring boundary.

I was pretty happy with that.

But I also want to be clear about what that number means.

This was synthetic validation of the assessment and scoring engine. It was not a clinical validation study, and it does not mean Figure Us Out can identify a real person’s personality with 98.5% certainty.

What it tells me is that when the system is given response patterns representing the profiles it was designed to identify, it can reliably recover those profiles, including across a lot of the messy and borderline cases I deliberately threw at it.

That’s what I wanted to know before putting it in front of real people.

From your personality types to your report

Once the assessment has finished, your settled personality types become the foundation for your report.

Your charts, trait positions, and comparisons describe patterns associated with those types. They are interpretations of your personality profile rather than separate measurements taken from your questionnaire answers.

The longer written guides work a little differently.

There is a lot to explore in the combination of your MBTI-style type, Enneagram core type, and wing. Turning that into something that sounds natural, connects the different pieces, and is actually useful to read is where AI helps.

The report pipeline receives your settled personality types, along with your name and pronouns, to interpret your profile and write the guide.

It is not being handed your raw questionnaire answers and asked to figure out who you are from scratch.

That distinction is important to how the system works.

The scoring engine figures out the profile.

AI helps explain what that profile may actually look like in a real person.

Why the full guide can be more specific than the free report

The free report is meant to give you a useful introduction to your personality patterns.

The full guide explores your combination of personality types in more depth, with more room to connect the patterns and explain how they may show up in everyday life.

That includes things like:

• Your broader personality type
• Your Enneagram core type and wing
• How your MBTI-style preferences and Enneagram motivations may influence each other
• Strengths and recurring difficulties associated with that combination
• Communication, decision-making, and emotional patterns
• Practical examples that help you recognize those patterns in everyday life

So sometimes the full guide may describe something with more nuance than the free report.

That doesn’t mean your assessment changed.

It means we’re exploring the same personality types in more detail. The full guide does not require another assessment.

When two people take it

For couples, each person takes the assessment separately.

We don’t blend the two people together into some kind of average “couple personality.”

Both individual profiles stay intact.

Then the system compares them.

That lets us look at things like:

• Where the patterns associated with your personality types are similar
• Where they’re different
• How those differences may show up in everyday situations
• Where you may naturally understand each other
• Where the same situation may feel completely different to each of you
• Communication differences
• Decision-making patterns
• Emotional differences
• Practical differences
• Traits that may complement each other
• Places where one person’s natural tendency may run directly into the other person’s
• Areas where a difference may be useful in one situation and frustrating in another
• Patterns that probably wouldn’t be obvious from looking at either personality type by itself

The point isn’t to decide whether you’re compatible.

And there isn’t some secret perfect combination of personality types that makes a relationship work.

The point is to take two complicated personality profiles and make the differences between them easier to see and understand.

A lot of the time, something that feels personal in a relationship makes more sense once you can actually see the pattern underneath it.

Personality is still messy

After putting this much work into building a personality assessment, I should probably also say this:

No personality assessment is going to completely explain a human being.

People are complicated.

Context matters. Experience matters. Culture matters. Relationships matter. People change. We behave differently at work than we do at home. Some of our tendencies are extremely strong and others are barely preferences at all.

And sometimes people just do weird things that don’t fit their personality profile.

That’s normal.

I don’t think personality types should be boxes that people have to squeeze themselves into.

I think they’re useful models.

If the model helps you recognize something about yourself, understand your partner a little better, or finally put words to a difference you’ve been running into for years, then it’s doing what I built it to do.

Figure Us Out is an educational personality tool. It isn’t a clinical assessment, psychological diagnosis, or substitute for therapy.

See what all of this turns into

The assessment and scoring system are really just the foundation.

The useful part is what you can do with the information once it’s understandable.

You can look through our example reports to see exactly what the results turn into before deciding whether you want the full package.

Read the example guides →

You can also take the free individual assessment to explore your 16-type personality and Enneagram.

Take the free assessment →

About 15 to 20 minutes. No payment required.