BCG Consulting Career Assessment (CCA): 2026 Guide

The BCG Consulting Career Assessment, or CCA, is an early recruiting assessment used in some BCG processes. BCG's public recruiting material describes it as a short, multidimensional assessment designed to surface potential beyond the application. The exact experience can vary by geography and cohort, and your invitation governs.

That distinction matters because third-party guides often present one vendor, one sequence, and one set of item counts as universal. The official BCG pages reviewed for this guide do not publish those details. Candidate reports are still useful for familiarization, but they should be labeled as reports, not converted into official facts.

This guide covers what the CCA is and how to prepare for it. If you have already decided to practice and want to compare the tools available, read which BCG CCA simulator to use instead. That page compares price, free access, and format match across providers.

Road to Offer's BCG CCA simulator focuses on what you can actually train: original, timed numerical practice across the reasoning families the CCA rewards, plus work-style familiarization so the ranking interface is not new on test day.

What does BCG officially say about the CCA?

Current BCG pages support three practical conclusions:

• Allow about half an hour unless your invitation says otherwise. BCG Mexico describes roughly 30 minutes, while office and cohort instructions can vary.
• Expect variation. Different BCG location and program pages describe the assessment differently, so a single internet blueprint should not override your invitation.
• Follow the invitation window. The current US page says candidates normally complete the CCA within 48 hours. Your email remains authoritative for timing, accommodations, device instructions, and technical support.

Public BCG material reviewed here does not establish a universal assessment vendor, item count, behavioral scoring model, camera rule, calculator rule, or section order. If your invitation includes an official practice link or system check, use that before any third-party resource.

Check you have the right assessment

This page covers the CCA. Third-party guides often use "BCG online assessment" loosely, and an invitation described that way can refer to the CCA, to the chatbot-style online case often called Casey, or in some processes to both. They reward different preparation, so confirm yours before you start:

• Your invitation names the assessment. It is the authoritative source for the name, the sections, the window, and the device rules.
• A short, multidimensional assessment is the CCA. Stay on this page.
• A chatbot or an interactive online case with exhibits is Casey. Use the BCG online case guide.
• Still unsure, or you were invited to both? The BCG online assessments overview walks through each one.

Preparing for the wrong format wastes a limited completion window. Confirm the name first, then prepare.

What format and timing are reported for the CCA?

BCG does not publish a section list or item counts. What follows is the structure consistently reported by candidates and by every major prep provider, which is the best available picture and is labelled as reported rather than official. Treat your invitation as the authority and expect variation by office and cohort.

Two things in that table are firmer than the rest. The 30-minute overall duration comes from BCG's own recruiting pages, cited at the end of this guide. The 8 questions in 12 minutes numerical block is the figure the named third-party prep pages in the sources below publish independently of each other.

Do not read that row as a description of Road to Offer's free diagnostic. Ours is also eight questions on one 12-minute clock, but it is four scored numerical questions plus four unscored work-style questions, because it practises both halves of the assessment in one sitting. The six timed numerical mocks behind full access are the surface that drills a numerical block at length.

The rest is softer. Nobody outside BCG publishes the scoring model, the number of dimensions behind the work-style sections, or whether the vendor is identical in every geography. A guide that states those confidently is guessing.

Where does the CCA sit in BCG's hiring process?

Candidates consistently report the CCA as an early screen, after the application and before live case rounds. Passing it is what moves you forward. It is not the case interview and it does not replace one.

That placement is the practical point: the CCA is worth preparing for because it gates everything after it, but it rewards timed numerical accuracy and a stable work-style reference frame, not case structuring. Preparing for it with case frameworks is preparing for the wrong stage. Once you clear it, the BCG case interview guide covers the live rounds and the BCG online case guide covers Casey.

Does everyone applying to BCG get the CCA?

No. Reports vary by office, programme and cohort. Some candidates receive the CCA, some receive the Casey online case, some receive both, and some receive neither. There is no public rule that predicts which you will get, so do not infer your process from someone else's Reddit thread. Your invitation is the only reliable signal, and it arrives with its own deadline.

If you have not received an invitation yet, preparing the transferable numerical skills below costs nothing and carries over to Casey and to live case math.

What do candidates report?

Across public candidate discussions, some North American candidates describe a numerical reasoning block alongside workplace-style choices. Ranking three plausible statements from most to least like you is a recurring report. Other accounts differ, which is why this guide does not promise that every CCA uses the same mechanic.

Candidate reports can help you avoid interface surprise. They cannot establish BCG's current scoring logic, hidden dimensions, or an ideal answer. A forced-choice screen also does not imply a simple personality score: the meaning of a response depends on the instrument and scoring model, neither of which BCG publishes for the CCA.

How should you prepare for CCA-style numerical reasoning?

The transferable skills are straightforward: percentage change, ratios and rates, capacity or scheduling constraints, and chart or table reading. The highest-value sequence is learn, then time.

1. Write the relationship before calculating.
2. Keep units visible throughout the setup.
3. Estimate the direction and range of the answer.
4. Calculate, then sanity-check the magnitude.
5. Under a shared clock, make a forward decision instead of endlessly rescuing one question.

Road to Offer's free diagnostic is eight original questions on one 12-minute clock with forward-only navigation: four scored numerical questions and four unscored work-style questions. The free result shows every outcome and a worked solution for all four numerical questions. The active bank covers four transferable families: scheduling, reallocation, quantitative reasoning, and logical patterns.

The signed-in learning path adds the full mock, focused family practice, and a deterministic next-family recommendation based only on numerical accuracy and time.

A worked reallocation question

Reallocation is the family candidates most often lose time on, because the trap is setup rather than arithmetic. A representative question, original to Road to Offer:

Four analysts each spend 6 hours a week on manual reporting. A new tool removes 75% of that reporting work. The saved hours are redistributed equally across the four analysts for client analysis. How many hours a week does each analyst gain?

Write the relationship before touching a number:

1. Total reporting time = 4 analysts x 6 hours = 24 hours a week.
2. Time saved = 24 x 0.75 = 18 hours a week.
3. Gain per analyst = 18 / 4 = 4.5 hours a week.

18 and 4.5 are both correct numbers, and only one answers the question. The question asked per analyst, so 4.5 is the mark and 18 is the trap sitting one line above it in your own working. Under a shared 12-minute clock the arithmetic is rarely what costs you. Re-reading the final sentence before you answer is worth more than calculating faster.

A worked scheduling question

Scheduling items look like logic puzzles and are usually arithmetic with a hidden addend. A representative question, original to Road to Offer:

A project room is free from 09:00 to 13:00. Three sessions must run inside it: discovery (90 minutes), synthesis (60 minutes), and a client readout (45 minutes). The readout runs last and must finish by 12:45. Synthesis cannot start until discovery has finished. A 15-minute reset is required between sessions. Does it fit, and what is the latest discovery can start?

Place the fixed points before the preferences, and work backwards from the hard deadline:

1. Session time = 90 + 60 + 45 = 195 minutes.
2. Resets = two gaps between three sessions = 2 x 15 = 30 minutes.
3. Total = 195 + 30 = 225 minutes, which is 3 hours 45 minutes.
4. Readout must end at 12:45, so it starts at 12:00. A reset before it puts the end of synthesis at 11:45, so synthesis starts at 10:45. Another reset puts the end of discovery at 10:30, so discovery starts at 09:00.

It fits, and it fits exactly. The latest discovery can start is 09:00, the moment the room opens, with zero slack. The 30 minutes of resets is the whole answer: drop it and the schedule appears to have half an hour spare, which is the single most common way this family is lost. Count the gaps, not just the blocks.

A worked logical-pattern question

Pattern items reward translating each condition literally and then refusing to go past what the conditions actually support. A representative question, original to Road to Offer:

Four proposals, W, X, Y and Z, are each either approved or deferred, and each is either domestic or international. Exactly two are approved. Every international proposal is deferred. W and Y are domestic. X is international. Z is approved. What is determined about Z, and what is not determined about W and Y?

Translate each condition, then combine:

1. X is international, and every international proposal is deferred, so X is deferred.
2. Z is approved. The contrapositive of "every international proposal is deferred" is "every approved proposal is domestic", so Z is domestic. That is forced, not inferred.
3. Exactly two are approved and Z is one of them, so exactly one of W and Y is approved and the other is deferred.
4. Nothing in the conditions separates W from Y. Which one is approved is not determined.

The mark here is knowing that step 4 is the answer rather than a failure to finish. Under time pressure the expensive mistake is picking one of W or Y because the question felt like it wanted a single name. Read the contrapositive, take what is forced, and stop.

A worked quantitative-reasoning question

Percentage items are usually one relationship written in one line, wrapped in a number that does not matter. A representative question, original to Road to Offer:

A regional team's revenue fell 20% in Q1, then rose 25% in Q2. Support costs were flat at 180,000 across both quarters. Q2 revenue was 900,000. What was revenue before the Q1 fall?

Write the relationship first:

1. Call the starting revenue S. After a 20% fall it is 0.8S. After a 25% rise it is 0.8S x 1.25.
2. 0.8 x 1.25 = 1.00, so the two moves cancel exactly and Q2 revenue equals S.
3. Revenue before the Q1 fall was therefore 900,000.

Two traps sit in one short question. The support costs never enter the calculation: they are there to be carried through four lines of arithmetic that change nothing. And a 20% fall followed by a 25% rise reads like a net gain, when it is exactly break-even. Estimating the direction before calculating catches both.

What mistakes cost the most on the CCA?

• Answering the wrong question. Per-unit versus total is the single most common slip on reallocation and scheduling items, as the worked example above shows.
• Rescuing one hard question. One clock covers all eight. A question you cannot set up in 30 seconds is costing the three you have not seen yet.
• Treating work-style questions as a personality puzzle. There is no published ideal profile to reverse-engineer. Inconsistency across similar items is more visible than an imperfect answer.
• Preparing with case frameworks. The CCA is not a case. Profitability trees and MECE will not help you here; timed numerical setups will.
• Missing the window. BCG's US page describes a normal completion window measured in hours, not weeks. Confirm yours the day the invitation arrives.

How should you approach work-style questions?

Treat work-style familiarization as reflection, not answer coaching. Research on workplace forced-choice assessments shows why simplistic gaming advice is unreliable: several statements may all sound desirable, response formats differ, and the scoring model matters.

Use one stable reference frame: how you typically behaved in recent, comparable work situations. Read every statement before ranking. Choose the closest description of your usual behavior, not the most impressive sentence in isolation. Normal context differences do not prove inconsistency, and repeating a practice set is not a way to discover a hidden ideal profile.

Road to Offer's work-style module uses 91 original items across rank-three and self-rating formats so both interfaces feel familiar on test day. Its output is descriptive: it surfaces aggregate themes and neutral paired-context reflections from that practice session. Raw rankings and item IDs are not retained.

A worked work-style item

Rank-three blocks are the part candidates most often try to game, so it helps to see why the format resists it. A representative item, original to Road to Offer, in the two-step ranking shape candidates report and the Road to Offer runner reproduces:

Rank these three statements. Pick the one most like you, then the one most like you of the two that remain.

A. I rework a deliverable until the detail is right, even when the deadline is close. B. I keep a stakeholder informed even when there is no progress to report. C. I take on the part of a project nobody has claimed.

Notice what is missing: a wrong answer. All three describe behaviour a consulting firm can reasonably want, and that is the design. A rating scale lets a candidate agree with everything; forcing a full ranking of three removes that escape and makes the block measure which trait you will trade away, not whether you recognise a good trait.

That is why "pick the most consulting-sounding option" fails here. Every option sounds consulting-sounding. A candidate optimising block by block also drifts: A over B in one block and B over A in the next is the inconsistency these formats are built to surface.

The workable approach is the one this guide recommends throughout. Fix a reference frame before you start, your typical behaviour in recent comparable work, and answer every block against that same frame even when an option feels unflattering. Choosing D honestly is worth more than choosing C strategically.

What is a practical four-week preparation plan?

Week 1: establish the methods

Learn the four numerical setups without a timer. For every miss, label the cause as setup, units, exhibit read, arithmetic, or pacing. Run the free mock only after the methods are familiar enough to produce a meaningful baseline.

Week 2: isolate the weak family

Practice the family with the lowest accuracy. When accuracy ties, use the slower average completion time as the tiebreaker. This is the same deterministic rule the Road to Offer progress loop uses; behavioral responses never influence numerical readiness.

If you are unsure which skill is weakest, run the free CCA-style numerical simulator first. Its report separates the four numerical families so you can practice the weakest one. Comparing tools? See the best BCG CCA simulator, compared.

Week 3: add work-style familiarization

Complete a short rank-three preview once. Notice where context changes your instinct, then write down the reference frame you will use on assessment day. Do not memorize rankings or rehearse a supposed BCG answer key.

Week 4: rehearse the whole routine

Run a fresh timed numerical set in one sitting, review each walkthrough, and practice the weakest family once more. Confirm the deadline, device requirements, accommodations, and support contact in your invitation. Then move your preparation toward the online case and live interview stages relevant to your process.

How is the CCA different from Casey and live cases?

The CCA, the chatbot-style online case often called Casey, and live case interviews are different experiences. The CCA is an early assessment whose implementation varies. Casey-style online cases focus more directly on structured problem solving and recommendations. Live cases add a human interviewer and test how you communicate and adapt in real time.

Do not assume every candidate receives every step or receives them in the same order. Use the BCG online case guide for Casey-style preparation and the BCG case interview guide for live rounds.

What should you verify before starting?

• The exact deadline and time zone in your invitation
• The assessment name and any office- or cohort-specific instructions
• Supported browser, stable connection, and charged device
• Official practice or system-check links
• Calculator, camera, and proctoring rules stated for your assessment
• Accommodation instructions and recruiter or support contact

If an instruction in your invitation conflicts with a third-party guide, including this one, follow the invitation.

Sources and evidence boundary

• BCG US Associate recruiting: current public description of the US CCA, duration, and normal completion window
• BCG Mexico recruiting: current location-specific description and approximate duration
• BCG Empower FAQ: an example of cohort-specific CCA wording
• BCG interview process and BCG case preparation: official context for later interview stages
• Brown & Maydeu-Olivares, 2011: forced-choice formats and scoring considerations
• Cao & Drasgow, 2019: meta-analysis of forced-choice formats and response distortion

Reported-structure table: the section counts in the format table above (around 76 work-style prompts, eight numerical questions in 12 minutes, around 15 further work-style prompts) are not published by BCG. They are the figures that independent third-party prep pages state publicly, including CaseBasix's BCG CCA guide and Prepmatter's BCG CCA pages, both checked 2026-08-25, and they match what candidates describe. They are presented as reported for that reason and are deliberately not hyperlinked here. Your invitation overrides all of it.

Candidate discussions and third-party prep pages informed the reported-structure table and the recurring interface reports above. They are not authority for BCG's official current structure, scoring model, or vendor, none of which BCG publishes.

BCG, Casey, and CCA are trademarks of Boston Consulting Group. Road to Offer is not affiliated with or endorsed by BCG.