McKinsey Problem Solving Test (PST): Is It Still Used?

Sources: McKinsey's Solve page, the August 2025 Solve FAQ and its July 2025 game-lab article, all listed at the end. The two practice items below are original and fictional.

Is the McKinsey Problem Solving Test still used?

McKinsey says its game-based assessment, Solve, replaced an earlier multiple-choice test, and none of the McKinsey pages reviewed on October 11, 2026 describes the Problem Solving Test (PST) as part of current recruiting; McKinsey publishes no retirement date and no list of offices or roles. That is the official position. Whether some office somewhere still sends something else is not published either way, so the reliable test is your own invitation: it names the assessment and its length.

If the email says Solve, the PST is not what you will sit. If it names something unfamiliar, ask your recruiter, because McKinsey sends non-technical questions there. Our guide to McKinsey online assessments by role covers what is known beyond Solve, and the PSG, Imbellus and Solve naming guide maps each label candidates use to what it refers to.

What does McKinsey itself say about the old test?

In a July 2025 article about its game-based innovation lab, McKinsey says it had "a multiple-choice test that was very quantitative" before Solve, and that the aim was to measure "how people think, as opposed to what they know." The speaker is Brian Mendoza, who leads the lab. This is an official description, but it gives no date, no length and no name for the older test, which is why every PST detail beyond "multiple-choice and quantitative" is candidate-reported.

Two things follow. First, the PST and the game are different formats in McKinsey's own telling, so a guide that calls Solve "the PST online" is wrong. Second, the reason McKinsey gives is about method, measuring thinking in context, not about maths being unimportant. Our guide to what Solve measures in each game sets that out skill by skill.

How did the PST differ from Solve?

The earlier test was multiple-choice and, in McKinsey's words, very quantitative, while Solve is a set of game scenarios with an untimed tutorial before each task and a length stated in your invitation. The drawing below sets the two side by side on the points that decide how you prepare. Where McKinsey is silent about the older test, the left side says so rather than borrowing a number from a prep page.

Scoring is the point people most want and nobody can give. Neither side of the drawing has a published pass mark, and the Solve FAQ says the recruiter decides on multiple factors, so Solve alone does not settle an application. Our pages on the Solve pass rate and on how the Solve result is used keep to what is documented. Figures comparing PST and Solve pass rates circulate on prep sites, unverified, and this guide does not repeat them.

Which PST habits still help in Solve, and which do not?

Estimating first, naming the base of a percentage and picking the right figure from a dense exhibit still help most in Red Rock, as candidates describe it, while Sea Wolf rewards checking against rules and the Sustainable Futures Lab rewards ranking and linked decisions. This is an editorial judgement drawn from the task descriptions candidates report, not a McKinsey map. McKinsey names only Sea Wolf among the games, so each row rests on reported formats.

Read the columns, not just the rows. Red Rock is the game closest to what the old test trained, because candidates report percentages, ratios and chart reading there; the Red Rock math guide works those calculation types and the Red Rock practice method covers the Study and Cases. Sea Wolf uses averages, but to decide whether three chosen microbes fit a range, which is a rule check; see Sea Wolf math and the Sea Wolf cheat sheet. The Sustainable Futures Lab has almost nothing in common with a quantitative multiple-choice test, and its ranking task is a good example of why.

What does a PST-style question look like next to a Red Rock style one?

A PST-style item, as candidates describe it, hands you every number in a short paragraph and asks you to pick one of five answers; a Red Rock style item, as candidates describe it, makes you find and label the numbers in an exhibit first. The two items below are original, fictional and ours, and they use the same arithmetic idea so you can see exactly what changes. Neither is McKinsey content, and McKinsey says Solve scenarios vary between candidates.

The PST-style item

Compute year-one revenue first, then the combined factor, then check it the long way. The factor route is the quick one: each change multiplies the base, so the two factors multiply each other.

Choice 10% sits among the answers on purpose; multiple-choice formats work by putting the common slip on the list. That is the habit worth keeping from PST practice: know which wrong answer the item is built to tempt, and you will rarely land on it.

The same idea as a Red Rock style item

Now widen it. A reported Red Rock item gives you an exhibit for two channels and asks for a figure you assemble yourself. Here is an original exhibit for a fictional tea seller, followed by three questions: the online change, the total change and the online share of year-two revenue.

The extra work is in the set-up: revenue is not given, so build it for each channel and year before you compare anything.

The total needs both channels added before dividing. Then the trap changes shape: instead of adding two percentages, you are tempted to average two channel changes.

The arithmetic never went beyond what the first item taught. What was new was the labelling, the choice of which year and channel belong in each fraction, and the habit of noticing when two answers disagree by an order of magnitude (12.5% against 1.25%). The Red Rock practice questions work six more mini-cases in this spirit, and the Red Rock mistakes guide collects the setup errors that recur. McKinsey says Solve requires no preparation, so this is practice, not a requirement.

How should you use old PST material now?

Use old PST material, if at all, as a short arithmetic warm-up and nothing more: McKinsey says no preparation is needed for Solve, publishes no PST format to match, and the habits that carry over are small. The table below sorts the usual kinds of old material by what each is still good for. If the invitation says Solve, spend most of your preparation time on the three games instead.

Pick the games in order of risk to you. If arithmetic is your weak spot, the 35-minute Red Rock time budget and the Red Rock calculator and journal guide are the natural next stops. If you are choosing which game first, Red Rock versus Sea Wolf compares them, and the Solve prep plan fits a schedule around your date. If you are weighing McKinsey against other firms' tests, the McKinsey, BCG and Bain assessments comparison shows how different the formats are.

Road to Offer covers four practice formats: Red Rock, Sea Wolf, SFL Classic and SFL Island, with one complete run of each free. Red Rock includes the full Study and six Cases on a 35-minute practice clock; Sea Wolf includes all three sites and four stages per site on a 30-minute Simulation clock; Classic includes all 13 linked decisions without a countdown; Island includes onboarding and all three project days on a 30-minute Simulation clock. Start without an account, then create a free account to finish and keep the score and worked debrief. The first run and its worked review require no payment.

The free review explains where points went and each submitted decision, with highlighted calculations and charts in Red Rock, constraint and treatment visuals in Sea Wolf, decision and dependency diagrams in Classic, and evidence chains, placement maps and team feedback in Island. Learning mode adds a step-by-step path with coaching and feedback as you go across all four formats, untimed in the clocked games; it is paid access included with the matching game pack, the Solve pack or qualifying Annual access. The separate Solve plan is free with a free account: your invitation length and test date select dated practice steps for up to 28 days.

Start a free independent practice run

What if your recruiter or a forum mentions the PST?

Ask your recruiter which assessment applies to you, because McKinsey's FAQ sends non-technical questions to the recruiter and publishes no role-by-role assessment list. A forum post that says "I got the PST this year" is a candidate report: it may be a loose use of the old name for Solve, or something role-specific, and you cannot tell which from the post. Use the invitation as the tie-breaker, and use the FAQ pages for the support routes if a technical problem arises.

Context matters too. Experienced applicants and MBA candidates ask this most; the guides on experienced hires and MBA candidates say what is and is not published for each route. For a quick lookup of other candidate questions, the Solve FAQ in 30 short answers and ten forum claims checked cover the rest. When you are ready to set up your sitting, the invitation guide and the test-day checklist take over from here, and the complete Solve guide ties the games together.

Sources and evidence boundaries

• Official, checked October 11, 2026: the McKinsey Solve page and the August 2025 Solve FAQ for no preparation needed, untimed tutorials, the invitation stating the length and how the result is used.
• Official, checked October 11, 2026: McKinsey's July 2025 article on its game-based innovation lab for the earlier multiple-choice test, the Imbellus acquisition, Sea Wolf, the 300,000 figure and the January pilot of a new version.
• Official, rules context: the assessment integrity expectations, which apply to outside tools in any assessment.
• Reported, unverified: the names PST and Problem Solving Test, the older test's length, question count and format details, and which games a 2026 Solve invitation contains.
• Not published: a PST pass mark or pass rate, a retirement date, which roles or offices ever used it, and whether any office still does.
• Original teaching work: the tea packer item and the two-channel exhibit were written for this article. Their values come from the accompanying data file and are recomputed by its test. The skills map is an editorial judgement.