McKinsey's Sustainable Futures Lab: the complete guide

What changed in 2026

The Sustainable Futures Lab is the third module in McKinsey Solve, the firm's gamified assessment. For its first months the module was a predictable exercise: 13 sequential multiple-choice questions in 20 minutes, each one a paragraph-length scenario about an environmental project, with every candidate seeing the same linear sequence. That original format is covered in full in the 85-minute section of this guide, and through mid-2026 it was the only version candidates received.

The redesign changes the genre. Test takers who sat the new version from August 2026 onward describe a role-play rather than a quiz: you manage a research project over three in-game days instead of answering 13 discrete questions. One taker who had prepared with written material for the old version put it plainly: the questions they studied were for the older version of the game, and what appeared on screen was something they had never seen.

Context that helps: Solve's lineup itself has shrunk. The original Ecosystem and Plant Defense games are retired from current invitations, leaving Redrock and Sea Wolf as the two constants and the Sustainable Futures Lab as the module that appears on longer invites. The 2026 redesign is the latest step in that churn, and it is still in phased testing rather than a confirmed permanent rollout. Since our first version of this page, enough detailed walkthroughs have landed, and we have reconstructed the format closely enough to build a simulator on it, that this guide now describes mechanics rather than just the broad shape.

The new format, day by day

You play the project leader of a four-person research team at an island field site. The theme rotates across environmental scenarios, coastal habitat work such as seagrass surveys, reef recovery, mangrove restoration or seabird colony counts, but the structure holds regardless of scenario. There are four workstations, for example a field survey station, a laboratory, a nursery or care site, and an analysis desk, and your four researchers map onto them: a senior field scientist who wants a clear goal and room to decide how to reach it, a meticulous lab technician who slows down when rushed, a keen junior researcher who is certified for more than you might assume and quick to feel left out, and a data analyst who is happiest at a screen and measurably weaker when posted to field stations. The three days build toward a deliverable, a funding review in one scenario, a ministry briefing in another.

The clock starts the moment you commit to the run, and the onboarding briefing plays on it, not before it. The briefing opens by asking you to rank its four sections in the order you want them answered, with the team roster of names and roles up front, and then serves the mission, the stations, the day rhythm and what carries forward in your chosen order. One 30-minute clock covers the briefing and all three days that follow. There is no per-day timer, so time you lose on day one comes straight out of day three, and at five minutes remaining a banner tells you plainly. The candidates who reported losing their first 5 to 10 minutes "just understanding the interface" were spending clock from the first screen. Read the onboarding properly and keep it moving: it is the cheapest time on the clock, and it tells you the loop before the loop starts.

Each day then runs the same rhythm: a morning briefing, then the four phases the game names Explore, Assign, Support and Reflect.

• Briefing. The day's objective, which stations carry the most weight (chips such as "Meadow Transects, 40%", heaviest first), how many explore questions your budget allows this morning, and your bench with each person's current energy and skills. On days two and three the briefing also restates where yesterday landed: each station's output and quality label, promises that come due today, and any lingering effects of your earlier decisions.
• Explore. Information comes from conversations, not cards. You pick a target from up to eight tiles, the four researchers and the four stations, with some targets closed for the day and clearly marked unavailable. Each available target offers exactly two fixed questions and answers only the one you pick, and each researcher answers at most one question per day: a researcher speaks to how they feel today or how they want to work with the others, a station to how its work is going or what new information has come in. The answer records itself in the Notes panel, and the good ones carry the gold: one might reveal the junior researcher is certified for work you assumed was off-limits, another that someone's morale has quietly dipped. The budget is three questions on days one and two and two on the final day, so the trained skill is who to talk to and what to ask, not which paragraph to read.
• Assignments. You place all four people across the four stations, or bench someone. Every station shows its weight in the day's output, the skills it needs, its energy cost and whether it is weather-sensitive; the game will not let you confirm a partial assignment. Confirming also settles yesterday's promises on the spot: a kept promise lifts the person's morale, a broken one costs double. Benching is a real tool, it recovers energy, and the analysis desk is where people get their breath back, so running everyone into the ground every day is a choice with a price.
• Support. Every day each of the four researchers brings you one request: twelve live requests across the lab. Pending requests show as badges on the person cards, and when more than one is waiting you choose who to address first; a "make another selection" affordance puts an open request back without spending it. Each request carries three options, and the options lean in different directions, deciding on the spot, consulting the team, deferring to the specialist, and the mix you choose is visible to the scorer (more on that below). Some options are commitments the game will check the next morning, honouring a promised lighter day, for example. You may also close the phase with a request still waiting: it is recorded as skipped, scores nothing, and being left waiting is itself something the person feels.
• Reflect. The day closes with the game asking each person one question: "After today's work on the project, I'm feeling...", and you answer Doubtful, Neutral or Confident, or I don't know. This is not a formality. Your answers are scored against how each person's day actually went, and the honest I don't know is a first-class answer: for a person you never engaged with it can beat a guess. The evidence is how people talked during the day, the pressure they described, the tone of their request for help, not the output numbers alone.

Day two opens with a disruption, a storm, an equipment failure or a supply problem, announced in a banner that states both numbers, and on the disruption day it also arrives as one researcher's support request: the station's output is cut (to half, in the storm case) unless the person you post there carries the stated skill, which softens the hit (to 80%, for the storm). The mitigation is knowable in advance, it is exactly the kind of fact the day-one questions can surface, so the disruption rewards candidates who asked the right questions rather than the lucky ones.

Two more texture details from taker reports. First, the onboarding is text-led and the early candidates were brutal about it, one called it the worst-explained tutorial they had ever sat, though the loop itself is simple once you have seen it once (it is on this page, which is the point). Second, the build is graphically heavy for a recruitment test, enough that one candidate's older laptop struggled and another had to borrow a computer after a graphics check blocked scheduling. Run the technical check on the machine you will actually use, well before test day.

Will your invite include it?

Your invitation email is the best predictor of what you will face. Candidates currently receive one of three Solve configurations, and the stated duration maps cleanly onto the module list. The 95-minute invitation is the one that carries the redesigned Sustainable Futures Lab; its email states the assessment takes approximately 95 minutes and suggests setting aside an additional 20 minutes to read all instructions and complete the technical check.

Which configuration you draw appears to be randomized during the phased rollout. One September taker polled their friends who applied around the same time and was the only one who received the new version. Candidates also report a post-assessment survey that asks them to review that new version in detail. Whether that session itself counts toward your Solve result has not been announced. Assume it counts.

What the new SFL is testing

McKinsey's official position on Solve has not changed: it is a fair, objective assessment of problem solving, taken alone, with results considered alongside the rest of your application. The firm publishes no mechanics or scoring breakdown for any module, and the redesigned Sustainable Futures Lab is no exception. What follows is our read of the format, from the walkthroughs and from building a working model of it, not official doctrine.

Read for what it rewards, six things stand out.

1. Output. Whether your stations actually produced. This is the matching question: did you put people with the right skills on the stations that carried the day's weight, with enough energy left to work them.
2. Plan fit. How close your assignment plan was to the best plan available on the information you had. You can produce respectable output with a mediocre plan; the scorer separates the two.
3. Judgment. The quality of your twelve live requests. The options are not equal, and the strong ones tend to share a shape: act, flag the uncertainty, keep the specialist trusted, fix it properly tomorrow.
4. Consistency and follow-through. Whether you keep the promises your options created, and whether your leadership style varies with the situation. Answering every request with "put it to the team" reads as a pattern, and the format notices.
5. Information use. Whether the questions you asked changed anything. Asking the high-value question and ignoring the answer is budget spent for nothing; the walk-through scores we have seen reward acted-on information most.
6. Team read. How well your end-of-day feeling reads match how each person's day actually went, and whether you knew when you did not know: the game credits an honest "I don't know" for a person you never engaged with. This is calibrated self-assessment: can you tell what you observed from what you assumed.

The behavioural traits the original version scored, prioritisation, decision-making under uncertainty, interpreting messy information, balancing trade-offs, and team effectiveness, all still apply. The difference is the medium: instead of answering questions about delegation, you delegate, watch workloads, keep people unblocked, and then report on how well you actually read your team. It is a work sample of a first week as an engagement junior, compressed into 30 minutes. That also explains why takers find it disorienting rather than hard. The complaints target unexplained mechanics and the interface, never the content.

How to prepare right now

Unlike when we first wrote this page, the new format is now practicable. In order of payoff:

1. Learn the loop and the clock before test day. Briefing, Explore, Assign, Support, Reflect, three days, one 30-minute clock that starts when you commit, before the onboarding briefing. Walking in knowing the loop is the cheapest time you will ever save, and the pacing rule from the earlier version still holds: read once, commit, move on. Rereading is how a shared clock disappears.
2. Practise the real thing. Our Sustainable Futures Lab simulator now includes a team lab mode that models the three-day format, the Explore questions, the support requests, the day-two disruption, carry-over between days and the daily Reflect reviews, alongside the original 13-question format. A few timed runs will teach you more about budgeting 30 minutes across three days than any amount of reading.
3. Ask like a consultant, not a tourist. Spend your questions on the targets that could change an assignment or warn you about a person, let the small talk go, and act on what you hear. Write down answer facts the day you learn them, because the questions expire and the state does not.
4. Assign to the day's weights, not to preferences. A researcher who wants the field station may still be the wrong person for it, the data specialist measurably is. Bench people deliberately to recover energy, and staff the day-two disruption's mitigation skill, not just its station.
5. Read the room from evidence. Track each person's energy, mood and workload from minute one: who was benched, who got skipped for the good station, whose request for help sounded strained. Keep a running promise list so nothing comes due unnoticed. By day three your Reflect answers should be evidence, not vibes, and for the person you never engaged with an honest "I don't know" is evidence too.
6. Keep the sister modules sharp. Redrock and Sea Wolf appear on every invite and are unchanged. Running a Redrock simulation and a Sea Wolf simulation trains the pacing discipline that carries across all three modules, and our Solve tips cover the timing rules.

One reassurance worth keeping in mind: the takers who struggled describe confusion about mechanics, not impossible content. The one mildly positive early read called the session less hard than expected, just unexpected. Candidates who understand the format before the clock starts are playing a different game from those deciphering the briefing live, and the briefing tells you less than this page does.

On an 85-minute invite: the original 13-question format

If your invitation says 85 minutes, the SFL you sit is the original version: a 20-minute (1,200-second), text-based project-management module with no calculations, data journals or spreadsheets. You are placed inside an environmental research team, themes rotate between coastal habitat recovery, air quality and wetland restoration, and what McKinsey actually scores is how you navigate team dynamics, respond to new information, and handle trade-offs when no option is clearly right.

The structure is linear and non-branching. Every candidate sees the same 13-question sequence for a given scenario, the timer runs continuously from the briefing, earlier decisions trigger short feedback cards but never alter the path, and you cannot go back to a previous question or re-read the initial briefing once you have advanced:

• Question 1, prioritisation ranking: a single drag-and-drop task ordering four operational actions from highest to lowest priority. It tests whether you establish a hypothesis-driven direction early or default to gathering more input first.
• Questions 2 to 13, action questions: twelve sequential multiple-choice decision points, each with 2 to 4 full-paragraph options starting "You recommend…" or "You decide…". With roughly 90 seconds per question on average, time management is an implicit scoring factor throughout.

The module evaluates five behavioural traits, each appearing multiple times across the sequence, and McKinsey is not scoring right-or-wrong answers, it measures how consistently you respond as situations evolve:

1. Prioritisation, focusing on what matters most and sequencing actions effectively.
2. Decision-making under uncertainty, moving forward on incomplete information without analysis paralysis or overreaction to a single input.
3. Interpreting messy information, extracting meaning from conflicting data and telling real discrepancies from noise.
4. Balancing trade-offs, making competing valid perspectives explicit before committing.
5. Team and stakeholder effectiveness, balancing progress with inclusion and protecting credibility under pressure.

Four consulting mindsets reward marks consistently across the 13 questions: pragmatism and impact over perfection (an 80% model that saves critical days beats 99% accuracy at the expense of deadlines), hypothesis-driven thinking (act on plausible assumptions rather than waiting for exhaustive research), communicative inclusiveness (engage sub-teams and protect stakeholders without letting the team become an indecisive debating club), and balanced compromise (when trade-offs are extreme, lean toward middle-ground options that satisfy critical compliance or safety needs while maintaining operational velocity).

Frequently asked questions

Is the Sustainable Futures Lab still multiple choice?

Not always. The redesigned 2026 version is a three-day team leadership simulation. Its live support requests still present multiple-choice options, three per request, but the module is a simulation you play, not a sequence of questions you answer. The original 13-question multiple-choice format still appears on some 85-minute invitations.

How long is the new Sustainable Futures Lab?

Roughly 30 minutes, on a single clock that starts the moment you commit to the run, before the onboarding briefing. The briefing plays on the clock, and a banner warns at five minutes remaining. It rides a 95-minute Solve invitation alongside Redrock and Sea Wolf, and the invitation suggests setting aside an additional 20 minutes for instructions and the technical check.

Why does my Solve invite say 95 minutes?

A 95-minute invitation means three modules: Redrock, Sea Wolf, and the redesigned Sustainable Futures Lab. A 65-minute invitation is Redrock and Sea Wolf only, and an 85-minute invitation adds the original 20-minute version of the Sustainable Futures Lab.

Does the redesigned SFL count toward my Solve score?

McKinsey has not said. Takers report a post-assessment survey that asks them to review the new format, which suggests it is still being evaluated. The safe assumption is that it counts, so treat it exactly like a scored module.

Does information carry over between days in the new SFL?

Most of it does. Energy, morale, skills, revealed facts, promises and the effects of your decisions all persist, and the next morning's briefing restates yesterday's station results. What does not come back is the questions themselves: targets you consulted or skipped are gone once the day closes, and the game's answer log wipes at each day boundary, so write down anything you might need later.

Can you practice the new Sustainable Futures Lab?

Yes. Our Sustainable Futures Lab simulator now includes a team lab mode that models the three-day format, the Explore questions, the support requests, the day-two disruption, carry-over between days and the daily Reflect reviews, alongside the original 13-question format.