Ecosystem Building is McKinsey's older Solve game: you place species into a location so every animal has a legal food source and every condition is satisfied. McKinsey has described the design in public ("build a reef, save a species"). It was the headline module when Solve launched around 2020. It is not the default 2026 two-module sitting.
In 2026 the common invite is Redrock (~35 min) plus Sea Wolf (~30 min). Ecosystem Building still shows up in some longer ~85-minute, three-game variants, sometimes with Plant Defense or the Sustainable Future Lab. Solve overview. Read your invitation.
Last full pass: 24 August 2026. CoachNed is independent practice material, not affiliated with McKinsey. Do not use leaked "real ecosystems." Transfer practice only.
Default 2026 vs this game
| Ecosystem Building | Sea Wolf | Redrock | |
|---|---|---|---|
| Core task | Build a closed food chain in one habitat | Match species to sites with traits + interactions | Interpret studies, calculate, report |
| Typical 2026 role | Alternate / three-game variant | Common module 2 | Common module 1 |
| Classic fail | Producer missing; species in illegal terrain | Single-trait match; competitors stacked | Wrong base; over-claim |
| Prep transfer | Constraint satisfaction | Constraint then optimize | Chart math + caveats |
If your invite is ~65 minutes and two games, do not spend this week drawing food webs. If it is ~85 minutes and names ecosystem / reef / food chain, this page applies.
Independent ecosystem-style practice on CoachNed sits with the Solve family of sims (see the assessments hub on the Solve guide). Still not McKinsey's exam.
How to think about a reef (3-step test-day)
- List producers first. If nothing converts sunlight / nutrients into biomass, the chain is dead no matter how many predators you like.
- For each consumer, name its food and confirm that food is present and legal in this terrain. "Eats small fish" is useless if you only placed a shark.
- Check abiotic rules last-mile: depth, temperature, current, substrate. A legally fed fish on the wrong rock still fails.
You are not maximizing a single "score species." You are satisfying all constraints at once. Unsustainable chains (no food, illegal condition) are the usual zero, not slow play.
Worked example: six cards, one missing producer
Habitat. Shallow reef, warm, rocky. Illegal: deep-only species.
Cards you may place (you will not use all).
| Species | Role | Eats | Terrain |
|---|---|---|---|
| Zooxanthellae coral | Producer (photosynthesis) | Light / nutrients | Shallow rock |
| Turf algae | Producer | Light | Shallow |
| Parrotfish | Grazer | Algae | Shallow |
| Small wrasse | Mid | Invertebrates / turf | Shallow |
| Reef shark | Apex | Small fish | Shallow or mid |
| Lanternfish | Mid | Plankton | Deep only |
Bad build (looks "complete"). Coral + wrasse + shark. Coral is a producer, but the wrasse's diet on the card is invertebrates/turf, not coral. Shark eats small fish — wrasse might count — but you never placed a legal mid that actually eats a producer the wrasse can use. The chain has a decorative producer and a hungry middle. Lanternfish is illegal on a shallow reef anyway.
Legal build. Algae (producer) → parrotfish (grazer) → shark only if the shark card allows parrotfish as prey. If it does not, add wrasse if wrasse may eat turf and shark may eat wrasse. Coral can sit as a second producer for stability if the round scores diversity, but coral does not feed the parrotfish if the parrotfish card says algae.
The tell. Count arrows. Every non-producer needs a chain down to a producer that exists on the board. Deep lanternfish is a decoy for people who "want eight species" because old blog posts said eight. There is no fixed species count to memorize — Solve guide.
If the producer cards are both terrain-illegal, you cannot brute-force a chain. Leave the round honest rather than placing a deep fish on a shallow rock.
Failure modes unique to this module
- Preparing this as "the Solve game" in 2026. You may sit Redrock and Sea Wolf instead and waste the week.
- Predators first. Apex species with no food is a fail.
- Ignoring terrain. Deep fish on a reef.
- Memorizing "8 species." The constraint set is the test.
- Confusing it with Sea Wolf. Sea Wolf is multi-site matching; this is one habitat, closed chain.
- Leaked maps. Don't. Integrity plus the map will not match.
If you might get the three-game variant
Skim this page and one constraint-chain puzzle. Skim Plant Defense as resource allocation under waves (Solve guide). Skim Sustainable Future Lab as multi-objective tradeoff, not max-one-KPI. Then confirm modules on the in-test screen before you start — you often see names there even if the email was vague.
McKinsey has described the reef idea in public recruiting essays; that is not a license to treat forum screenshots as the 2026 item bank. Transfer the constraint habit. Do not host or chase "real" maps.
When Solve is done, the offer still lives in cases and PEI. Ecosystem Building does not replace either. A calm drill on filtering infeasible options is closer to this game than another unstructured YouTube walkthrough.