From analysis to solution — the architecture of a recommendation
Issue tree → prioritised levers → one decisive recommendation with numbers attached.
Losing teams present "three options for the client to consider." Winning teams say: "We recommend B. Here is the math, here is the plan, and here is why A and C don't hold up." Judges reward decisiveness — optionality reads as fear.
TL;DR · Key Takeaways
Key takeaways
- Structure MECE, generate 8–12 levers, then kill 70% on an impact × feasibility matrix — and show the kill visibly in the deck.
- Recommend in 2–3 pillars (quick win, core move, structural bet) carried by a financial spine: cost, return, payback, sensitivity.
- Phase the roadmap, name your top risks before the judges do, and pre-write the kill question into the appendix.
The build, in four moves
Move 1–2: structure, then kill your darlings
Structure the ask with the same MECE discipline as a case interview — the core frameworks and issue-tree method transfer directly. Competitions differ in one way: you have days, not minutes, so the tree should generate 8–12 candidate levers before you choose. Then prioritise ruthlessly on an impact × feasibility matrix and show the matrix in the deck — judges score the visible discipline of rejecting your own ideas almost as highly as the ideas you kept.
Move 3: the financial spine
Every recommendation needs four numbers: what it costs, what it returns, when it pays back, and what happens if your key assumption is half wrong. "This will improve profitability" is air; "₹14 crore investment, ₹46 crore incremental revenue by year 2, payback in 11 months, still positive at half the assumed adoption" is a recommendation. Build the model bottom-up in a sheet, keep it simple enough to defend live, and put the headline numbers on the recommendation slide — the workings go to the appendix.
The innovation–feasibility trade
Rubrics often weight "innovation" ~30% — which tempts teams into metaverse-flavoured fantasy. The trick: innovate inside one pillar (the structural bet) while keeping the other pillars boringly executable. The deck reads as both imaginative and adult.
Move 4: roadmap and risks
Phase the plan
30/90/365 days, each phase with an owner function (sales, ops, digital), a milestone, and the KPI it moves. A Gantt-style strip on one slide is enough.
Name the top three risks yourself
For each: likelihood, impact, mitigation, and the trigger at which the mitigation fires. Every risk you name and answer is a Q&A grenade defused in advance.
Pre-write the kill question
Ask: "if a judge wanted to destroy this in one question, what would they ask?" Write that question and its answer into the appendix. It will be asked.