Evidence-based case analysis
導入済み| English | 意味 |
|---|---|
| fact/fækt/ | 事実 |
| assumption/əˈsʌmpʃn/ | 仮定 |
| conclusion/kənˈkluːʒn/ | 結論 |
| review metrics/rɪˈvjuː ˈmetrɪks/ | レビュー指標 |
Which sentence can you trust?
- A report to the Nimbus owner says: "Customers love us. Demand is growing. We should open a third shop." Every sentence sounds confident.
- Yet none of them says how it is known. A case analysis is useful only when each claim shows its evidence and the recommendation can be checked later. This lesson puts the earlier tools together.
Fact, assumption or conclusion?
- A fact 事实 has a source, a measure and a period: "55% of customers bought again within a month (shop tills, March)". An assumption 假设 is treated as true for now: "the third shop will pay back in two years".
- A conclusion 结论 follows from facts and assumptions under stated criteria. Numbers without a source are assumptions wearing a fact's clothes.
Fact, assumption or conclusion?
Sort each sentence from the Nimbus report. A fact shows its source, measure and period; an assumption is a guess treated as true for now; a conclusion follows from the others under stated criteria.
Which of these are facts, ready to use as evidence? Select all that apply.
Difficulty 2/5 · 140 coins + a small bonus on first completion today
The first two give a source, a measure and a period. The third has none of these, and the fourth is a projection, so both are assumptions.
Five parts, in order
- Situation: the key sourced facts. Problem: one decision question with real options. Analysis: the frameworks, each carrying evidence.
- Recommendation: one option, with its risks named. Review metrics 复核指标: the numbers and dates that will show whether it worked.
Put the parts of a case analysis in order.
Difficulty 3/5 · 200 coins + a small bonus on first completion today
Each part builds on the one before. The review metrics come last because they test the recommendation.
Weights before scores, then test them
- Fix the criteria and weights before scoring, as in the decision lesson. With weights 0.5, 0.3 and 0.2, pre-order scores 6.2 and the third shop 5.6.
- Then try other reasonable weights. If the winner never changes, the choice is robust. If it changes, you have found the real question to argue about.
With weights profit 0.4, risk 0.4 and staff fit 0.2, what does pre-order (scores 5, 7, 8) total?
Difficulty 2/5 · 140 coins + a small bonus on first completion today
$0.4 \times 5 + 0.4 \times 7 + 0.2 \times 8 = 2.0 + 2.8 + 1.6 = 6.4$. The third shop totals $2.4 + 1.6 + 1.4 = 5.4$, so putting more weight on risk widens pre-order's lead.
Where the answer flips
- With weights 0.7, 0.2, 0.1, which put more weight on profit, both options total 5.7: a tie. Giving profit even more weight would favour the third shop.
- A good report says this openly: "pre-order wins unless profit matters much more than risk". The reader can then judge the weights instead of trusting a single total.
If the winning option changes when the weights change, the scoring method has failed.
Difficulty 1/5 · 100 coins + a small bonus on first completion today
The method has worked. It has shown exactly which judgement, here how much profit matters compared with risk, decides the answer.
A recommendation you can prove wrong
- A useful recommendation names a threshold and a deadline that could show it was wrong. That is what makes it reviewable.
- "Run a Saturday stall for 8 weeks; stop if daily profit stays below £100" can be checked. "Expand, because demand is growing" cannot.
- Is the Saturday stall worth running?
- Expected sales: 150 items at £2.60 contribution each = £390 a day (assumption: 150 items, from the March trial).
- Costs: £45 pitch fee + £120 wages = £165 (fact: the market's price list and the rota).
- Profit: £390 − £165 = £225 a day. The review metric: stop if profit stays below £100 for three Saturdays in a row.
Which recommendation can be reviewed and proved wrong?
Difficulty 1/5 · 100 coins + a small bonus on first completion today
It names a threshold (£100 a day) and a deadline (three Saturdays, within 8 weeks). The others give no number or date that could show them wrong.
The stall sells 150 items at £2.60 contribution each. The pitch costs £45 and wages £120. What is the stall's profit for the day, in £?
Difficulty 2/5 · 140 coins + a small bonus on first completion today
$150 \times 2.60 = 390$, and $390 - (45 + 120) = 225$. Break-even is $165 \div 2.60 \approx 64$ items, so the 150 assumption has a wide safety margin.
Check the numbers behind the words
- Every claim about a business model can be tested with arithmetic: sales × contribution must cover fixed costs and leave a profit. Revenue, profit and cash must also be kept apart.
- If a claim cannot be turned into a number, ask what evidence would change your mind. If nothing would, it is an opinion, not analysis.
A manager says "the third shop is obviously right" but cannot name any evidence that would change their mind. What is the claim?
Difficulty 1/5 · 100 coins + a small bonus on first completion today
Analysis can be tested; if no result could change the view, it is not open to evidence.
Marks that slip away
- Unsourced numbers treated as facts. Give the source, the measure and the period.
- A recommendation with no review metric. Without a threshold and a date, nobody can tell whether it worked.
- Weights chosen after the scores. Fix them first, then test other weights openly.
- Listing frameworks with no evidence. Each one must carry data that affects the decision.
You've got it
- Sort every claim into fact, assumption or conclusion; a fact needs a source, a measure and a period.
- Structure the case as situation, problem, analysis, recommendation and review metrics.
- Fix the weights before scoring, test where the answer flips, back claims with arithmetic, and give the recommendation a threshold and a deadline.