Data Interpretation Prompt Templates

**Customer Acquisition Cost Analysis: Q1 2024**

**Bottom line**: Our cost to acquire customers rose 20% year-over-year ($98 vs $82). Paid search is the biggest driver of the increase.

**Key finding**: We're over-reliant on paid search, which is getting more expensive. It's up 35% YoY and now costs 3x more than organic ($145 vs $45). Meanwhile, our most efficient channel (referrals at $28) drives only 12% of customers.

**What the data suggests**:

| Channel | CAC | Share of Customers | Trend |
|---------|-----|-------------------|-------|
| Paid Search | $145 | 45% | ▲ 35% YoY |
| Social Ads | $80-200 | 23% | Volatile |
| Organic | $45 | 20% | Stable |
| Referral | $28 | 12% | Stable |

**Recommended actions**:

1. **Invest more in organic**: Stable, low CAC, but under-indexed. SEO and content investment pays off over time.

2. **Scale the referral program**: Best CAC at $28, but only 12% of volume. Can we 2x this with better incentives?

3. **Audit paid search**: Which keywords/campaigns drove the 35% increase? Cut low performers before increasing budget.

4. **Stabilize social**: High variance ($80-200) suggests inconsistent execution. Focus on replicating winning campaigns.

**If we shift 10% of paid search budget to referral and organic**, we project blended CAC could drop to ~$88, saving approximately $150K annually at current volumes.

**Data note**: Attribution uses 30-day last-touch model. Multi-touch would show organic contributing more, strengthening the case for investment there.

**Checkout Flow Test Results**

**Plain English summary**: The new checkout flow (B) converts better than the current one. We're confident this is a real improvement, not random chance.

**The numbers**:
- Old checkout: 2.0% of visitors completed purchase
- New checkout: 2.3% of visitors completed purchase
- That's a 15% relative improvement (0.3 percentage points)

**How confident are we?**

Very. There's only a 3% chance we'd see this difference if the new checkout wasn't actually better. In statistical terms, p=0.03, which is below the typical 5% threshold for significance.

Think of it this way: If we ran this test 100 times when there was no real difference, we'd expect to see a result this strong only 3 times. So it's unlikely we're just seeing noise.

**What's the range of impact?**

The true improvement is probably between 0.1% and 0.5% (additional conversions). Our best estimate is 0.3%, but it could be a bit smaller or larger.

At our current traffic:
- 0.1% lift = ~500 extra conversions/month
- 0.3% lift = ~1,500 extra conversions/month
- 0.5% lift = ~2,500 extra conversions/month

**My recommendation**: Ship the new checkout. The evidence is strong enough to act on, and even the low end of the range (0.1%) is worth capturing.

**Caveat**: We tested for 2 weeks. There could be novelty effects that wear off. I'd suggest monitoring conversion rate for the next month after launch to confirm the improvement holds.

**Presentation Structure: The Onboarding Problem**

**Opening hook (1 slide, 30 seconds)**

*"We have a Feature X problem. Not with the feature itself, but with the fact that most customers never discover it."*

Show: Two numbers side by side
- Customers who use Feature X in week 1: 82% still with us at day 90
- Customers who don't: 28% still with us at day 90

This grabs attention with the contrast before explaining why.

---

**The mystery (2 slides, 2 minutes)**

*"We've been investigating why early churn is so high. Here's what we found."*

Slide 1: Churn curve showing steep drop in first 30 days
- Annotation: "47% of churn happens before day 30"

Slide 2: Comparison of churned vs retained users
- Churned users: Avg 2 logins, 1.5 features used, 0 Feature X usage
- Retained users: Avg 8 logins, 4.2 features used, Feature X in week 1

*"The retained customers aren't just more active. They're doing something specific."*

---

**The revelation (2 slides, 2 minutes)**

*"Feature X is the difference. But most users never see it."*

Slide 1: Feature X usage breakdown
- 60% of churned users: Never used Feature X
- 85% of retained users: Used Feature X

Slide 2: Current onboarding timeline
- Day 1: Account setup
- Day 3: Basic features intro
- Day 7: Settings walkthrough
- Day 14: Feature X mentioned (too late!)

*"We're burying our best feature. By day 14, half these users are already gone."*

---

**The proof point (1 slide, 1 minute)**

*"Users who discover Feature X early behave completely differently."*

Show: Retention curves split by Feature X adoption timing
- Used Feature X in week 1: Best retention
- Used Feature X in week 2-4: Medium retention
- Never used: Worst retention

*"This isn't correlation. Feature X delivers our core value. Without it, we're just another tool."*

---

**The ask (1 slide, 1 minute)**

*"We need to redesign onboarding to get users to Feature X in their first session."*

Proposed change:
- Day 1: After signup, guided Feature X experience (5 min)
- Success metric: 50% of new users try Feature X within 24 hours
- Expected impact: If we move from 25% to 50% week-1 adoption, we project 15% reduction in 30-day churn

*"This is our highest-use retention fix. I'm asking for priority in the next sprint."*

---

**Story arc summary**:
1. Hook: Dramatic contrast to create curiosity
2. Mystery: Show the problem without the answer yet
3. Revelation: Deliver the insight (Feature X is the key)
4. Proof: Demonstrate this isn't just correlation
5. Ask: Clear, specific action with expected outcome

Total: 7 slides, 7 minutes, one clear ask.

**Communicating Q2 Revenue Forecast to the Board**

**Recommended framing**:

*"Our best estimate for Q2 revenue is $4.2 million. Planning should account for a realistic range of $3.8M to $4.6M, depending on deal timing and economic conditions."*

---

**How to present the uncertainty**:

**Visual**: Show the forecast as a range, not a point

```
Q2 Revenue Forecast

 $3.8M $4.2M $4.6M
 |============|============|
 Downside Expected Upside
 (if deal estimate (if pipeline
 slips) accelerates)
```

**Language that works**:

✓ "We expect Q2 revenue of $4.2M, with a realistic range of $3.8M to $4.6M"
✓ "We're 80% confident revenue will fall between $3.8M and $4.6M"
✓ "The biggest swing factor is the [Customer] deal, which may close in Q2 or Q3"

**Language to avoid**:

✗ "Revenue will be $4.2M" (false precision)
✗ "Revenue could be anywhere from $3M to $5M" (too vague, undermines credibility)
✗ "We're not sure" (unhelpful)

---

**Scenario planning for the board**:

Instead of one number, present three scenarios tied to decisions:

| Scenario | Q2 Revenue | Likelihood | Hiring Implication |
|----------|------------|------------|--------------------|
| Conservative | $3.8M | 20% | Hold at current headcount |
| Expected | $4.2M | 60% | Add 2 roles in Q3 |
| Optimistic | $4.6M | 20% | Accelerate 4 hires |

*"I recommend budgeting for the expected case but not committing to hires until we see June pipeline."*

---

**Addressing specific risks**:

Don't hide them; frame them as known variables:

*"Three factors could move us within this range:*

*1. The [Customer] deal ($400K): 70% likely to close in Q2. If it slips, we're closer to $3.8M.*

*2. Renewal timing: Two renewals ($150K combined) are in final negotiation. Risk of delay is low but non-zero.*

*3. Economic conditions: Our model is based on current pipeline velocity. A slowdown in close rates would push us to the lower end."*

---

**Credibility statement**:

*"Our forecasting model has been within 12% of actual results for the past 6 quarters. The range I'm presenting accounts for that historical variance."*

This shows you know the model's track record without apologizing for it.

---

**Key principle**: Uncertainty isn't weakness. Presenting a range with clear drivers shows you understand the business. A single number with hidden assumptions is more risky for decision-making.