Practice — Hot · Solution
1–2 · The instrument, and what it says
| Month (2026) | YoY Growth % |
|---|---|
| January | +14.33% |
| February | +24.68% |
| March | +23.07% |
| April | +10.13% |
| May | +13.02% |
| June | +15.92% |
The sequence is not climbing. It rises, peaks in February, falls to its low in April, and drifts back up. Six matched windows scatter between +10% and +25% around a center in the mid-teens. "Accelerating… and climbing" describes a staircase; the data is a wobble.
3 · February, explained
February's +24.7% is real arithmetic on a flattering base: February 2025 was the year's weakest month — $1.14M, the post-holiday trough from Lesson 1's very first line chart. Growing from the softest base of the year produces the tallest percentage of the year without any special performance in 2026. The one-sentence version for the CFO: "February looks spectacular because it is being compared against last year's worst month — that is a base effect, not a breakout." (March's +23.1% carries a milder dose of the same medicine.)
4 · The read
"The monthly series does not support 'accelerating': the six matched month-over-month-of-prior-year figures run +14, +25, +23, +10, +13, +16 — a scatter around the mid-teens, not a climb. February's +25% is a base effect against last year's weakest month, and April's +10% would falsify 'climbing' on its own. What the data does support is strong: growth in every single month, averaging comfortably above last year's +12%. I would say 'growth is running ahead of last year and broad-based across all six months' — every word of which survives an analyst's audit, which 'accelerating and climbing' would not."
Reflection answers
- Two true numbers, unsafe conclusion: a trend claim needs a series, and two points make a line through anything. The preventing visual is exactly the one you built — monthly YoY, matched windows, six points instead of two. The instrument is cheap; the retraction is not.
- Prior sightings of the same disease: May's file "inflation" in L03 (bigger than January because of season, not duplication — the comparison base was the wrong month), and the broken percent-of-total in this week's Medium (2025 revenue against an all-time denominator). A number is only as meaningful as what it is measured against — arguably the single most repeated lesson in this course, now with a proper name: the base.