Month 9 · Compare campaigns cautiously
Four relative self-study weeks, fictional aggregate sources and one monthly activity. No advertising purchases, campaigns, real customer data or causal promises.
This is the developed Month 9 learning resource, not the entire twelve-month course. The four exercises form one monthly activity M09. Reading, personal marks and self-scores do not verify academic completion.
Complete fictional source F9
A and B are fictional sets. No real dates or comparable observation durations are supplied. Counts are not unique people or a traced funnel.
| Set | Visits | Inquiries | Other information |
|---|---|---|---|
| A | 100 | 10 | Spending/sales/origin comparability unknown |
| B | 80 | 12 | Spending/sales/origin comparability unknown |
Complete separate variants: V9A retains A and changes B to 80 visits/8 inquiries; V9B retains A and changes B to 0 visits/12 inquiries. No amounts, payments, tool intervention or customer identities are supplied.
Week 1 · Check comparability
Goal: Prepare a comparison without filling unknowns.
F9 supplies two fictional campaigns: A has one hundred visits and ten inquiries; B has eighty visits and twelve inquiries. These are aggregate counts, not identified people. Spending, sales and comparable source conditions are unknown. Campaign is an exercise label, not a connected advertising account.
State the question before calculating: what inquiry-to-visit ratio describes each set? Aggregate ratios do not establish individual journeys. Deduplication, attribution, channel, duration, prices, availability and season are unspecified. Never invent equality across those factors.
Preserve differences. A has more visits; B has more inquiries. Neither observation establishes profitability. Choosing only a favorable figure hides part of the source. Keep both counts and label unknowns instead of replacing them with zero or convenient assumptions.
Case for review
Draft D0: “B won because AI generated more sales with less investment.”
Your activity
Check support for winner, AI, sales and investment. Prepare a dictionary for visits, inquiries, ratios and missing data.
Worked feedback — compare after trying
Only visit and inquiry counts are supplied. There is no defined winner, identified AI intervention, sales or investment. Fewer visits is not lower investment: different units. Describe A/B without attributing benefit or cause.
Supporting artifact 1: F9 dictionary and unsupported-claim matrix.
Week 2 · Calculate ratios and differences
Goal: Distinguish counts, percentage points and relative change.
Inquiry/visit divides inquiries by visits. A is ten over one hundred, ten percent. B is twelve over eighty, fifteen percent. B has a higher observed ratio and two more inquiries, but twenty fewer visits. Show bases so another person can reproduce the calculation.
Fifteen percent minus ten percent is five percentage points, not five percent relative change. Taking A’s rate as the base gives five divided by ten, or fifty percent relative rate difference. Name that base; it implies neither revenue nor causation.
Inquiry change is two units, or twenty percent relative to A’s ten. Visit change is twenty fewer, or twenty percent lower relative to A’s hundred. These are different descriptive comparisons. Do not equate twenty percent inquiry growth with fifty percent rate difference.
Case for review
A card says “B improved five percent and earned fifty percent more money.”
Your activity
Show fractions, decimals and percentages. State rate difference in points and inquiry/visit differences in units. Name A as the base for any relative change.
Worked feedback — compare after trying
A10/100=0.10=10%; B12/80=0.15=15%. Difference:5 percentage points. Inquiries:+2,20% relative to10; visits:20 fewer,20% relative to100. Relative rate difference:50% against10%. None calculates money. Present original counts and ratios first.
Supporting artifact 2: calculations and complete labels.
Week 3 · Explain limits and propose minimal data
Goal: Ask useful questions without collecting real information.
Observed differences can coincide with prices, stock, season, channel or counting methods. F9 cannot separate these explanations. Conclude neither that B caused more inquiries nor that a tool was responsible. A future comparison proposal is a work hypothesis, not an executed experiment.
Define purpose and minimum information. To check comparability, propose period, aggregate channel, visit/inquiry definitions, deduplication rules, stock availability and price changes. Spending and sales would answer separate economic questions but are absent. Do not request names, phone numbers, addresses or individual histories here.
Proposed data requires authorization and coherent methods before collection. Write a local plan only. Open no accounts, buy no advertisements, export no customer lists and contact nobody. More figures alone cannot eliminate all uncertainty: measurement quality and design still matter.
Case for review
Someone proposes downloading customer contacts and launching ads immediately to explain B’s result.
Your activity
Replace this with five aggregate fields, purpose, required authorization and a remaining limit. Write two F9 observations and two interpretation limits.
Worked feedback — compare after trying
Observations: B has12 versus10 inquiries and15% versus10% observed ratio; A has100 versus80 visits. Limits: no spending/sales and unknown comparable conditions. Minimum plan: period, channel, visit definition, inquiry definition and deduplication rule; authorized review before collection. It collects nothing and proves no cause.
Supporting artifact 3: observations, limits and unexecuted aggregate plan.
Week 4 · Test conclusions and assemble M09
Goal: Deliver a reproducible, cautious comparison.
Use separate variants. V9A changes B to eight inquiries and eighty visits, yielding ten percent. V9B changes B visits to zero while retaining twelve inquiries: the ratio is not calculable and source review is needed. These are not real periods or extra observations to add to F9.
A summary should respond to changed inputs. A template that always says “B wins” fails V9A and overstates F9. Describe observations and unknowns. Keep source, calculations, corrected draft and data plan with versions. Publish no commercial recommendation from this exercise.
Combine four artifacts into one M09 activity. Approved rubric v1: facts/sources, artifact/process, data/human review and reasoning/limits; each 0–2. Zero absent/wrong, one partial, two sufficiently justified. Minimum 6/8 without zeros requires authorized human review, never checklist graduation.
Case for review
The report retains “B is more profitable” after admitting costs and sales are unknown.
Your activity
Correct it in two sentences and test F9,V9A,V9B. Attach operations, two observations, two limits, aggregate plan and evidence-based self-review.
Worked feedback — compare after trying
Summary: “F9 shows A10% and B15% inquiry/visit ratios, with10 and12 inquiries. It establishes neither profit nor causal effect because spending, sales and comparable conditions are missing.” V9A8/80=10%,same rate as A with fewer inquiries; V9B12/0 not calculable. A 6/8 without zeros may be reviewed;6/8 with a zero fails. No statements or benefits.
Single M09 submission: descriptive comparison, variant tests and unexecuted plan with self-review.
Assemble your single monthly package
Download M09 package guide (TXT)
Keep your work locally. Formal review requires actual artifacts and an authorized human reviewer. No certificate, enrollment or benefit is issued by this page. The other eleven months and final portfolio remain separate parts of the annual program.
Optional AI tutor — read consent and limits first · My personal study marks, not academic verification