Learn with Phoebe Β· Data & Analytics

Learn Marketing Analytics with Phoebe

Data analysts who serve marketing, brand and BD get two kinds of question wrong. The first kind is about the numbers: a follower count that decides nothing, an open rate inflated since 2021, a platform ROAS the platform grades itself. The second kind is about the job: "why won't data do our market research?" These ten sessions teach both. Six Angles sessions read every channel in the same four beats, what it is, what marketers watch, the vanity trap, and where data adds value, then cohorts, segment value and lift. Four Partnership sessions stand in marketing's shoes, tag every ask by the knowledge it needs, and draw the line the whole course argues for: market research and penetration strategy belong to marketing and BD, and data is consulted with a named deliverable and a date. The running company, Kettle & Co, is constructed.

10sessions, two tracks
11channels in four beats
24asks on the triage desk
2failure modes, measured
45min per session

Angles track 🧭 · six sessions, no code needed

The angles from which a data team looks at marketing. How marketers think and what they really ask, eleven channels read in the same four beats, acquisition cohorts by channel, segment value and payback, and lift tests read honestly, including the result that says nothing worked.

🧭Angles session 1 · start here

How marketers think, and the question behind the ask

The funnel as marketers draw it, brand against performance and the contested 60:40, and the habit of turning a vague ask into a decision question with an owner.

β–Ά Start here
πŸ”ŽAngles session 2

Owned and earned channels: search, content, social, PR

SEO, content, organic social and creators, and PR, each in four beats: what it is, what marketers watch, the vanity trap, and where data adds value.

four beats per channel
πŸ’ΈAngles session 3

Paid and lifecycle channels: auctions, inboxes and partners

Paid search and paid social, email and CRM, affiliates and influencers, events. Platform ROAS is self-reported; opens broke in 2021; a scorecard where every metric names its decision.

metric beside its decision
πŸͺ£Angles session 4

Acquisition cohorts by channel

Group customers by the month and channel that brought them, and read where each cohort leaks. The channel that wins the first order is often not the one that wins the second.

cohort by channel
🎯Angles session 5

Segment and value: who deserves the next dollar

CAC, payback and margin by segment, CLV as an input, and the difference between the cheapest customer to win and the best one to keep.

CAC and payback
πŸ§ͺAngles session 6

Campaign lift and brand health

Test against control, holdouts, and reading a lift test honestly, including the no-effect result; brand tracking waves and their intervals; and where attribution takes over.

the honest no-effect read

Partnership track 🀝 · four sessions, the working relationship

How to work with marketing, brand and BD: their scoreboards and calendar, whose job each ask is, the front door and the service menu, and how to say no with a deliverable and a date. The bench is a request triage desk computed in your browser.

foundations core stretch bench night

Your passport 🎫

Clear all three quiz questions in a session and it gets stamped. Ten stamps means you can read any marketing channel's numbers for what they decide, tag any ask by whose job it is, and say no with a deliverable and a date. Stamps live in this browser only.

What this is not, so you land in the right place 🧭

If your question isGo here instead
I am a marketer: how do I position, plan and budget?Marketing β†—
Which touchpoint gets the credit, and how do MMM and geo-lift work?Marketing Attribution β†—
How do I use AI to write the SEO, content, ads and emails?AI + Marketing β†—
How do I compute CLV or build a churn model?Customer Retention β†—
How do I read the store funnel and the checkout?E-commerce Metrics β†— Β· User Journey Analytics β†—
How do I run the discovery interview itself?Asking the Right Questions β†—
I am the analyst marketing, brand and BD come to, and I want to read their numbers well and know whose job each ask is.you are in the right place

Marketing analytics as a working relationship. Kettle & Co is constructed; platform facts are read from the platforms' own help pages and dated; contested claims, such as the 60:40 brand split, are taught as contested.

Choose your path πŸ—ΊοΈ

Three doors into the same conviction: read the number for the decision it informs, and know whose decision it is.

🧭 I read marketing's numbers A1β†’A2β†’A3β†’A4β†’A5β†’A6 the channels, cohorts, value and lift
🀝 I am drowning in asks P1β†’P2β†’P3β†’P4 their shoes, the map, the front door, the scripts
πŸŽ›οΈ I lead the data team A1β†’P1β†’P2β†’P3β†’P4 the question behind the ask, then the whole partnership

What is honest here 🧾

The running company is constructed and says so on every page. Platform facts are read from the platforms' own pages and dated, because they change. Where a claim is contested, the page teaches both sides.

ClaimStatus
Kettle & Co, its people, its 24 asks, its hours and its channel dataConstructed, and labelled each time
Every number on the triage deskComputed in your browser from the constructed stream; a Python reference prints the same integers; the queue rule and hours are modelled
How search, attribution windows, conversion modelling and Mail Privacy Protection workRead at the platforms' own pages, dated, and flagged to re-check before you rely on them
The 60:40 brand and activation splitContested, and taught as contested: the original claim, its category caveats and its critics
Market research and penetration strategy belong to marketing and BDA stance, argued with its reason, next to the alternative model that some good teams run

The knowledge map 🧠

Both tracks at a glance - hover a session to spotlight its concepts, click any node to jump in.