Domain 3 of 6

Reading Product Behaviour

What the instrumented product is saying. Funnels and where people leave, cohorts that separate a change from a trend, retention curves and whether they flatten, churn measured in customers against churn measured in revenue, the activation moment that predicts staying, how engagement is counted, the segments hiding inside an average, and why a generative feature needs measures of its own.

Concepts

~22%

Of the exam

Practice questions

Concepts in this domain

01Funnels and conversionA funnel as a sequence of steps with a count at each one, the three choices that decide what it measures, a worked conversion calculation from forty thousand visitors down to four hundred and eighty customers, and why the denominator has to be stated before any rate means anything.02Cohort analysisGrouping people by when they arrived or by what they did and following each group forward, the difference between an acquisition cohort and a behavioural cohort, a worked case where a blended figure falls while no cohort gets worse, and the three directions a cohort table is read in.03Retention curvesThe shape a cohort traces over time, what a curve that flattens says about a durable audience, what a curve that reaches zero says about the absence of one, and how far apart day N, unbounded and bracket retention put the same cohort on the same data.04Churn and revenue retentionChurn counted in customers set beside churn counted in money, the formulas for gross and net revenue retention, two worked cases where the two counts point in opposite directions, and what net revenue retention above one hundred per cent does to the growth a company has to buy.05ActivationThe early action that separates people who stay from people who leave, how a candidate action is found by comparing retained and churned cohorts, the two properties a usable definition needs, and the correlation at the centre of every activation finding.06Engagement and frequencyHow often people come back, daily over monthly active users as a ratio with a worked case where two very different products report the same 20 per cent, counting active days in a period, and the natural frequency that sets the ceiling for any product.07Segmentation in analyticsCutting a population so an average stops describing nobody, the cuts that usually pay, and a worked case of Simpson's paradox where a signup rate improves on every device and the combined figure falls from 14.8 per cent to 12.9 per cent.08Measuring an AI featureAcceptance rate and the unit it is counted in, edit and override rate, the error rate a team has to build because nothing emits an event for a wrong answer, the arithmetic of the trade between latency and quality, and why ordinary engagement measures point the wrong way for a generative feature.

Try a question from this domain

A flooring supplier's trade portal recorded 40,000 visits to the pricing page, 8,000 trials started, 2,400 invitations sent, 1,200 data sources connected and 480 accounts paid over a thirty day window. The first step loses 32,000 people, so a designer argues that rebuilding the pricing page is obviously the most valuable work available. What does the arithmetic say?

AThe designer is right, because 32,000 people is by far the largest loss in the sequence and recovering even a small share of them outweighs anything further down.
BA one tenth gain at any single step produces 528 paying accounts instead of 480, so the extra 48 arrives wherever in the sequence the improvement happens and the choice rests on which step is cheapest to move.
CThe step converting at 30 per cent is the one worth working on, because it is the weakest rate below the public page and the weakest rate has the most room in it.
DNo comparison is possible until the funnel is rebuilt with an enforced order, because the steps as listed count anybody who took them in any sequence.

20 questions on this domain.

One per page, with a worked explanation.

Start the set