Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Growth Marketing › Intermediate

Capstone: A Growth Strategy and Experiment Roadmap

The capstone of the intermediate growth level asks you to write a quarter's growth strategy for an invented podcast hosting service. It draws on unit economics by channel, cohort reading, finding the stage that limits growth, choosing a growth motion, building a loop and running an experimentation programme, and is judged against explicit criteria.
Capstone projectIntermediate levelModule 3: Loops, models and experimentation at scale6 minUpdated

Brief

Driftwood Audio is an invented podcast hosting service. Podcasters start a free trial, upload episodes, and Driftwood distributes them to listening apps and hosts an embeddable player. Every player shows a small "Start your own podcast" badge. Customers pay monthly on a Starter or Pro plan. The founder has asked you, the new growth lead, for a strategy for next quarter because new paying podcasters have stopped rising even though spend has.

Four channels bring customers: paid search, sponsored ad reads on other podcasts, content and search optimisation, and the player badge. The founder's instinct is to buy more ad reads, since they "feel on brand". Sales, which is one account manager, wants an enterprise tier after several podcast networks asked for multi-show accounts.

Inputs and constraints

Illustrative example — not a real result

Unit economics by channel, per quarter, twelve-month window

ChannelCost incl. people and incentivesNew paying podcastersCACTwelve-month contribution per podcasterRetained at month sixPayback
Paid search£36,000600£60£11861%Month 5
Podcast ad reads£24,000300£80£9649%Month 9
Content and search optimisation£15,000500£30£13266%Month 2
Player badge£3,000250£12£14068%Month 0
All channels£78,0001,650£47

The last £10,000 of quarterly paid search spend brought 110 podcasters, £91 each. Ad read customers are attributed from promotion codes and a sign-up question; the channel has never been tested for incrementality. Average contribution is £14 per paying month.

Illustrative example — not a real result

Funnel, cohorts and research

StageMeasureMonthly average
AcquisitionTrial starts2,400
ActivationTrials publishing a first episode within the 14-day trial912 (38%)
RevenueTrials converting to paid550: 55% of activated trials, about 3% of the rest
RetentionRetained at month six, latest complete cohort58%, against 64% a year earlier
ReferralPaying podcasters from the badgeAbout 83

Over the same year the ad reads' share of new podcasters rose from 10% to 25%. Podcasters who publish four or more episodes in their first two paying months are 76% retained at month six; the rest, 34%.

Research notes: podcasters moving an existing show from another host raise more trial support tickets than any other group, most about redirecting their feed. Listeners who click the badge land on the generic homepage. Three podcast networks asked for multi-show accounts with one invoice and shared admin.

Constraints: the quarter is 13 weeks; the Starter price rises in week 8, so no tests on pricing or checkout pages in weeks 7 to 10; no test may run longer than six weeks; finance sets a six-month payback ceiling per channel and a fixed budget. Team: you, an engineer half time, a lifecycle marketer, a part-time designer and an analyst one day a week.

Task

Write a strategy document with the parts below, in this order. The growth marketing hub sets the frame: the whole journey as scope, experiments as method and unit economics as the scoreboard.

  1. Channel economics. For each channel, state LTV to CAC ratio, payback against the ceiling and a decision: scale, hold, cap or test. Follow the definitions in the guide to LTV and CAC, use marginal cost where spend would rise, and say why the blended figure cannot settle any channel. The break-even calculator checks individual rows.
  2. Cohort reading. Explain the fall in month-six retention, separating mix from product, and define the retention KPI as a cell of the cohort table.
  3. The limiting stage. Lay out the stages with the AARRR framework, name the stage that most limits growth and show the evidence, including which group of trial users it hits hardest.
  4. Growth motion. Decide how to answer the networks, using the comparison of product-led and sales-led growth, and say what you will not build this quarter.
  5. A loop to build. Map the badge loop as it is and as you would change it, with a number on every arrow, a cycle time and the measure of its return, as the growth loops framework requires.
  6. The experiment roadmap. Write each idea as a card on the experiment hypothesis template, check feasibility against trial volume and the test ceiling, score and sequence what survives, and lay it across the quarter around the price change. Include an Incrementality test for the ad reads and a holdout, read at fixed cohort ages, for any retention change.
  7. Cadence and decisions. Assign the roles and rhythm of the experimentation programme playbook to Driftwood's team, name who alone declares results, and write a decision rule for every test before it starts.
  8. The founder's page. One page: what you will scale, cap and stop, the stage you will fix, the loop you will build and how the founder will know in a quarter whether it worked.
Illustrative example — not a real result

One row of a strong channel table, for comparison

ChannelLTV ÷ CACPayback against ceilingMarginal viewDecision
Paid search£118 ÷ £60 = 2.0Month 5, inside six£91 at the margin: £118 ÷ £91 = 1.3, and slower paybackHold spend at the level before the last £10,000; reinvest that sum where payback is faster

The row shows its arithmetic, uses the marginal figure for the spending decision and ties the decision to the ceiling.

Evaluation criteria

Assess your strategy against each row, or ask a peer to. A strong strategy meets every row; most first attempts miss two or three.

CriterionWhat a strong answer showsCommon gap
Honest channel economicsRatios and payback per channel on contribution, marginal cost used for increases, the blend set aside for budget decisionsRanking channels by the blended figure, or by cost per sign-up
The ad readsCapped at the ceiling until an incrementality test, such as pausing reads for a planned period, shows what they addScaling them on the founder's instinct, or cutting them with no test
Cohort readingThe retention fall traced to the shift towards ad read customers before any product cause is claimed; KPI defined as a cohort cellBlaming the product for a mix change
The limiting stageActivation chosen with evidence, the switching podcasters singled out, and an early-publishing leading metric validated before usePicking acquisition because it is the stage marketing owns
Growth motionNetworks served with a self-serve multi-show plan and usage-led outreach by the account manager, with no sales team built this quarterAn enterprise sales motion bolted on for a handful of requests
The loopA dated badge loop map with numbers, a landing page for badge clicks, and its return compared with paid channelsA loop drawn as intended, with no return per cycle
Roadmap feasibilityTests sized to trial volume, inside the ceiling, sequenced around the price change and the team's capacityMore concurrent tests than the engineer can build
Decision disciplineDecision rules written first, one person declaring results, holdouts with an end date, a learning logResults argued after the fact
Fit to the businessA page the founder can read once and act onA strategy that needs a larger team

Submission checklist

0/11

When you have finished, compare your roadmap with the phases in the experimentation programme playbook, then return to the intermediate growth level to mark the level complete. If your strategy changed a channel decision because of what its cohorts showed, the level has done its job.

Pages this lesson applies

Terms used on this page

Also related

Related