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

Growth Marketing › Beginner

Capstone: A First Growth Model and Experiment Plan

The capstone brings the beginner level together. You build a growth model for an invented photo book maker: stage definitions, a dashboard read as a chain, a north star with its inputs, one mapped loop, and an experiment plan with scored ideas and complete cards. The result is a first growth plan a small team could start running next week.
Capstone projectBeginner levelModule 3: Running first growth experiments7 minUpdated

Brief

Pagecraft Photo Books is a fictional online business. Customers upload photos, arrange them in a browser editor and order a printed hardback book, usually as a gift or a keepsake after a holiday. Every book carries a small "Made with Pagecraft" line and web address on the back page.

The founders have grown the business with paid search and paid social, and with a big December peak each year. Growth has stalled outside the peak, and paid costs are rising. A friend has told them to "do growth marketing". They have asked you to write their first growth plan: how the business grows, where it leaks, what single measure the team should watch, and which experiments to run first and how.

Your role is the growth lead for a quarter. One founder will review your plan; the other will approve it. The deliverable is a written plan of a few pages, not a slide deck.

Inputs and constraints

Everything below comes from the founders' analytics, order system and a short customer survey. Treat it as your evidence and say where you would want more.

Illustrative example — not a real result

Pagecraft Photo Books: a typical non-peak month

MeasureValue
Site visits52,000
Accounts created3,900
Projects started (at least one photo uploaded)2,600
First books ordered780
Paid spend£15,600
First-book customers attributed to paid channels520
Average order value£34
Contribution per book after printing and delivery£14
Visits from the back-page web address1,100
First books ordered by back-page visitors40
Behaviour cutValue
Projects using the auto-layout button900 started; 414 ordered (46%)
Projects not using it1,700 started; 366 ordered (about 22%)
Last year's first-book customers who ordered again within 12 months22%, averaging two further books each
December ordersroughly four times a typical month
Survey of customers who abandoned a project (48 replies)Count
"Arranging the photos took too long"21
"I didn't know what it would cost until checkout"9
"I meant to come back and forgot"11

Constraints you must respect:

  • Two founders and a part-time designer; developer time is a few days a month.
  • Traffic outside December is modest, so only a small number of tests can run in a quarter.
  • No checkout changes in the six weeks before Christmas.
  • Paid budget is fixed for the quarter; it can be moved between channels but not raised.

Task

Work through the steps in order. Each produces a section of the plan.

  1. Write the model sentence. In one sentence, say how a customer arrives, what they must do to get value, why they come back and how they bring the next customer, as the growth marketing guide describes for a first plan. If part of the sentence is weak, say so.

  2. Define the stages. Using the pirate metrics framework, write each stage as an event, a population and a period. Justify the activation event from the behaviour cut, not by assumption, and choose a retention period that matches how often people buy photo books.

  3. Read the chain. Lay out the month as a stage dashboard with a rate from each stage to the next. Name the stage where a feasible improvement would move the end of the chain most, and explain why more paid spend would not fix it, with reference to CAC and first-year value.

  4. Choose the north star. Test candidates against the criteria in the north star metric and objectives framework, choose one, define it and break it into inputs with an owner each. Add at least one guardrail. Write one objective with key results for the quarter, each with a baseline.

  5. Map one loop. Use the method in the growth loops in practice guide to map either the back-page gift loop or the repeat purchase loop. Put a rate on each arrow, mark guesses as guesses and name the weakest arrow.

  6. Build and score a backlog. List at least six experiment ideas tied to the weakest stage or loop arrow. Score them with impact, confidence and ease as the experimentation loop describes, with a reason for each score, and check the top of the list against what the traffic can detect.

  7. Write the first cards. For the top two ideas, complete every field of the experiment hypothesis template, including the decision rule. Say which can run as an A/B Test and which should be shipped as a considered change and watched.

  8. Set the rhythm. Using the experimentation cadence guide, describe the weekly read-out, the monthly review and the learning log entry format the founders will keep.

  9. Summarise for the founders. In plain words: where the business leaks, what the team will watch, what it will test first, and what it will not do this quarter.

Illustrative example — not a real result

Checking your reading of the chain

Visits 52,000 → accounts 3,900: 7.5%. Accounts 3,900 → projects started 2,600: about 67%. Projects started 2,600 → first books 780: 30%, a loss of 1,820 projects in which the customer had already uploaded photos.

Paid acquisition cost: £15,600 ÷ 520 = £30 per first-book customer. First-year contribution per customer: £14 + (22% × 2 × £14) ≈ £20, below the £30 it cost to acquire.

Back-page loop: 40 first books from 1,100 visits, about 0.05 first books per first book ordered that month (40 ÷ 780), a figure to treat as a rough guess until it is read by cohort.

Your figures may differ if you make different, stated assumptions. What matters is that every step is shown and that the weakest stage is chosen on where a feasible change moves the chain, not on which rate is smallest.

Evaluation criteria

Review your own plan against the table, then swap with someone else if you can.

CriterionWhat a strong answer showsCommon gap
Model sentenceArrival, value, return and referral in one sentence, with the weak part namedA list of channels instead of a model
Stage definitionsEvent, population and period for each stage; activation justified by the auto-layout and ordering data; a retention period of months, not daysActivation set at account creation, or retention read weekly for a product bought a few times a year
Chain and weakest stageThe project-to-order step named as the leak, with the abandonment survey as supporting evidencePaid acquisition named as the problem because it is the line the founders spend on
EconomicsFirst-year value read on contribution, not revenue, and set against acquisition cost; spend held until the leak improvesLTV quoted as order value, which makes every channel look profitable
North star and inputsA count of value delivered, such as books delivered to customers each month, with inputs and owners; a guardrail on margin or refundsRevenue as the north star, or inputs nobody owns
Objective and key resultsOne objective; key results that are readings with baselinesKey results written as tasks
Loop mapEach arrow rated, guesses marked, weakest arrow named, loop treated as a subsidy rather than an engineA funnel with a referral afterthought called a loop
Backlog and scoringIdeas tied to the leak, scored with reasons, checked against traffic and the December freezeScores with no reasons, or a homepage redesign at the top
Experiment cardsEvery template field filled, one change per card, a mechanism, a guardrail and a decision rule written in advanceExpected effect copied from hope; no decision rule
CadenceA weekly read-out, a monthly review and a learning log entry format the founders can keepA plan with no rhythm, which becomes a one-off project

Submission checklist

Before you call the plan finished, confirm each item:

  • The model sentence is written and its weak part named.
  • Every stage has an event, a population and a period, and activation is justified from the data.
  • The stage dashboard shows a rate at each step, with the leak named and the arithmetic shown.
  • First-year value is on contribution and compared with acquisition cost.
  • The north star is defined, with inputs, owners and at least one guardrail.
  • One objective has key results with baselines.
  • One loop is mapped with rates, guesses marked and the weakest arrow named.
  • The backlog has at least six scored ideas with reasons.
  • Two experiment cards have every template field filled and a decision rule.
  • The cadence names the rituals and the learning log format.
  • The founders' summary fits on one page and says what will not be done this quarter.

When the plan is complete, mark the lesson as done on the beginner growth marketing level page. If a criterion exposed a gap, the quickest revision is the matching exercise: mapping meal kit metrics to pirate metrics stages, choosing a north star for a language learning app or writing an experiment card for a coworking space.

Pages this lesson applies

Templates and checklists used

Terms used on this page

Also related

Related