My progress
Menu

Digital Marketing · Focus area 3: Advertising and growth engineering

DM9: Growth engineering: referrals, sharing and lifecycle messages

Goal: design product features that help a business grow by itself: growth loops, referral programmes that resist abuse, share links that look good and measure well, lifecycle messages people want, and feature flags for testing growth ideas, without dark patterns.

You'll need: DM4, DM5, DM7, Module 19 (Browser Interactions), Module 44 (Authentication), Module 88 (Workflow Automation), Module 97 (Product Development), and the MittiMart practice kit (for the capstone design notes).

Where this is used: payment apps grew with cashback for inviting friends, food apps with referral credits, and shops with WhatsApp share buttons and abandoned-cart reminders. Growth engineers build these features, measure them with experiments, and stop them being gamed.

How it differs: advertising buys attention from outside. Growth engineering changes the product so each customer brings the next one, or comes back by themselves. It's product work with a marketer's measures, and it needs an engineer's care for abuse and privacy.

Time: reading about 2½ hours · practice (the Try it tasks) about 2 hours · assignment about 3 hours.

Lessons

  1. 1Growth loops, activation and retentionReading 15–25 minutes, plus practice✓
  2. 2Referral programmes that resist abuseReading 15–25 minutes, plus practice✓
  3. 3Share links, previews and deep linksReading 15–25 minutes, plus practice✓
  4. 4Lifecycle messages people wantReading 15–25 minutes, plus practice✓
  5. 5Feature flags, growth experiments and ethicsReading 15–25 minutes, plus practice✓

🛠️Assignment: Design MittiMart's growth features

  1. Draw one growth loop for MittiMart and define its activation event.
  2. Write the rules for a two-sided referral programme, with three abuse controls and the cost per referred customer (illustrative).
  3. Design a share button for product pages: Open Graph tags, the share link with campaign tags, and the share event.
  4. Design an abandoned-cart message with consent, a frequency cap, a re-check before sending and a holdout measure.
  5. Review one proposed growth tactic against the 2023 dark patterns guidelines and write a verdict.

Self-check before you finish:

✅Module check

Explain in a few sentences, in your own words:

How do you build features that help a product grow without tricking anyone?

Show a good answer

A good answer mentions: growth loops that feed customers back in, a measurable activation event, and retention before acquisition; referral rewards after delivered, kept orders with limits against self-referral and fraud, and cost compared with other channels; Open Graph previews, tagged share links and a share event without personal data; lifecycle messages with separate marketing consent, frequency caps, re-checks and easy opt-out, measured with holdouts; feature flags with owners, end dates and kill switches; experiments with guardrails; and rejecting dark patterns such as false urgency, basket sneaking and confirm shaming before they're built.

📚Official references

Official documentation and guidance for this module's topics, chosen on 2 November 2026. Tools, laws and standards change, so check the current version before you rely on a detail. If a link has moved, search the same site.

📝Module quiz

10 questions · 10 marks each · 100 marks in all · pass mark 70% · retry as often as you like ·

1. Which is a growth loop rather than a funnel?

Why each option is right or wrong

  • “Visitors see an ad, view a product, add it to the cart and buy it” (Incorrect): A one-way path is a funnel.
  • “The marketing team increases its ad budget every month” (Incorrect): Spending more is acquisition, not a loop.
  • “A shop sends the same newsletter to all customers each week” (Incorrect): A broadcast doesn't feed new customers back in.
  • “Buyers post photo reviews, which bring search visitors, who buy and post reviews” is correct.

    The output (reviews) brings new customers who produce more of it.

    • Measure: visits to reviewed products.
    • Test: with a review prompt experiment.

📖 Revise: Lesson 1: Growth loops, activation and retention

Question ID sp-digital-marketing-09-q1 · Report a mistake in this question
2. A start-up's referred customers rarely order twice because deliveries are late. What should come first?

Why each option is right or wrong

  • “A bigger referral reward, so more friends try the service” (Incorrect): More new customers would leave for the same reason.
  • “More ads, so the loss of customers is made up by volume” (Incorrect): That fills a leaky bucket at greater cost.
  • “Remove the option for customers to leave reviews” (Incorrect): Hiding feedback doesn't fix deliveries.
  • “Fix delivery so customers return, then grow referrals” is correct.

    Retention comes before acquisition; a product worth recommending makes referrals work.

    • Check: cohorts' second orders.
    • Then: restart rewards.

📖 Revise: Lesson 1: Growth loops, activation and retention

Question ID sp-digital-marketing-09-q2 · Report a mistake in this question
3. When should a referral reward be paid?

Why each option is right or wrong

  • “As soon as the friend signs up on the website or app with the referral code” (Incorrect): Sign-up rewards attract fake accounts.
  • “When the friend first opens the shop's app on their phone” (Incorrect): Opening an app proves little.
  • “When the referrer shares the code on social media” (Incorrect): Sharing doesn't bring a customer by itself.
  • “After the friend's first order is delivered and the return period ends” is correct.

    Delayed rewards reduce fake and cancelled orders.

    • Plus: limits per device and address.
    • Measure: extra customers with a holdout.

📖 Revise: Lesson 2: Referral programmes that resist abuse

Question ID sp-digital-marketing-09-q3 · Report a mistake in this question
4. One address has created 40 accounts through referral links in a week. Which control would help most?

Why each option is right or wrong

  • “A larger reward, so honest customers are encouraged to refer more” (Incorrect): That rewards the abuse too.
  • “Removing the friend's reward but keeping the referrer's” (Incorrect): The abuser is both sides; it doesn't stop them.
  • “Asking each new account to confirm they're a real person by email” (Incorrect): Creating email addresses is easy.
  • “One reward per delivery address, device and payment method, with monthly limits” is correct.

    Self-referral uses many accounts but few addresses and payment methods.

    • Also: delay rewards.
    • Tell: customers the rules.

📖 Revise: Lesson 2: Referral programmes that resist abuse

Question ID sp-digital-marketing-09-q4 · Report a mistake in this question
5. A product link shared on WhatsApp shows only the site's logo, the same for every product. What's the fix?

Why each option is right or wrong

  • “Shorten the link with a link shortener before anyone shares it in a group” (Incorrect): Shortening doesn't change the preview.
  • “Ask customers to add a photo when they share the link” (Incorrect): The page should provide the preview.
  • “Add more keywords to the product page's meta description” (Incorrect): Previews use Open Graph tags; keywords don't help.
  • “Set each product page's own og:image, og:title and description” is correct.

    Open Graph tags tell apps what to show in the preview.

    • Image: the product's own photo.
    • Plus: og:url as the canonical.

📖 Revise: Lesson 3: Share links, previews and deep links

Question ID sp-digital-marketing-09-q5 · Report a mistake in this question
6. A share link is built as ?ref=asha.rao@example.com. What's wrong?

Why each option is right or wrong

  • “Nothing, as it helps the shop thank the person who shared the link with friends” (Incorrect): It exposes the sharer's email to everyone who sees the link.
  • “The link is too long and should use a shortener” (Incorrect): Length isn't the problem; the email is.
  • “Referral links can't include any parameters at all” (Incorrect): A referral code is fine.
  • “It puts personal data in the link; use a referral code instead” is correct.

    Links are forwarded, logged and stored; contact details shouldn't be in them.

    • Code: random and not guessable.
    • Measure: campaign tags.

📖 Revise: Lesson 3: Share links, previews and deep links

Question ID sp-digital-marketing-09-q6 · Report a mistake in this question
7. An abandoned-cart email "influenced" 240 orders from 2,000 emailed carts. A 10% holdout bought at 9%, against 12% for the emailed group.

Why each option is right or wrong

  • “240, all the orders placed after the email was opened” (Incorrect): Many of those buyers would have bought anyway.
  • “0, since the held-out group also bought” (Incorrect): The emailed group bought more.
  • “180, the buyers who would have bought anyway” (Incorrect): Those are the orders the email didn't add.
  • “About 60, as 3 extra points on 2,000 carts” is correct.

    The holdout shows what would have happened without the email.

    • Check: unsubscribes and complaints.
    • Note: illustrative figures.

📖 Revise: Lesson 4: Lifecycle messages people want

Question ID sp-digital-marketing-09-q7 · Report a mistake in this question
8. A shop sends abandoned-cart reminders by email, SMS and WhatsApp within an hour, with a discount. What's the likely long-term effect?

Why each option is right or wrong

  • “More orders every week, with no downside for the shop at all” (Incorrect): Over-messaging has real costs.
  • “Customers stop using the website and switch to the app instead” (Incorrect): That isn't the main effect.
  • “Search engines rank the shop lower because of the messages” (Incorrect): Messages don't affect search ranking.
  • “Complaints rise, and customers learn to abandon carts for discounts” is correct.

    Too many messages annoy people, and automatic discounts train them to wait.

    • Fix: one channel, a delay, a cap.
    • Measure: with a holdout.

📖 Revise: Lesson 4: Lifecycle messages people want

Question ID sp-digital-marketing-09-q8 · Report a mistake in this question
9. What's the main purpose of a feature flag's kill switch?

Why each option is right or wrong

  • “To show a feature only to staff members for testing before each release” (Incorrect): That's a targeting use, not a kill switch.
  • “To speed up the website by removing old code” (Incorrect): Removing old flags helps code, but that's not a kill switch.
  • “To decide which variant wins an experiment” (Incorrect): The analysis decides that.
  • “To turn a feature off at once if something goes wrong” is correct.

    A kill switch stops harm without waiting for a release.

    • Also: an owner and an end date.
    • Use: with guardrail alerts.

📖 Revise: Lesson 5: Feature flags, growth experiments and ethics

Question ID sp-digital-marketing-09-q9 · Report a mistake in this question
10. A team proposes testing an "Only 2 left!" badge on every product, though stock is plentiful. What should happen?

Why each option is right or wrong

  • “Test it for two weeks and keep it if conversions rise” (Incorrect): A test doesn't make a false claim acceptable.
  • “Show it only to new visitors, who won't notice that the badge doesn't change from day to day” (Incorrect): Targeting doesn't make it honest.
  • “Use it only during festivals, when urgency is expected anyway” (Incorrect): It's still false.
  • “Reject it: it's false urgency, a dark pattern named in the 2023 guidelines” is correct.

    Misleading tactics shouldn't be built or tested.

    • Instead: real delivery dates or real stock.
    • Check: long-term measures.

📖 Revise: Lesson 5: Feature flags, growth experiments and ethics

Question ID sp-digital-marketing-09-q10 · Report a mistake in this question

ProLearners Assist