4. Growth Experiments

_Growth hacking here is not a bag of tricks — it's a disciplined loop around one metric._

4.1 The model

   ACQUISITION          ACTIVATION            HABIT                REVENUE
   see the ad    →   sign up + verify   →  FIRST UPLOAD   →   repeat uploads   →   Premium
                                            ▲
                                    the whole game is here
                                            
   REFERRAL:  satisfied uploaders invite others → back to the top of the funnel

Stage	Metric	Where it's measured	Current state
Acquisition	Reach, clicks, cost per click	Platform ads + UTMs	Campaign start
Signup	Verified accounts	Admin → Users	Live
Activation	Verified → first receipt upload	Admin → OCR Quality (receipts processed)	Primary KPI
Habit	2nd, 4th, 8th upload; streak length	Admin → Users, points data	Live
Referral	Signups per referrer; referral acct → first upload	Referral data	Live
Revenue	Active Premium, conversion rate, MRR	Admin → Overview	Live

Everything the campaign does is judged on activation and habit. A viral post that brings 5,000 signups who never scan a slip is a failure.

4.2 Built-in growth levers (already in the product — use them)

Lever	What it does	How marketing uses it
Automated re-engagement email	Emails verified users who haven't uploaded after 2 days	The single highest-leverage asset. Campaign traffic lands, then this catches stragglers automatically
Upload-credit codes	Admin-issued coupon adding extra uploads on top of the free allowance	Launch giveaways, creator codes, apology/bonus gestures, competition prizes
Referral system	Every user has a referral code; referrer earns +25 points when the referral uploads their first receipt	Campaign layer on top (see 4.4)
Points & streaks	+50 first upload, +10 per upload, streak bonuses at 2/4/8 weeks	Content hooks ("8 weeks = 100 points") and retention prompts
Free monthly limit (10 uploads)	Creates a natural upgrade moment	Where Premium messaging belongs — not in cold ads
Leaderboard	Social proof of habit	Monthly "top savers" content (with permission)

4.3 Experiment backlog

_Effort: S ≤ 1 day, M ≈ 3 days, L ≥ 1 week. Run no more than three at once._

Acquisition

#	Experiment	Hypothesis	Primary metric	Effort
A1	Creator seeding — 10 local creators, each with a tracked link + credit code for their audience	Local faces beat brand ads for a trust product	Signups → first upload per creator	M
A2	"Guess the price" series on TikTok (weekly, real basket, real answer)	Native engagement format drives profile visits and cheap reach	Profile visits, link clicks	S
A3	Store-point distribution — QR flyer at till points with partner stores	Highest-intent moment is standing in the shop	Scans → signups	M
A4	Group seeding — share Basket Watch into Facebook/WhatsApp community groups, value-first, no hard sell	These groups are where the Household Manager already talks about prices	Referral traffic, signups	S
A5	Hero message test on landing: outcome ("See what you overpaid") vs mechanism ("Scan your grocery slip")	Outcome-led beats mechanism-led for cold traffic	Landing → signup rate	S

Activation (highest priority)

#	Experiment	Hypothesis	Primary metric	Effort
B1	Second nudge email at day 5 for anyone who ignored the day-2 nudge, with a worked example and a "reply with your shop name, we'll show you its prices" hook	The first email proves intent; a second angle converts a further slice	Verified → first upload, day 5 cohort	S
B2	SMS/WhatsApp nudge for users who verified by phone and haven't uploaded (consent permitting)	SMS is read far more than email in Botswana	Verified → first upload, SMS cohort vs email-only	M
B3	"Scan one slip on us" — after signup, show a single-screen promise: "your first slip takes 10 seconds" with a one-tap camera prompt	Reduces the blank-state hesitation that kills activation	Signup → first upload within 24h	S
B4	Welcome credit — every new verified user gets one bonus upload credit	Framing the free tier as "10 + 1 bonus" increases first-attempt rate	Signup → first upload	S
B5	Onboarding checklist in-app: verify → scan → see insight → share	Makes the next action obvious	Completion of each step	M

Habit & retention

#	Experiment	Hypothesis	Primary metric	Effort
C1	Streak reminder email at day 6 of the week: "your streak is at risk"	Streaks are the strongest retention mechanic we have	Week-over-week upload rate	S
C2	Two-slip challenge — "scan twice this week, earn double points"	Sets the repeat behaviour that becomes a habit	Second upload within 7 days	S
C3	Monthly savings recap email: what you overpaid vs what you'd have saved	Makes saved value visible and justifies Premium	30-day retention, Premium conversion	M

Revenue

#	Experiment	Hypothesis	Primary metric	Effort
D1	Limit-moment upgrade — when the free 10 uploads are hit, show a value-framed screen with their own measured savings so far	Conversion belongs at felt value, not in cold ads	Free → Premium rate at limit	S
D2	🔴 Payment friction audit — Premium is currently activated manually after an EFT/Orange Money transfer that the user initiates. Measure how many users who intend to upgrade actually complete	Every manual step leaks conversions; this is likely our biggest revenue leak	Intents → activations	M
D3	Annual-first pricing test — lead with P249/year (save 29%) rather than monthly	Higher LTV per activation, fewer renewals to chase	Plan mix, revenue per activation	S
D4	Referral give-and-get — also give the new user a bonus (e.g. +5 uploads) on their first upload	Two-sided rewards roughly double referral participation	Referral signups → first upload	M

4.4 Referral campaign layer

In-product today: referrer earns +25 points when the referred user uploads their first receipt.

Campaign layer on top (test one at a time):

Two-sided bonus (D4) — give both sides something concrete. Points alone are

abstract; uploads are tangible.

Referral leaderboard — top referrers this month win a free Premium quarter.

Fuels the Points Collector segment and generates content.

Creator referral codes — each creator gets a code and a tracked link so we know

exactly which creator drives activations, not just clicks.

Family/group framing — referrals spread through family and church groups far

better than through individuals. Write for "your cousin", "your group".

Critical measurement: referral success = referral uploads a receipt, not referral signs up. The bonus already fires on the right event — keep the reporting aligned.

4.5 Credit-code giveaways

Use the admin Credits tool to issue codes granting extra uploads:

Use	Mechanic
Launch week	"First 100 signups get +10 uploads" — announced socially, code auto-emailed
Creator codes	Each creator's audience gets +10 uploads through a shared code
Competition prize	Weekly "best saving found" wins credits (and content rights)
Win-back	Users dormant 30 days get +5 uploads with a "come back" email
Apology	If a user reports a bad receipt, a credit is a cheap, sincere gesture

4.6 Weekly growth ritual

Every Monday, one page, three questions:

What did activation do this week? (verified → first upload, by source if data allows)
Which experiment moved its primary metric, and by how much?
What do we kill, and what do we double?

Then commit the next week's three experiments. Keep a running log so the agency and Breadwinner share one version of the truth.

4.7 First 90 days, prioritised

Weeks	Focus	Why this order
1–2	Foundation: tracking, UTMs, 8 creatives, 3 blog posts, creator shortlist	Nothing can be measured without this
3–4	B1, B3, A1 — activation emails, first-slip promise, creator seeding	Activation is the bottleneck; creators test the message cheaply
5–8	A2, A4, B4, C1 — content cadence, group seeding, welcome credit, streak reminders	Build habit while acquisition scales
9–12	D1, D2, D4, C3 — limit-moment upsell, payment friction fix, referral give-and-get, savings recap	Only optimise revenue once activation and habit are working

Do not start paid acquisition at scale before B3 and the day-2 email are proven.
Pouring traffic into a leaky activation step just buys wasted signups.