Skip to content

exploration-006.md

Latest commit

 

History

History
136 lines (90 loc) · 7.07 KB

exploration-006.md

File metadata and controls

136 lines (90 loc) · 7.07 KB

Exploration: Issue-006 — Ozi Reorder Experiment

Agent: Research Agent Date: 2026-03-20 Issue: issue-006

Problem Analysis

Is the problem real?

Yes. The problem is real, specific, and well-scoped.

The issue correctly identifies that dark-store churn in the baby essentials category is not driven by delivery failure — it is driven by absence of re-engagement. The user experience is structured for discovery-first browsing, which is mismatched to the mental model of a parent buying consumables on a predictable cycle.

Pain signals:

  • Baby essentials (diapers, formula, wipes) are non-substitutable, non-deferrable, and cycle-predictable. A size-3 Pampers purchase today is a near-certain size-3 or size-4 purchase in 18–22 days.
  • A parent running out of diapers at 10pm is not making a considered purchasing decision — they are executing an emergency. The competitor who sends a timely reminder owns the considered decision window.
  • The current Ozi experience requires the user to initiate: open app, search, add to cart, checkout. For a product category that runs on autopilot in the user's mind, this is 4 steps too many.

Frequency: High. Parents of 0–3 year olds reorder baby essentials every 2–4 weeks.

Existing workarounds: Competitors. Blinkit and Amazon Subscribe & Save already intercept this need. The current Ozi workaround is the user leaving the app permanently.

Assessment: Problem validity — Confirmed.

Market Scan

Competitor Approach Strength Weakness
Blinkit App notifications + reorder shortcut on order history High frequency reminders, tight delivery SLA Generic nudges, not timed to consumption cycle
Amazon Subscribe & Save Subscription-based auto-delivery Zero friction once set up Requires upfront commitment; over-serves casual buyers
Zepto Repeat order button in order history Simple UX Passive — waits for user to open app
Swiggy Instamart Order history + occasional push nudge Familiar UX No prediction logic; nudges are date-based, not cycle-based
Firstcry (baby-specific) Email-based replenishment reminders Category-aware Slow delivery SLA; not a dark-store model

Unserved gap: None of the quick-commerce players use consumption-cycle-aware timing for push notifications — they push based on elapsed time since last app open or order recency, not on estimated product depletion. A notification timed to Day 18–20 post-delivery for diapers is qualitatively different from a promotional nudge. Ozi is small enough to instrument this precisely.

User Pain Level

Classification: Critical problem for Emergency Parent and Routine Replenishment Parent.

  • Emergency Parent: Pain is acute and time-sensitive. A Day 18–20 reminder is the only intervention window before the emergency happens. Missing it = guaranteed competitor order.
  • Routine Replenishment Parent: Pain is moderate but cumulative. Behavioral lock-in risk grows with each missed reminder cycle.
  • First-Time Parent: Moderate-to-high. The pre-filled cart is a trust signal ("Ozi knows what I need"), not just a convenience feature.

Opportunity Assessment

Does solving this create meaningful value?

Yes, on three dimensions:

  1. Revenue retention: A nudge that lifts 21-day repeat purchase rate by even 10–15% across active users is a material LTV improvement at Ozi's current stage.
  2. Behavioral lock-in: A parent who receives two or three well-timed reminders begins to associate Ozi with "handles itself." That is a habit. Habits are hard for competitors to displace.
  3. No infrastructure dependency: This experiment does not require ML, personalization, subscription architecture, or any new data models beyond order_delivered event + SKU category tag.

Market size: Delhi-NCR dark-store parent cohort. Focused experiment — right size to generate learnable signal without over-building.

User willingness to adopt: Very high for Routine Replenishment Parents.

Distribution difficulty: None. Retention feature. Users already have Ozi installed.

Proposed MVP Experiment

Core feature (minimum to test the hypothesis):

  1. A trigger that fires a push notification 18–20 days after order_delivered for diaper/essential SKU categories.
  2. A deep link from the notification that opens a pre-filled cart with the user's previous order items and quantities.
  3. A control group (50% of eligible users) who receive no reminder.
  4. 7 instrumented events to measure the funnel end-to-end.

Intentionally excluded from V1:

  • ML-based timing (use fixed Day 18–20 rule)
  • Personalization by sub-segment
  • Quantity adjustment in reminder flow
  • Subscription opt-in
  • A/B testing of notification copy
  • Multi-channel (push only)

What this experiment must answer:

  • Does a timely reminder materially increase 21-day repeat purchase rate vs. control?
  • Do users who click the notification complete checkout at a meaningful rate?
  • Is Day 18–20 the right trigger window?

Minimum 7 instrumented events:

  1. reminder_triggered — notification sent (user_id, order_id, sku_category, trigger_day)
  2. reminder_delivered — push received
  3. reminder_opened — notification tapped
  4. cart_prefilled — deep link resolved, cart loaded
  5. checkout_started
  6. order_placed — conversion
  7. control_order_placed — organic conversion in control group

Risks

Technical Risk: Medium

  • Push notification infrastructure may not support event-triggered firing based on order_delivered + N-day delay without backend work.
  • Pre-filled cart requires either cart persistence API or URL scheme encoding order items. May not exist yet.
  • Mitigation: Scope discovery call with Ozi engineering to confirm push infra, SKU category tagging, and pre-filled cart deep link feasibility before committing to sprint scope.

Market Risk: Low

  • Parents are high-intent repeat buyers by necessity. A well-timed reminder for a genuinely needed product will feel useful, not intrusive.
  • Mitigation: Use Day 18–20 as conservative starting point. Instrument timing data to refine.

Distribution Risk: Low

  • Retention feature. No new acquisition cost.
  • Caveat: If push notification opt-in rates are low, reachable cohort shrinks. Validate opt-in rate before setting sample size targets.

Execution Risk: Medium

  • Clean A/B control group requires feature flag or experiment framework. If none exists, use user_id modulo 2 for cohort split.

Final Recommendation

Build.

The problem is real, the pain level is critical for the primary cohort, the competitive gap is real and time-limited, and the intervention is technically simple relative to its potential impact.

Before writing the plan, resolve two factual questions:

  1. Does Ozi's stack support event-triggered push notifications (order_delivered + N-day delay)? Or is custom cron + notification API required?
  2. Is there an existing deep link scheme or API endpoint for cart pre-population from order history?

These two answers determine whether the MVP is a 1-sprint or 2-sprint build.