Amplitude for Beauty Box Subscriptions

Beauty box subscriptions run on one metric that matters above all others: lifetime value per subscriber. Amplitude gives you the behavioral data infrastructure to protect and grow that number — but only if you instrument it correctly for the subscription model from the start.

This guide covers exactly how to do that.

---

Why Amplitude Fits the Beauty Box Model

Most analytics tools are built for transactional ecommerce. Amplitude is built for product behavior — which maps directly to how subscription retention actually works.

Your subscribers aren't making a one-time purchase decision. They're making a recurring decision, every billing cycle, to stay or leave. Amplitude's event-based architecture lets you track every behavioral signal between those billing moments — the box openings, the product reviews, the referrals, the pauses — and connect those signals to who churns and who doesn't.

That's the foundation of lifecycle optimization in this industry.

---

Setting Up Your Event Taxonomy

The most common mistake beauty brands make in Amplitude is importing too many events without a naming schema. You end up with a bloated, unstructured event library that nobody trusts.

Use a [Subject] [Verb] naming convention for all events. This keeps your taxonomy readable and filterable.

Core Events to Track

These are the events that directly map to subscriber lifecycle stages:

Acquisition & Onboarding

• `Subscriber Signed Up` — with properties: plan type, acquisition channel, promo code used
• `Quiz Completed` — beauty profile quiz completion, with properties: skin type, shade range, product preferences
• `First Box Shipped` — timestamp this as the true lifecycle start
• `Unboxing Video Viewed` — if you host this content in-app or on a portal
• `Product Review Submitted` — critical for curation feedback loops

Engagement Between Boxes

• `Portal Logged In` — frequency here is a leading churn indicator
• `Customization Made` — subscriber swapped a product before shipment cutoff
• `Wishlist Item Added` — intent signal for future curation
• `Loyalty Points Redeemed` — engagement with your retention mechanics
• `Referral Link Shared`

Retention Risk Events

• `Cancellation Flow Started` — track which step they entered from
• `Pause Requested` — with properties: pause duration, reason selected
• `Payment Failed` — first failure vs. repeated failures behave differently
• `Skip Month Selected`

Reactivation

• `Win-Back Email Clicked`
• `Resubscription Completed` — distinguish this from first-time subscribers in all your analysis

Event Properties That Matter

Every event should carry subscriber tenure as a property — calculated in days or billing cycles from their first box. Without this, you can't distinguish a new subscriber's churn from a 12-month veteran's churn, and those are entirely different problems with different solutions.

Other essential user properties: current plan tier, quiz profile data, cumulative boxes received, average product rating given, and referral status.

---

Segments to Build First

Amplitude's segmentation becomes powerful when you build around behavioral patterns rather than demographic categories.

The Four Lifecycle Segments

New Subscribers (Boxes 1–3)

This is your highest-churn window. Segment anyone in their first three billing cycles and track their quiz completion rate, portal login frequency, and whether they submitted at least one product review. Subscribers who do all three in the first 60 days have significantly higher 12-month retention.

Engaged Core (Boxes 4–12)

These subscribers have survived initial drop-off. They're your referral and upsell audience. Segment by customization frequency and loyalty point activity — the most engaged within this group are your natural brand advocates.

At-Risk Subscribers

Define this behaviorally, not by billing status. A subscriber who hasn't logged into the portal in 45 days, hasn't customized in two cycles, and gave a product rating under 3 stars in their last box is at risk — even if their payment is current. Build this as a dynamic cohort in Amplitude.

Lapsed and Win-Back

Anyone who cancelled within the last 90 days. Track which win-back touchpoints they interact with before resubscribing, and what their resubscription behavior looks like compared to original acquisition.

---

Funnels and Charts That Drive Decisions

The Onboarding Funnel

Build this as a strict-order funnel in Amplitude:

1. `Subscriber Signed Up`
2. `Quiz Completed`
3. `Portal Logged In` (within 7 days)
4. `First Box Shipped`
5. `Product Review Submitted` (within 21 days of shipment)

Measure conversion at each step by acquisition channel. You'll almost certainly find that one channel — often influencer traffic — drives high signup volume with significantly lower quiz completion, which predicts poor curation fit and early churn.

The Cancellation Flow Analysis

Map every path through your cancellation flow as a funnel. Where subscribers drop off before completing cancellation matters — those are your save points. If 40% of subscribers who start the cancellation flow exit at the pause offer screen without cancelling, that's your most effective retention mechanic and you should invest in it.

Retention Curves by Cohort

Use Amplitude's Retention Analysis chart to build monthly cohorts based on `First Box Shipped` as the starting event and `Subscriber Renewed` as the return event. Compare curves by: acquisition channel, quiz completion (yes/no), and whether the subscriber customized their first box. These three cuts will show you the behavioral levers with the biggest impact on Month 3 and Month 6 retention.

---

Automations and Integrations

Amplitude doesn't send messages — it generates the signals. You need to connect it to your CRM or lifecycle tool (Klaviyo, Braze, Iterable) via the Amplitude Data sync or your CDP layer.

Behavioral Triggers Worth Automating

• At-risk alert: Subscriber enters at-risk cohort → trigger personalized "how is your box?" outreach within 48 hours
• Post-review signal: Subscriber submits a 5-star review → trigger referral program invitation within 24 hours
• Customization window reminder: Subscriber has not made a customization with 72 hours before cutoff → trigger reminder with their most-rated product categories surfaced
• Payment failure sequence: `Payment Failed` fires → suppress all promotional messaging and route to recovery flow exclusively

The Industry-Specific Challenge: Box Delivery Lag

Unlike SaaS products, your subscribers have a physical experience gap — the time between order and unboxing. Most engagement analytics tools don't account for this. In Amplitude, create a computed property called `Days Since Last Box Received` and use it to normalize all engagement metrics. A subscriber who logged in once in 30 days looks disengaged — unless their last box arrived 28 days ago, in which case that login is actually post-unboxing engagement and a positive signal.

---

Frequently Asked Questions

How do I track offline behavior like in-person unboxing events?

Use Amplitude's HTTP API to send server-side events for any offline engagement you capture. If you run pop-up events or retail activations where subscribers scan a QR code or provide their email, you can fire a custom event like `Live Event Attended` and associate it with their existing Amplitude user ID. This keeps your subscriber record complete and lets you measure whether in-person touchpoints affect downstream retention.

Our subscriber base is under 10,000 — is Amplitude worth the complexity?

Yes, with one condition: keep your event taxonomy small. At under 10,000 subscribers, you don't need 80 events. You need 20 instrumented correctly. The segmentation and retention analysis Amplitude provides is valuable at any subscriber volume because subscription economics make even a 5% retention improvement material to your LTV.

How do we handle subscribers who pause vs. those who cancel?

Treat these as entirely separate behavioral tracks. Create distinct user properties — `Current Status` with values like `active`, `paused`, `cancelled`, `reactivated` — and update them via your billing system integration. Paused subscribers who return within 60 days behave more like active subscribers than like win-back conversions. Mixing them into the same cohort distorts your retention curves and your reactivation attribution.

Can Amplitude replace our customer data platform?

No, and you shouldn't try to make it. Amplitude is an analytics and instrumentation layer. It answers behavioral questions. Your CDP (or your CRM used as one) handles identity resolution, audience orchestration, and message delivery. The right architecture is: behavioral events flow into Amplitude for analysis, cohorts sync out to your messaging tools for activation. Keep these responsibilities separate and your data stays clean.