From Advanced Discounting to Global Offer Platform

Context

Postscript merchants use offers everywhere: campaigns, automations, keyword replies, pop-ups, and AI-driven experiences like Shopper. Requests for advanced discounting came up repeatedly, especially from larger customers.

They flagged the gap as critical, so shipping a solution before Black Friday and Cyber Monday (BFCM) was essential to retaining Postscript’s most valuable merchants.

Key gaps:

Capability: Offer creation was limited to basic, single-use discount codes.
Platform: Offers were created inside individual messages, with no global visibility or source of truth.

Problem

Complex promotions are limited in two ways

Merchants couldn’t execute sophisticated promotional strategy at scale inside Postscript. This showed up in two ways:

Offer creation was too basic

Merchants needed capabilities that are standard in Shopify: Buy X Get Y, offer stacking, subscription, cashback (Fondue), and offer tiering (explored; de-scoped for v1 due to complexity).

Without these, merchants were either simplifying their strategies to fit Postscript’s limitations or moving work to third-party tools like Monocle.

Existing offer creation

Unable to manage offers globally

Offers lived inside individual messages, with no global view. Updating an offer meant editing every instance where it appeared.

Without a global framework, this also blocked Shopper, Postscript’s SMS AI agent, which needed visibility into available offers to send to subscriber.

Design approach

Expand the capability + Establish the platform

This initiative had two layers: advanced discounting, which unlocked Buy X Get Y, offer stacking, subscription logic, and cashback, and centralized offer management, which created a global offer framework and migrated offers from message-level creation to a centralized model.

Constraints

Modal was the only viable format. Offer creation had to be fetched dynamically by FlowBuilder, Segments, Campaigns, and Popups. A full-page experience would have broken the integration points across the product.
Shopper needed this fast. Postscript’s SMS AI agent was blocked until the new modal shipped. It couldn't generate unique codes per subscriber without it.
Shifting business strategy. The offer roadmap changed repeatedly due to team shifts and business changes. That uncertainty made it difficult to design with confidence for what the system would need to support long-term.

Advanced discounting

Design the capability surface

I translated Shopify discount concepts into a Postscript-native offer creation experience. I partnered closely with product and engineering to ensure the UI accurately reflected the underlying Shopify GraphQL capabilities.

New offer type: Buy X Get Y

Fast path for common cases + Progressive complexity

The core challenge was adding power without slowing basic discount creation. I kept a fast path for common cases, and used progressive disclosure by grouping advanced options into a dedicated section.

Pushed back on the basic versus advanced split: We considered using a lightweight default modal for discount creation and a separate “advanced” configuration page for more complex setups. I recommended against this because we didn’t yet have a centralized offers surface where that page could live. More importantly, stacking rules and subscription logic are standard parts of Shopify’s discount configuration, would've added unnecessary complexity and inconsistency.

Simple discount creation and advanced settings

Initial screen groupings

As new offer types were added, including cashback, the selection screen grew. Without structure, the longer list would have felt overwhelming and harder to scan. I introduced groupings, with “Most Popular” at the top for common use cases and “More ways to reward” below for less frequent types.

Consistent IA

Consistent information architecture across offer types: Same layout and grouping patterns reduce cognitive load.

Design iterations highlights

1. I tested multiple structures

I explored a multi-step modal and a full page layout, then landed on a long modal with clean sections. This kept the experience integrated, helped chunk advanced settings, and preserved a fast path.

Layout iterations: page and multi-step modal

Key tradeoff: The stepper reduced overwhelm but made the flow feel longer and introduced a non-standard modal pattern for Postscript. The page approach improved clarity but risked feeling disconnected when users create offers mid-workflow.

2. Iterated on Buy X Get Y based on user feedback.

Buy X Get Y was the most complex offer type, and it took several iterations to arrive at a clean solution. After research, I renamed it from Gift with Purchase to Buy X Get Y, a more inclusive label that matches Shopify’s terminology and more clearly describes the configurations it supports.

Buy X Get Y iterations

Centralized offer management

Establish the platform layer

In parallel, I shaped the direction for moving from message-level offers to a global model. Research with five customers showed that a full reporting dashboard was not a day-to-day need.

Global table with key information

A global table with core attributes (name, type, applies to, stacking, used, sent) and actions to create, edit, clone, and delete.

Shift based on research: My PM and I decided to skipped a “phase 1 view-only table” because customers did not find it valuable. We prioritized shipping a global version with meaningful controls.

Offer management

Unique code visibility JTBD

The core JTBD were seeing where an offer was used across campaigns, popups, and messages, and diagnosing unique code issues by finding a specific code, checking its status, and identifying the instance it was tied to; as a result, I prioritized search and filters over showing a fully populated view by default.

JTBD: View offer used in and unique code (diagnostic workflow)

Fondue x Postscript Integration

Setting the vision

Beyond what shipped, I created concept explorations showing what it would look like to bring Fondue fully into Postscript (cashback creation, analytics, all settings, etc).

Integration discussions had been abstract and inconclusive for months. These concepts gave engineering a concrete view of where cashback creation would live, how settings would be organized, and what the navigation model could look like. That clarity unlocked more specific technical conversations about feasibility, backend requirements, and what we could ship quickly. The team left energized and aligned.

Click here to see more examples

Impact 🎉

Launched

Shipped a new offer creation modal with advanced discounting ahead of BFCM, reducing churn risk for 12 high-GMV merchants representing $869M in cumulative GMV and $1.3M in annual net revenue.
Unlocked Shopper capability by enabling per-subscriber unique codes (not just static codes) and giving Shopper visibility into available offers through the global offer object foundation.

Defined and validated foundation

Validated the offer management table design with users, and used research findings to refine the scope.
Shared Fondue x Postscript vision concepts that accelerated technical alignment and energized the team.

Execution ended due to an org restructure (RIF)

Reflection

The biggest lesson was balancing speed with reversibility. When the path ahead was unclear, I tried to make choices that were easy to extend or adjust later, while staying concrete enough for engineering to ship confidently on a tight deadline.

If I could do one thing differently: push for the user research on the offer management table earlier. Having that data sooner would have helped us skip phase 1 from the start rather than discovering it wasn't useful partway through.

From advanced discounting to global offer management