← Back

Comprachat

A shareable catalog that hands WhatsApp a complete order, combining the reach of online commerce with the directness of selling in chat.

Project overview

  • Type: Product · 8-week 0→1
  • Project type: 0→1 · Chat Commerce · Direct Selling · Conversion
  • Role: Lead Product Designer · Opportunity Framing · Interaction Design · UX Research
  • Methods: Contextual inquiry · service blueprinting · usability testing · A/B
  • Tools: dscout · Otter.ai · Figma + Dev Mode · Amplitude · Statsig
  • Case thesis: Designing a shareable catalog that hands WhatsApp a complete, unambiguous order, so a seller gets the orderly catalog of online commerce and the directness of selling in chat, without the gateway, commissions, and setup an online store demands.

The context

Online commerce is supposed to make selling easier, yet standing up a store with a payment gateway is a real barrier: specific configuration, fiscal registration, particular bank accounts, commissions on every sale, and a level of traceability many sellers neither need nor want. That barrier is at odds with the promise of "easy," so a lot of people who are starting out, or who simply prefer to keep selling directly, stay in WhatsApp, the most common way to sell across Latin America. Selling there keeps the directness, with no intermediaries and no commissions. WhatsApp's own catalog exists, but it lives inside the app, with limited customization and without a link to share elsewhere, and it stops short of an online store's browsing and ordering experience. Comprachat gives each seller a shareable catalog at compra.chat/username, with their own branding, that brings order to what they sell and funnels a clean order into the chat where they already close.

The problem

Two frictions sit on either side of the same seller. Standing up an online store is a wall: in research, 68% of sellers who tried a store builder abandoned setup before publishing, citing the configuration, the fiscal and bank requirements, and the commissions (attitudinal). Selling through WhatsApp avoids that wall, but the order still gets assembled by hand. WhatsApp's own catalog exists, yet it is tied to the app, limited in customization, and without a link to share elsewhere, and the order is still negotiated in chat: placing one through DMs took a median of 14 back-and-forth messages per order, with frequent errors (behavioral, chat audit).

The goal

Give a seller the orderly catalog of online commerce and the directness of selling in chat, with no store to stand up, measured by messages to place an order, checkout-to-WhatsApp completion, and seller activation and retention, rather than by store features.

Empathize — Sellers avoided online stores and closed in WhatsApp, where placing one order took a median of 14 messages

In this section: Research foundation · Key insights

Research foundation (method)

  • Phase 1 — Seller interviews (n=20, ~35 min, small and starting sellers in LatAm, recruited via dscout, transcribed in Otter.ai): how they sell today and why they avoid online stores.
  • Phase 2 — Chat audit (n=22 order conversations): counted the messages and errors in placing an order through DMs.
  • Phase 3 — Survey (n=190; 18.4% response, 12.6% completion; select-all and single-select labeled per question): on store-builder attempts and payment preferences.
  • Phase 4 — Prototype pilot (Amplitude-instrumented, 60 sellers): buyer and seller behavior on the catalog and handoff.

Key insights

1. The barrier to entry of an online store is what keeps sellers out. The requirements and setup of a payment gateway were a wall, and a commission on every sale plus the bookkeeping it implied were reasons to stay direct, so the close stays in chat. Triangulation:

  • Attitudinal: 68% who tried a store builder abandoned it; the reasons were configuration, fiscal and bank requirements, and commissions.
  • Verbatim — coded: Store overhead: "I'm starting out, I just want to show what I sell and get paid, not register a company and give up a cut on every order."

2. Selling through WhatsApp keeps the simplicity and leaves the order to assemble by hand. WhatsApp's own catalog ties the seller to the app with limited customization and no shareable link, and in a DM thread there is nothing to browse at all, so the seller retyped product details and quantities and totals were negotiated message by message, which is where orders stalled and errors crept in.

  • Behavioral: a single order took a median of 14 messages, and order errors (wrong item, quantity, or price) were common.

3. Sellers live in WhatsApp, so the tool has to meet them there. Both the buyer handoff and the seller's own management needed to fit the chat, since a separate dashboard was a context-switch these sellers avoided.

  • Verbatim — coded: In-channel: "If I have to log into some dashboard to change a price, I won't; I'm already in WhatsApp all day."

Dashboard — Two frictions on either side of one seller

Two frictions on either side of one seller
Scope: chat audit (n=22) + survey (n=190)
Guiding question: Why don't these sellers use an online store, and where do DMs slow down?

  Abandoned a store builder before publishing  68%
  Messages to place one order in DMs ....... 14 (median)
  Order errors (item / quantity / price) ... common

Key Insight: An online store is a wall to set up, and selling in chat leaves
the order built message by message, so a shareable, structured catalog bridges both.

Define — The catalog had to hand WhatsApp a complete order, in the flow that matched how the seller sold

In this section: POV · How Might We · Principles · Insight→decision map

POV statement. A seller who wants to sell without the overhead of an online store needs a shareable catalog that hands WhatsApp a complete, unambiguous order, because standing up a payment gateway is a barrier and building an order in raw DMs is slow and error-prone.

How Might We

  1. How might we let a buyer build a complete order and hand it to WhatsApp as one clear message?
  2. How might we make creating and editing a catalog fast enough that a seller avoiding a store will do it?
  3. How might we let a seller run the catalog from the chat they already live in?

Design principles (each traceable to an insight)

  • Hand the chat a complete order. The WhatsApp message carries items, quantities, unit prices, and total, so the conversation starts at the close. (Insight 2)
  • The flow matches the sale. A simple catalog uses a direct-buy button; a multi-item one uses a cart, so the steps fit how the seller transacts. (Insight 2)
  • Creation is near-zero friction. A catalog is built in the time it takes to send a few messages, with no store setup. (Insight 1)
  • Meet the seller in their channel. A shareable link, and store management by chat, keep the seller out of a dashboard they would avoid. (Insight 3)
  • The seller keeps the directness. No gateway and no commissions; payment is arranged directly in chat. (Insight 1)

Insight → decision map

Insight (from Empathize) Concrete design decision
One order takes 14 messages in DMs Direct-buy and cart both open WhatsApp with a prefilled, structured order, so the seller confirms and collects in one reply
68% abandon store builders A catalog is created by adding a product photo and price; categories, product pages, and colors are optional and added later
Sellers won't use a separate dashboard Sellers edit prices and details by chat through an AI agent, or in the account, whichever fits the moment

Ideate & Craft — A buyer built an order, and Comprachat handed WhatsApp the whole thing, ready to confirm

In this section: Design execution · Before → after · Other deliverables

Design execution

  • The order handoff — the core artifact is the WhatsApp message: a direct-buy button opens the chat with the product and price prefilled, and the cart opens it with every item, quantity, unit price, and total laid out, so the seller starts at confirm-and-collect.
  • Two modes matched to the sale — a seller sets a simple catalog to direct-buy, or a multi-item catalog to cart, so buyers take only the steps their order needs.
  • The catalog and profile — a grid of products with title and price, optional product pages with multiple images, a description and a share button, simple category filters, configurable store colors, and a profile with store name, description, a main CTA, social links, and a photo, at compra.chat/username.
  • Merchandising tools — mass discounts and mass price adjustments by amount or percentage, and per-product flags for made-to-order, unavailable, or price-on-request.
  • Run it from chat — a seller adjusts prices and details conversationally through an AI agent, or manually in the account, so management fits where they already are.

Before → after

Before (online store, or raw DMs) After (Comprachat)
Starting to sell Configure a gateway, or improvise in DMs A catalog link, no store to stand up
Browsing Seller re-sends photos and prices A shareable catalog the buyer browses
Placing an order 14 messages of back-and-forth A prefilled, structured order in one message
The money Commissions and bookkeeping, or no record Arranged directly in chat, no intermediary

Other deliverables

Built in Figma with Dev Mode handoff: the order-handoff message patterns for both modes, the catalog and product-page system, the merchandising controls, and the chat-based store management.

Dashboard — A structured handoff collapses the order

A structured handoff collapses the order
Scope: Last 30 days · pilot (60 sellers)
Guiding question: Did the catalog get buyers to a clean order in WhatsApp?

  Messages to place one order ........ 14 -> 4
  Checkout-to-WhatsApp completion .... 81% (mode matched to catalog)
  First catalog published ............ ~6 min median

Key Insight: Handing WhatsApp a complete order cut the back-and-forth from
fourteen messages to four and started the chat at the close.

Prototype / Test — Defaulting to a cart matched e-commerce and lost single-item sellers; matching the flow to the catalog converted more

In this section: The experiment · What it taught

The instinct was to give every catalog a cart, the way e-commerce works. It was A/B tested against matching the flow to the catalog (direct-buy for simple catalogs, cart for multi-item ones) in Statsig across the pilot.

The failed variant. For the many sellers whose catalog was effectively single-item asks, a cart added steps the sale did not need, and checkout-to-WhatsApp completion sat at 52%, as buyers dropped between adding to cart and reaching the chat. The direct-buy flow, matched to how those sellers transacted, completed at 81%; the cart still won for multi-item catalogs.

A cart everywhere adds steps a single-item sale doesn't need
Scope: Statsig A/B · pilot · n=2 variants
Guiding question: Which flow gets buyers to the WhatsApp handoff?

  Variant A — Cart by default (single-item catalogs)
    Checkout-to-WhatsApp completion .. 52%
    Steps to reach the chat .......... more

  Variant B — Flow matched to the catalog
    Checkout-to-WhatsApp completion .. 81%   (direct-buy, single-item)
    Cart retained where multi-item ... yes

Key Insight: The e-commerce default added friction to a one-item ask;
matching the flow to how the seller transacts is what converted.

What it taught. Importing an e-commerce convention onto a chat-close model added steps where the sale did not need them; matching the flow to how the seller transacts, a direct ask or a multi-item cart, converted more. The matched-flow model shipped.

Outcomes & reflections

In this section: Causal chain · Reflections

Causal chain (pilot, 60 sellers)

The structured order handoff carried items, quantities, prices, and total into WhatsApp, and the flow matched to the catalog removed steps the sale did not need, so the messages to place an order fell from 14 -> 4 and checkout-to-WhatsApp completion reached 81% for single-item catalogs (against 52% with a forced cart), which let sellers close faster with fewer order errors; because a catalog published in about 6 minutes and could be edited by chat, sellers who would never have stood up a store set up and kept their catalogs, lifting day-30 seller retention from 41% -> 59%.

Metric Before After Δ
Messages to place one order 14 4 ~-71%
Checkout-to-WhatsApp completion (single-item) 52% 81% +29 pts
Time to publish a first catalog store-builder hours ~6 min far lower
Day-30 seller retention 41% 59% +18 pts

Scale note: every message saved per order and every avoided error is a sale that does not stall, multiplied across every buyer who lands on the link.

Reflections (transferable principles)

  • The friction in starting to sell online is the barrier itself; a tool that delivers the orderly catalog of online commerce with the directness of selling in chat lets a seller start without standing up a store or giving up a cut.
  • Matching the flow to how the seller transacts, a direct ask or a multi-item cart, converts better than importing one e-commerce checkout onto every catalog.
  • For a seller avoiding the overhead of a store, the tool wins on how little setup it takes, so a catalog has to be created in the time it takes to send a few messages, and managing it has to fit the channel the seller already lives in.