Blum

Built spot trading experience for casual users and experienced traders, scaled
to 50.8M MAU at peak

Blum is a crypto trading app that scaled to 95M+ users. I led the end-to-end design
of Spot Trading. I built core user flows, a scalable design system, and a UX architecture
that gave each type of user their own trading experience: casual users got simplicity
without friction, experienced traders got full control.

ProductWeb3 / Exchange / DEX / DeFi

RoleLead Product Designer

YearsSEP 2024 – FEB 2026

Context

Blum started as a tap-to-earn product inside Telegram with tens of millions of users. The next strategic step was launching Spot Trading, bringing full token trading on TON and Solana directly into the app.

Problem

Blum's users were there for the airdrop, not for trading. Once the airdrop was over, there was nothing to keep them in the app.

Goal

Convert the existing tap-to-earn audience into active traders on TON and Solana. Establish a sustainable source of trading volume and fee revenue, and keep users trading within the Blum ecosystem.

Benchmarking

I started with a competitive analysis of two types of products: CEX platforms (Bybit, OKX, Binance) and on-chain tools (Axiom, GMGN, BullX). I looked at how they structure information, how accessible they are for casual users, and how well they serve experienced traders.

Binance

Bybit

Axiom

GMGN

Binance

Bybit

Axiom

GMGN

Blum vs. other CEX and DEX platforms

Blum vs. other CEX and DEX platforms

Benchmarking Insight

The key insight was that all these tools are built for experienced traders and not designed for casual users. This was an opportunity for Blum to fill a gap that nobody had closed. Blum was the only product that allowed users to trade directly inside Telegram with no registration and no KYC. Users could trade inside the app they use every day.

First version: Fast Launch

The first version of Spot Trading shipped fast. Speed was the priority. The team needed to validate that users would trade inside Blum at all. The hypothesis worked: users bought and sold tokens.

Key problem

The first version had one structural problem: the trading flow was built as a single interface with no separation by user type. A casual user who just wanted to buy a token quickly, and an experienced trader who needed metrics and fast execution, both landed on the same screen.

Discovery

To understand how to solve this, I ran in-depth interviews and CustDev sessions (8 casual users, 5 experienced traders) and analyzed metrics. This revealed key problems:

Casual users didn't understand basic trading terms. The amount of data on screen made them nervous and scared of doing something wrong.
Discover didn't help users figure out what to buy. Every token looked the same and there was nothing to guide the choice.
Experienced traders needed everything in one place: key metrics, a chart, and an order form, just like on CEX platforms. Fast execution and a clear view of all open positions, without jumping between screens.

The key insight — a single interface couldn't serve both user types without failing one of them. Based on this, I formed two hypotheses:

Hypothesis 1

If I design Discover around different types of user intent, it will help users find a token faster, increase conversion to a trade, and turn Discover into the top of the trading funnel.

Hypothesis 2

If I split the interface into two trading modes, Lite and Degen, each user will get an interface built around how they trade, increasing retention, trading volume, and daily activity.

Success criteria

⁕ Increase conversion to first trade

⁕ Increase average order size

⁕ Increase retention rate

⁕ MAU / DAU growth

⁕ Grow total trading volume

⁕ Both modes are actively used

The experiment is complete when each mode reaches 1,000+ token page sessions and
500+ executed trades, with statistically significant differences across key metrics (p < 0.05).

New Discover

Before a user can trade, they need to find something worth trading. The original Discover was a token list, functional, but passive. It didn't help users answer the core question: "What should I buy?"

The updated Discover introduced three ways to find a token, each targeting a different user intent. Each entry point leads directly into the trading flow. Discover became the first step of the trading funnel, not just a browsing screen.

1. Spotlight

Spotlight surfaces algorithmically curated feeds of trending tokens, top gainers, and notable market movements. For users who don't have a specific token in mind but want to find what's worth buying right now.

2. Ask AI

Ask AI turns token discovery into a conversation. Instead of scrolling through lists, the user asks what they're looking for and gets a structured result ready to act on.

AI helps users find tokens but doesn't provide financial recommendations. The decision to buy is always the user's own.

3. Copy Trade

Copy trading lowers the barrier for users who don't know what to buy. Instead of researching independently, the user follows real trades made by other traders and copies them in one click.

Token, position size, and PnL are visible at a glance. A trade can be copied directly from the feed with a preset amount that can be changed at any time.

Lite and Degen trading modes

The core architectural decision was to build two separate trading modes, each designed around a different way of trading. Lite removes complexity for casual users, Degen gives experienced traders full control. The mode persists across sessions and switching between them feels natural, like adjusting a preference, not changing products.

Lite Mode

Lite mode is built for users who want to buy and sell tokens quickly without diving into market data. It reduces cognitive load and shows only what's needed to make a decision.

In Live mode, other users' purchases appear
in real time to highlight trading activity.

Degen Mode

Degen mode is built for traders who need metrics, a chart, and fast order execution. Everything is on one screen: the order form, chart, and key metrics without extra taps. The trader makes a decision and executes it instantly, without losing sight of the market.

More filters for search results, displayed directly on the page • One-click buy amount setup • Trading Presets Setup • Market Cap as the key token metric • Customizable token info display: Volume / Price / Liquidity

Key metrics such as market cap,
price, liquidity, volume and holders always on screen

Collapse chart to focus on trading, expand to track live price action

A widget with three buying modes: • Instant for quick one-click buy • Market / Limit with Take Profit and Stop Loss options

More filters for search results, displayed directly on the page: One-click buy amount setup, trading presets setup, key token metric. Customizable token info display: Volume / Price / Liquidity.

A widget with three buying modes: Instant — for quick one-click buy. Market / Limit with Take Profit and Stop Loss options

All open positions in one place: full portfolio view with current value, average buy price,
and unrealized PnL for each token.

A live feed of other users' buys and sells in real time. Helps assess trading activity and make decisions based on current market context.

A quick token safety check covering mint, freeze, blacklist, and holder concentration.
All without leaving the app.

All open positions in one place — full portfolio view with current value, average buy price, and unrealized PnL for each token.

A live feed of other users' buys and sells in real time. Helps quickly assess trading activity and make decisions based on current market context.

User testing

I tested the updated design and new flows through Useberry and Maze, and analyzed heatmaps in Hotjar to see where users focused their attention.

Casual users found Lite mode simple and approachable. The updated Discover helped them find tokens without endless scrolling. They also appreciated seeing their token balance directly on the page.

Experienced traders said Degen mode became a proper trading tool for their needs. One-click presets and the collapsible chart were among the most appreciated features.

Results

I designed Spot Trading from the ground up, converting Blum's tap-to-earn audience into active traders. The two-mode trading UX architecture and redesigned Discover turned the product into a sustainable source of trading volume and fee revenue.

50.8M MAU

At its peak

20.6M DAU

At its peak

+13.2%

Conversion to first trade

+11.6%

Growth in average trade size

28.4% → 36.7%

Retention rate (7d)

72% / 28%

Casual users / Experienced traders

Key Learning

When one interface tries to serve two opposite user types, it will always work better for one of them. The solution is to give each user type their own interface, built around their specific tasks.