How to Segment Crypto Users in DeFi

Quick answer: Crypto user segmentation is the process of grouping wallets by shared onchain and in-app characteristics, transaction behavior, holdings, lifecycle stage, acquisition source, or onchain role, so a team can measure and treat different groups differently. The most useful starting split combines two signals: what a wallet does (behavioral activity) and how recently it did it (lifecycle stage). Combining the two usually gives teams a more useful retention signal than relying on either dimension alone, and every dimension after that (value, acquisition source, bot status) refines the picture from there.

Introduction

This guide is for DeFi growth, marketing, and product teams at early-to-mid-stage protocols and apps, especially those operating without a dedicated data analyst. Most teams do not wake up wanting to build user segments. They want to understand why retention slipped, whether a campaign brought in valuable users, or which wallets represent real users instead of bots and one-time visitors.

That is where segmentation helps. It gives teams a way to turn messy user activity into groups they can analyze, compare, and act on. In practice, it helps answer questions like which users come back, which ones convert, and which acquisition channels bring in wallets that matter.

Crypto makes this harder. Wallets are pseudonymous, so teams cannot rely on the same identifiers and profile data common in Web2. At the same time, onchain data makes segmentation more precise when handled well, because teams can see real holdings, protocol activity, and transaction patterns instead of relying only on survey responses or declared preferences.

This guide covers

• What crypto user segmentation means in practice

• Which growth problems stronger segmentation can solve

• Eight practical dimensions for segmenting crypto users

• Segments versus cohorts, and why the distinction matters

• A six-stage lifecycle framework for tracking users over time

• How to build your own segments step by step, and how to measure whether they actually work

• Segmentation patterns by DeFi product type

• The mistakes that tend to cause problems, including the wallet-versus-user gap

• Real examples of DeFi teams using wallet segmentation

• How the main crypto analytics tools compare, briefly

What Is Crypto User Segmentation?

Web2 starts with identity, Web3 starts with behavior. A traditional Web2 segmentation model starts with who a user says they are: age, location, job title, stated preferences. A Web3 model starts with what a wallet has actually done: what it holds, when it last made a transaction, which protocols it touches, and how often it comes back.

That difference is not cosmetic. Web2 profile data is self-reported and often stale within months. Onchain data provides an objective behavioral signal that complements first-party and self-reported data, since it's a direct, timestamped, public record.

Pseudonymity is why behavior carries more weight than declared identity here. A wallet address alone reveals nothing about the person behind it, but it reveals a great deal about what that address has done. Many onchain actions impose observable economic costs, though bots and Sybil operators can still manufacture activity, which is why bot filtering is its own dimension in the taxonomy below rather than an afterthought.

Why DeFi Growth Problems Require Better User Segmentation

Flat or declining retention hides two very different problems. Without segmentation, a team cannot tell whether a retention drop reflects airdrop farmers leaving after claiming rewards, or real users churning for product reasons. Those require opposite responses.

Segmentation addresses four recurring growth problems directly:

• Retention that looks flat or declining, with no way to isolate the cause. A blended number cannot distinguish farmers who were never going to stay from genuine users who left because of a product issue.

• Broad campaign spend with no visibility into which channel brought sticky users. Two campaigns can produce the same wallet count while one brings users who make transactions for months and the other brings users who vanish after a single claim.

• Transaction counts or TVL that look healthy while revenue or fees lag. Volume can rise from a small number of large wallets or from incentive-chasing activity that never converts to fee-generating usage.

• No way to separate whales, bots, and one-time farmers. This turns every growth decision, from campaign budget to incentive design, into a guess.

Signs You Have a User Segmentation Problem

• You treat a wallet that made a transaction once the same as one active for a year.

• Your only user metric is a single aggregate number, such as DAU, TVL, or total wallet count.

• You cannot say which users actually drive revenue without a one-off manual analysis.

• Two campaigns brought in the same number of new wallets, and you have no way to say which one worked better.

Eight Ways to Segment Crypto Users

No single dimension below is sufficient on its own. Effective segmentation combines several, since a wallet can score high on one and low on another. This isn't an exhaustive list, governance participation, profitability, and other product-specific dimensions can also be useful, but these eight provide a practical starting framework.

Wallet infrastructure is a useful acquisition-source and technographic signal. A wallet automatically provisioned through embedded infrastructure such as Privy or Dynamic represents a different onboarding journey from a user who deliberately connects an existing wallet, worth treating as a distinct signal rather than folding both into a single "connected wallet" count. Wallet choice can also correlate with geography and user value, Dune's Wallet Report v2 shows substantial differences between wallet user share and balance share across markets, worth checking if regional segmentation is relevant to your acquisition strategy.

Segments vs. Cohorts: What's the Difference?

These two terms get used interchangeably, and that confusion causes real analysis mistakes.

A segment groups wallets by shared attributes or behavior, evaluated now. "Wallets with more than $50,000 in net worth and 5 or more active days this month" is a segment. It's a snapshot, membership can change day to day as a wallet's balance or activity shifts.

A cohort groups wallets by a shared starting event or time period, then tracks them over time. "Wallets whose first deposit occurred in July" is a cohort. Membership is fixed at the moment of the defining event, a wallet that joined the July cohort stays in it permanently, even if its activity changes later.

In practice, most useful analysis combines both: build a cohort to measure how a group behaves over time, then segment within that cohort to see which sub-group (by acquisition source, wallet value, or entity type) is driving the retention or churn you're seeing. For a deeper dive on building cohorts specifically, see Formo's guide to DeFi cohort analysis.

Formo's Six-Stage User Lifecycle Model

Generic power, core, and casual splits miss most of the useful signal. Formo's lifecycle model classifies every wallet into one of six mutually exclusive stages, based on first-seen date, activity frequency, and recency. Thresholds are configurable per project, so the defaults below are a starting point, not a fixed rule.

These stages are not a fixed sequence a wallet moves through in order. When a wallet meets more than one condition, Formo applies a precedence rule so each wallet receives one clear lifecycle label rather than belonging to multiple lifecycle stages simultaneously.

How to Build DeFi User Segments: Step by Step

1. Define your usage-success metric first. Deposits, swaps, TVL contribution, or whatever matches your protocol's actual goal, before you touch any data dimension.

2. Choose which data dimensions from the taxonomy above actually matter for that metric. Not every dimension is relevant to every goal.

3. Set your segment thresholds. Wallet age, holding size, and activity recency all need concrete cutoffs, not vague categories.

4. Tag entity type and bot/Sybil status before drawing conclusions. A segment built on unfiltered data will misrepresent whatever it's measuring.

5. Validate segments against actual outcomes. Confirm that your defined "power users" actually retain or convert better than other segments, rather than assuming the label is doing its job. See the measurement framework below.

6. Re-segment on a regular cadence. Wallets move between tiers over time, and a segment built once and left alone goes stale.

What this looks like in practice, per Formo's segments documentation:

Behavioral segments like these combine event logic with AND/NOT conditions rather than relying on a single lifecycle label alone.

Common Segmentation Mistakes to Avoid

• Segmenting without a clear goal attached. A segment built before you've defined the usage-success metric it's supposed to serve tends to sit unused, since there's no decision it's connected to.

• Relying on a single active/inactive binary as the whole framework. This collapses meaningful differences between a wallet that churned last week and one that churned six months ago, and misses segments like At Risk, where an intervention still has a chance to work.

• Expecting every wallet to fit cleanly into one bucket. Wallets can score high on one dimension, such as holding value, and low on another, such as recency, or behave differently across chains or protocols. Forcing a single blended score loses this nuance, and can hide a segment that looks strong on paper but is concentrated in a few large wallets rather than broad engagement.

• Treating wallet count as user count. One person can control several wallets, a hardware wallet, a hot wallet, a testing address, and an institutional entity can control many more. Unless you have first-party identity links or a defensible wallet-clustering methodology, treat "unique wallets" as a proxy for users, not a precise count. This matters most when reporting acquisition or retention numbers externally, where an inflated wallet count can misrepresent real reach.

How to Measure Whether a Segment Is Actually Useful

A segment isn't valuable just because it can be built. It needs to behave differently from the baseline, otherwise it's a label, not an insight.

Segment retention Returning wallets in segment ÷ eligible wallets in segment. Compare against your overall retention rate.

Revenue per wallet Segment revenue ÷ wallets in segment. Compare across segments to see where value actually concentrates.

Segment lift Segment conversion (or retention) rate ÷ overall conversion (or retention) rate. A lift near 1.0 means the segment isn't meaningfully different from your user base as a whole, the definition probably needs tightening.

Ask four questions of any segment before you rely on it: does it retain better, generate more revenue, activate faster, or respond differently to campaigns than the baseline? If the answer to all four is no, the segment isn't doing useful work yet.

Segmentation Patterns by DeFi Product Type

The dimensions that matter most shift depending on what your product actually does.

Use these as starting points, and combine dimensions where doing so creates a segment that better reflects the outcome you care about.

Real Examples: How DeFi Teams Use Wallet Segmentation

How WalletConnect Verifies Onchain Activity to Segment Community Creators

WalletConnect Pay needed to verify onchain actions from community creators before accepting content submissions, something Google Forms couldn't do. Formo's token-gated forms auto-verify eligibility at submission using wallet intelligence, and setup took hours. Helena, Community Lead at WalletConnect, put it this way: "We use Formo for creator submissions. The most valuable part is being able to check the onchain actions such as staking the token. It is not possible with Google Forms." The segmentation angle: WalletConnect could segment creators by what they actually held and did onchain, rather than what they claimed in a form.

Read the full WalletConnect story

How Ammalgam Identifies Power Users Before Mainnet Launch

Pre-mainnet on Sonic and Ethereum, Ammalgam's testnet activity gave no way to distinguish real power users from noise or bots, with no defined target personas and no visibility into drop-off patterns. Per Formo's customer account, Ammalgam's dashboards and user-level tracking replaced that gap in a single day, against an estimated 2 to 4 weeks to build custom analytics, that comparison reflects Ammalgam's own reported build-versus-buy estimate, not a general implementation benchmark. Will Fey, Founder of Ammalgam, said: "We chose Formo because it provided the fastest path to actionable insights without adding tooling overhead." Formo's Wallet Intelligence segmented Ammalgam's users into segments such as "Power Users," based on in-app activity, wallet labels, and onchain data, replacing guesswork with behavioral segmentation.

Read the full Ammalgam story

How CoW Protocol and a 2023 OpenSea/LooksRare Comparison Show Retention Diverging

CoW Protocol built a user analytics dashboard on Dune to segment active users, retention, and lifetime value directly from onchain data. Despite a small user base, it found it had "the highest retention rates of their competitive set, with users returning month after month."

A separate, older comparison illustrates a related pattern from a different angle: in a 2023 cohort retention analysis of NFT marketplaces built on Dune, OpenSea's monthly retention held steady across roughly 20 cohorts, near 35% at month 1, 9% at month 10. LooksRare, which ran on heavy token-incentive farming during the same period, collapsed to close to 3% retention across all windows. This is a dated illustration, not current NFT-market data, but it shows how incentive-driven acquisition can produce very different retention patterns from more organic usage.

The Compound Finance Retention Numbers

Castle Labs published a retention analysis of Compound Finance's participation in Arbitrum's Long-Term Incentive Pilot Program in August 2025, built on a Dune dashboard tracking wallet-level retention. Compound received 1.8 million ARB under the program, and during the incentive period (June to August 2024), Compound V3's supply TVL rose 180%, from roughly $90 million to roughly $260 million.

The retention split is the more useful number for segmentation purposes. Wallets that stayed active for 6 or more months and returned within the past 6 months averaged $154,000 in first-month deposits. Wallets that went inactive averaged $9,000, a gap that illustrates why treating all early deposits as equally valuable can hide differences associated with longer-term retention. Castle Labs used the term "inactive" rather than "churned," since DeFi usage is continuous and non-contractual, the same reasoning behind Formo's "At Risk" and "Resurrected" stages.

Choosing a Crypto User Segmentation Tool

These tools overlap with wallet-based segmentation but approach the problem from different categories, including product analytics, marketing analytics, acquisition, and identity resolution. This comparison excludes investor and research tools like Nansen or Dune, which solve a different job again: market and wallet research rather than in-product user segmentation.

Pricing across this category changes often enough that we're pointing to each tool's own comparison page rather than hardcoding figures here.

When you don't want to build the enrichment, identity, lifecycle, and event-logic layer yourself, a platform like Formo combines first-party product events, attribution data, and onchain wallet enrichment into reusable segments, reducing the need to maintain separate indexing, enrichment, and identity-stitching infrastructure specifically for user segmentation.

Growth Strategies Matched to Each User Segment

Get Started With Formo's Wallet Intelligence

Formo's Audience Insights feature builds on the lifecycle model above, surfacing wallet net worth distribution, wallet labels, lifecycle stage, and filterable views by app, token, and chain. Teams can save any combination of these as a reusable Segment rather than rebuilding the same query repeatedly.

Learn more in Formo's documentation.

Try Formo for free.

Final Takeaways

Segmentation only earns its place in a growth workflow when it changes a decision, which campaign to scale, which wallets to exclude, which segment gets a retention push. The taxonomy, lifecycle model, and measurement framework above give DeFi teams a way to move past a single aggregate number and toward cohorts and segments that actually connect to retention and revenue.

The Compound, CoW Protocol, and OpenSea/LooksRare examples show what this looks like in practice: teams that segment by real onchain behavior consistently find retention and value concentrated differently than their raw wallet count would suggest. A segment is only as good as its validation, if it doesn't retain better, convert faster, or generate more revenue than your baseline, it's not done yet.

Frequently Asked Questions

What is crypto user segmentation?

Crypto user segmentation is the process of grouping wallets or users based on shared characteristics such as transaction behavior, holdings, product activity, acquisition source, lifecycle stage, or onchain role. Unlike traditional segmentation, crypto teams can use public wallet activity alongside first-party product data. The goal isn't simply to create categories, but to identify groups that behave differently enough to justify different product, retention, acquisition, or messaging strategies.

How do you segment crypto users by wallet activity?

Start with activity frequency and recency, separating wallets that transact weekly from one-time users, then distinguishing currently active wallets from those inactive for 14, 30, or 90 days. Add product actions such as swaps, deposits, or liquidity provision where relevant. Validate the resulting segments against outcomes like retention or revenue rather than assuming more activity automatically means more value.

What data should DeFi teams use for user segmentation?

Combine product and onchain data. Product data includes sessions, feature usage, referrals, and campaign attribution. Onchain data includes token balances, transaction history, DeFi positions, and wallet age. Identity or reputation signals, such as wallet labels and proof-of-personhood scores, help separate genuine users from bots or Sybil wallets. The right combination depends on the decision the segment is meant to support.

What is the difference between a segment and a cohort in crypto analytics?

A segment groups users by shared current characteristics, while a cohort groups users around a common starting event or time period and tracks them over time. "Power users with more than $50,000 in wallet value" is a segment. "Wallets that made their first deposit in July" is a cohort. Segments are useful for targeting and comparison; cohorts are particularly useful for measuring retention and behavioral change.

How do you identify high-value DeFi users?

Don't rely on wallet balance alone. Combine financial value with product behavior and retention, such as a minimum wallet value, repeated protocol usage, recent activity, and evidence that the wallet generates fees or deposits over time. Validate the definition against actual revenue or retained TVL so large but inactive wallets aren't incorrectly classified as your most valuable users.

How do you distinguish real users from bots and Sybil wallets?

Use multiple signals rather than one rule: transaction timing, wallet age, repeated behavioral patterns, funding relationships, proof-of-personhood signals, and wallet labels. No single heuristic is perfect. Flag probable bots or Sybil wallets separately before calculating acquisition, retention, or campaign performance, since incentive farmers can otherwise inflate user counts without contributing durable value.

How should DeFi teams define churn?

Churn should reflect the normal usage frequency of the product. A DEX might treat 30 days without activity as churn, while a lending protocol with longer-lived positions may need a longer window. It's often more useful to distinguish at-risk, inactive, and churned wallets rather than a single active/inactive binary, and to test whether your chosen inactivity window actually predicts materially lower future engagement.

How do you measure whether a user segment is useful?

Compare the segment against measurable outcomes: retention rate, activation rate, revenue per wallet, transaction frequency, and lifetime value. You can also calculate segment lift by dividing the segment's conversion or retention rate by the overall user-base rate. If a supposedly valuable segment doesn't behave differently from the baseline, its definition probably needs to change.

What is Formo's six-stage lifecycle model?

Formo classifies wallets into six mutually exclusive stages, New, Power user, Resurrected, At Risk, Returning, and Churned, based on first-seen date, activity frequency, and recency, with a precedence rule ensuring each wallet gets one clear label. Thresholds are configurable per project.

What are examples of user segments for a DeFi protocol?

Useful DeFi segments include high-value power users, new wallets that completed their first transaction, at-risk users whose activity has slowed, Sybil-filtered campaign users, churned high-value wallets, and users grouped by protocol role such as depositor, borrower, or liquidity provider. The best segment combines attributes that correspond to a specific decision, then validates the group against retention, activation, or revenue.

Check out our related articles

• How Web3 User Segmentation Transforms Blockchain Analytics

• Wallet Profiling Explained: Boosting User Targeting in 2026

• Wallet Scoring: The Data-Driven Standard for Qualifying and Segmenting Web3 Users

• Wallet Analytics: How to Unlock Growth in Web3 with Onchain Data

• How to Target and Segment Audience in Web3

• Power Users vs Dormant Users in DeFi: How to Identify and Retain Them

• How to Create Custom Audiences from Crypto Wallet Data

• Complete Guide to Web3 Customer Personas: How to Create and Use Them

• Web3 Product Analytics Guide for Product Managers in Crypto

• Wallet Analytics: The Ultimate Guide for Crypto Teams

• The Crypto Founder's Guide to Product Management

• Changelog: user lifecycles, activity chart, and funnel breakdown