On-Chain Analysis Primer
How to Read Glassnode / Dune / Nansen

Traditional financial markets have public volume data but you can't see who is trading. Crypto's on-chain data makes every single transfer publicly verifiable, which gives traders a unique edge — you can see what large players are actually doing, instead of guessing.

This article covers the core concepts of on-chain analysis, the differences between the three major tools (Glassnode / Dune / Nansen), and how to incorporate on-chain signals into a trading system.

1. Why On-Chain Analysis Matters

1.1 Traditional Markets vs. Crypto

Stock market: you only see volume, price, and filings. Institutional positioning, insider trades, and whale intent are invisible.

Crypto: for BTC / ETH and others, every wallet and every transfer is public. As long as you know which address belongs to whom (via crowd-tagging / behavior-pattern analysis), you can track what big players are doing.

1.2 How Much Earlier Than TA Can You See Signals?

Classic example: before FTX collapsed in November 2022,FTX wallets were continuously draining BTC, but price hadn't cracked yet. On-chain analysts spotted the signal and warned 1–2 days in advance, saving a lot of people.

2. Three Major On-Chain Metrics

2.1 Exchange Inflow / Outflow

Inflow to exchanges = preparing to sell (medium-term bearish)

Outflow from exchanges = moving to cold wallets / DeFi (medium-term bullish, "Hodl" signal)

This is the easiest on-chain signal to understand, and one of the strongest. One of the starting points of the 2024 bull market was BTC continuously flowing out of Coinbase, a sign of long-term institutional accumulation.

2.2 UTXO Age / HODL Waves

Every BTC's "last moved" timestamp is queryable. Group all BTC by how long ago they last moved, plot it as a color band chart — that's HODL Waves.

• Long-term holder share rising = strong conviction, stable supply
• Long-term holders starting to move = old BTC coming out of cold storage, often a top signal

2.3 SOPR (Spent Output Profit Ratio)

For each "spent UTXO," was it sold at a profit or a loss? SOPR > 1 = on average selling at a profit, SOPR < 1 = on average selling at a loss.

• SOPR breaking above 1: bottom signal (people start feeling confident enough to take profit)
• SOPR breaking below 1: capitulation phase at the bottom

2.4 MVRV (Market Value / Realized Value)

Market cap ÷ on-chain realized value (each BTC's price when last moved × quantity).

• MVRV > 3.5 → historical top signal
• MVRV < 1 → historical bottom signal

MVRV nailed the tops and bottoms across past cycles, but past performance doesn't guarantee future results.

3. Comparing the Three Tools

3.1 Glassnode

Veteran on-chain analytics firm. Strength: complete BTC / ETH metric library. MVRV / NUPL / SOPR and other academic-style metrics were either invented or popularized by them.

Best for: long-term investors who want to read "market-structure signals"

• Free tier: basic metrics available
• Advanced: $39/month (plenty for retail)
• Professional / Enterprise: institutional use

3.2 Dune Analytics

Strength: custom SQL queries, with a huge library of community-shared dashboards. Best for analysts who need custom metrics or want to track specific protocols.

Typical uses:

• Querying liquidity history for a specific Uniswap pool
• Tracking holder distribution for a specific NFT collection
• Watching stablecoin mint / burn records

The free tier is enough for retail; paid tiers mainly unlock higher query limits.

3.3 Nansen

Strength: address labeling system. They have 140M+ tagged addresses (Smart Money / VC / Exchange / DEX, etc.), allowing named tracking of specific whales.

Typical uses:

• Watching recent activity from VC wallets like a16z / Paradigm
• Tracking which tokens "Smart Money" labeled addresses are accumulating
• Seeing who's buying and selling new tokens

Pricey: starts at $150/month. Best for active investors; retail users can substitute the free Nansen Query alternatives.

4. How to Combine All Three

Professional workflow:

1. Glassnode for the big picture: MVRV / SOPR / NUPL to determine the overall market phase (accumulation / distribution)
2. Nansen to find specific opportunities: what new tokens is Smart Money accumulating
3. Dune for custom verification: write SQL to confirm flows for specific wallets

5. How to Plug On-Chain Signals Into a Strategy

On-chain data usually updates slowly (hourly / daily), so it's not suited for scalping. But it works well as a medium-term directional filter:

# Conceptual logic
if MVRV < 1.5 and Exchange_Inflow_7d < threshold:
    # On-chain bottom signal + no more selling pressure inflows → medium-term bullish
    bias = "long"
elif MVRV > 3.5 and Long_Term_Holders_Selling:
    # On-chain top signal + old BTC coming out of cold storage → medium-term bearish
    bias = "short"
else:
    bias = "neutral"

# Entries still use TA signals (breakout / RSI / EMA), but bias dictates directional preference

6. Free Alternatives

• LookIntoBitcoin: free BTC on-chain metrics, similar to Glassnode lite
• Cryptoquant: partly free, particularly strong on exchange inflow / outflow
• Etherscan: free lookup for individual ETH address activity
• Dune community dashboards: search "Smart Money" or "Whales" to find popular queries
• Arkham Intelligence: free address-label database (a Nansen competitor)

7. Common Retail Mistakes

1. Treating on-chain signals as short-term entries. On-chain = big trend direction, not entry timing
2. Chasing whatever "Smart Money" just bought. By the time you notice, they may already be up 30%. You want the early accumulation phase, before they move
3. Only watching BTC on-chain metrics. Altcoin on-chain structure is completely different — BTC MVRV signals don't translate to SOL / SUI
4. Treating "whale transfers to exchanges" as an automatic sell signal. It might just be a wallet move / custody change, not a sale

8. Three Key Takeaways

1. On-chain analysis is a crypto-only edge — other asset classes can't see "who's moving"
2. The three tools each have their strengths: Glassnode for the big picture, Nansen for specific opportunities, Dune for custom queries
3. On-chain is a "directional filter" — entry timing still belongs to TA. The two combined is what makes the system complete