Holder behaviour by tier

Highlights the mean and median months-to-expiry for buckets of wallets based on the number of ENS names that they own.

Retention by holder tier (incl Dropped)

Tier	Wallets	Names held	Mean (months)	Median (months)
1 name	350,439	350,439	40.4	14.8
2–5 names	65,482	173,234	49.2	28.6
6–10 names	8,642	64,302	40.0	23.7
11–50 names	6,822	141,024	31.0	17.1
51–100 names	836	58,070	22.6	13.0
100+ names	621	160,951	22.4	12.1

For each tier (where tier = current names-held tier of the original registrant), how many names that wallet ever first-registered are still active today? The Dropped row tells you what fraction of names registered by exited wallets persisted in any form.

"Still held by original" is the stricter signal - the registrant didn't just see their name get re-registered by someone else, they actually still hold it. Big gap between "still alive" and "still held by original" means lots of secondary-market churn.

Registration duration choice by tier

Tier	Names ever registered	Still alive (any holder)	% still alive	% still held by original
Dropped (0 current)	1,690,618	155,026	9.2%	0.0%
1 name	574,554	306,311	53.3%	46.4%
2–5 names	423,803	167,818	39.6%	29.7%
6–10 names	157,861	60,089	38.1%	27.9%
11–50 names	291,088	111,974	38.5%	29.5%
51–100 names	84,702	38,143	45.0%	36.6%
100+ names	158,043	85,925	54.4%	47.8%

When wallets in each tier register a name, what duration do they pick? % of each tier's registrations.

Renewal duration choice by tier

When wallets in each tier renew, what duration do they pick? Renewer = the address that signed the renewal transaction; their tier is their current names-held bucket.

Methodology

Tier classification: each wallet's CURRENT count of active .eth 2LDs determines its tier. A wallet that holds zero names today is classed as 'Dropped' regardless of past activity. Tiers don't move historically - a wallet that registered names in 2022, then dropped them all, is in 'Dropped' for all four metrics.
Wallet-weighted expiry: each wallet's own mean remaining months across its names → tier mean. Sibling to Holder concentration's name-weighted mean (which lets a 100+ wallet dominate the tier average via volume).
Identifying the actor: tx_sender_address (the EOA that signed the registration/renewal transaction) was added to the indexer's event capture relatively recently, so older events have it NULL. For NameRegistered we fall back to owner_address (the address the name was registered to) - reliably populated and almost always the same as tx_sender for typical self-registrations. They only diverge for relayer/proxy registrations, where owner_address is the more semantically meaningful 'intended holder' anyway. NameRenewed events don't carry an owner_address; renewals without tx_sender are simply excluded from the duration-by-tier view, so renewal coverage is reduced to the post-enrichment era.
Retention 'still alive' = the label currently has an active expiry, regardless of who holds it. 'Still held by original' = active expiry AND the current owner address is the original registrant (the registrant for that label, identified via the rule above). Differences between the two columns reflect secondary-market churn.
Duration picks: NameRegistered duration = expiry timestamp − block timestamp at registration. NameRenewed duration = new expiry − immediately prior expiry for that label. Bins use a ±0.05y slack around 1 / 3 / 5 years to absorb block-time slop.
Mainnet .eth controllers only.

Wallet-weighted remaining expiry by tier

Retention by holder tier (incl Dropped)

Registration duration choice by tier

Renewal duration choice by tier