Every study showing “older domains rank better” has the same fatal flaw. It only samples survivors.
The invisible graveyard
Domains registered in 2010 that failed don’t appear in any dataset. They expired. They got penalized. They were abandoned. They’re not ranking because they don’t exist.
You’re looking at 2010 domains that succeeded and concluding age caused success.
That’s like surveying 90-year-olds about longevity habits and concluding everyone should do what they do – ignoring the millions who did the same things and died at 50.
What the data actually shows
Sample: domains currently ranking in top 10.
Observation: average registration date is 8 years ago.
Wrong conclusion: age helps rankings.
Right conclusion: domains that did things right for 8 years are still here. Domains that didn’t are gone.
Survival came first. Age is a byproduct.
The selection mechanism
What kills young domains:
- Poor content → no traffic → abandoned
- No link acquisition → no authority → invisible
- Bad UX → high bounce → algo demotion
- Penalty → deindexed → gone
- Business failure → domain expires
Domains that avoided all these for 10 years demonstrated sustained quality. The ones that didn’t are dead.
Age correlates with survival. Survival correlates with quality. Quality causes rankings.
Age is two steps removed from causation.
Why this error persists
Survivorship bias is cognitively invisible. You can’t see what’s not there.
SEO tools only show existing domains. Search results only show ranking domains. Case studies only feature successful domains.
The entire observable universe is survivors. The graveyard requires deliberate excavation.
How to correct for survivorship bias
Compare domains of same age with different outcomes.
Take 100 domains registered 2015. Some rank well. Some rank poorly. Some died.
What differentiates them? Not age – they’re identical age. Content quality. Link profiles. User signals. Business execution.
Age held constant reveals age is not the variable.
The cohort analysis approach
Track a cohort from registration forward, not from success backward.
Register 1000 domains today. Follow for 5 years. Document what happens.
Prediction: success distribution will correlate with effort and quality, not time passage. Failed domains won’t rank regardless of reaching 5 years old.
This study structure eliminates survivorship bias. It’s also why no one does it – takes years and delivers uncomfortable conclusions.
What survivorship bias means for your strategy
Stop benchmarking against survivors and concluding you need their age.
Benchmark against their inputs:
- How much content did they create?
- How many links did they earn?
- What was their investment level?
- How consistent was their execution?
Match inputs, not age. Inputs are controllable. Age is not.
How should SEO research be designed to avoid this bias?
Include failures in the sample. Pull domains from historical Wayback data that no longer exist. Compare to survivors.
Control for confounding variables. Age without controlling for backlinks, content volume, and investment tells you nothing.
Use longitudinal tracking. Follow domains from birth, not from success.
Until research does this, treat all “age correlates with rankings” claims as methodologically invalid.