SEO Analysis Via Aggregated Data: Importance of Domain Age | Hacker News Reader

SEO Analysis Via Aggregated Data: Importance of Domain Age

serpiq.com

11 pointsdchuk19 comments

> Over 93% of 1st position domains are over a year old. This not only confirms the fact that Google favors older domains, but heavily stresses the importance of seeking aged domains for your projects if possible.

No, it doesn't. Correlation != Causality.

you're absolutely right, that's why we included this disclaimer in the post:

Important Note: For this data, along with every other post in this series, we're by no means claiming that the key to ranking a site is simply to buy an old domain, nor are we claiming that a new domain can't be ranked. What we're trying to do is provide data that, when used holistically in your SEO campaigns, can give a competitive edge.

We're just trying to add to the list of best practices you can follow when trying to optimize your site for the search engines.

It's good that you included that disclaimer, but that line quoted above is simply false. The data you posted is interesting, but it does not confirm that Google favors old domains.

We're going to edit our post to clarify that and take a step back from our assertion because we do agree with you. Thanks for pointing this out guys.

EDIT: We've amended our post to clarify the point you guys have brought up. Thanks again for the feedback.

More likely older domains have simply acquired more backlinks.

This seems like a poor analysis to me.

Did you control for other factors? It's likely that older domains have had more time to build links, trust, authority, etc.

It's entirely possible that Domain Authority has a more causal relationship with ranking position, and that domain age is unrelated.

I'm not even sure what the baseline is. What percentage of domains with any relevant content at all are >1 yr old? For it to be surprising that 93% of first results are from >1yo domains, it would have to be the case that >1yo domains make up much less than 93% of the relevant population, so we can conclude that they're actually overrepresented in the top results.

Good point. Unfortunately, we don't have any realistic means to collect that data so we can't really accommodate such an analysis. We're working with what our product has collected and doing our best to share that and apply our own SEO-based interpretation of the data.

There are a handful of unsubstantiated claims that float through the SEO world that we're trying to provide some actual numbers for. If we could without a doubt show you what works in SEO we'd be sipping pina coladas on an island somewhere :). All we're saying with this series of posts is "out of all of the keywords people have analyzed using serpIQ, this is what we make of this specific assertion/data point/metric"

We're not by any means claiming that we made a purely scientific experiment as you can't particularly run with controls against Google anyways. We're simply taking a look at the data our product has collected in the aggregate and publishing it and adding some interpretation of the data so that people can take what they wish from the data and apply it to their own campaigns.

It would be useful to compare to a raw graph of the domain age of your customers. There's a chance that it looks similar to this - and that even if Google picked the #1 ranking site at random, then this chart would look the same.

sorry I don't think I understand what you mean by raw graph

SEOmoz don't appear to feature "domain age" as a specific point in their ranking factors survey - http://www.seomoz.org/article/search-ranking-factors#predict....

I think the general opinion there is that excepting an initial possibility of sandboxing for new domains the domain-age boost is merely a misreading of incoming linkjuice.

Logically this seems right as a domain with equal content (by algorithmic analysis) that acquires links from high authority domains, and more incoming links generally, in a shorter time and beats down an older domain in doing that, such a site must surely be better in some way for users? [run on sentences are a speciality of mine].

We're not really referencing anyone else's surveys or reports on these ranking factors, we're more trying to substantiate what we feel are common around the SEO and internet marketing worlds.

One way to look at the data we posted: it's much more likely that an old domain has built up trust, content, and links than a fresh domain, so they are a worthwhile investment for your SEO campaigns.

Hypothetical situation:

Two domains with identical volume/quality of content and equally good back links, Site A being 10 years old, Site B 1 year old. ( IE remove all considerations other than time )

My money is on site B showing higher in the serps.

interesting conclusion there, what makes you lean that way?

I suspect the newer domain would look more active, more current and therefore more likely to host content useful now/of contemporary interest.

I suspect ranking 'degrades' slightly over time - don't keep adding content and lose places in the SERP.

Having said that, it's pretty much hypothetical. A ten year old domain will have more links and more content, and I think _that_ is why old domains rank well. It usually takes time to develop a great quantity of useful, ranking content and links, hence the perception that "aged domains" are more valuable.

Anyone else have thoughts / different view on this?

I agree with you on everything but your first point. It's just as reasonable to assume an old domain has fresh content as a new domain, there's no connection between domain age and how fresh the content is on it/how active it is. Look at Reddit, or for that matter HN. Older domains, fresh content every second.

But yeah, old domains basically have had more time to accumulate trust/links/shares/content, so they're more likely to do better in the search engines.

If you purchase an old domain, wouldn't Google notice the change of IP address and whois information?

I'm only speculating that buying an expired domain, for example, would be flagged.

Well there's nothing inherently wrong or bad about buying old sites or dropped domains, so it shouldn't raise too much of a red flag

Reply on news.ycombinator.com