Comment by acangiano | Hacker News Reader

Comment by acangiano | Hacker News Reader

Parent Full thread

Here are a few positive factors that influence your ranking:

- Exact match keywords are an absolute killer when it comes to ranking

- Local beats global if your target is geographic (e.g., mysite.ca will perform better than mysite.com in Canada, particularly if the .ca is hosted in Canada)

- Number of backlinks to the specific page

- Number of backlinks to the whole domain

- Popularity of backlinks

- A variety of backlinks. Your backlinks need to look natural to Google. You want a backlink pyramid, so don't have 100 PR5 backlinks, and 10 PR0 ones. The web doesn't work that way, and Google knows it.

- For the same reason, have no follow backlinks as well. Whether they carry PR juice or not, it's unnatural for your site not to have a mix of no follow and do follow backlinks.

- Backlinks from .edu/.gov sites

- An aged domain name

- Include the keyword in the url, title tag, description, etc.

- Metatags are still important, even if Google doesn't care about them.

- Make your site as fast as you can

- Have an XML sitemap

- PubSubHubbub is excellent to get your site noticed/indexed by Google

- The number of indexed pages on your site

- How frequently is your site updated (often is better than rarely)

- Make your content unique. Duplicate content is not as much of a big deal as they say it is, but you definitely don't want to look like an autoblog in the eyes of Google (particularly in the early days of your site).

- A public whois record

- A domain registered for more than one year

- Get your site indexed in DMOZ and Yahoo Directory

Why this list is no good:

http://www.seo-theory.com/2011/01/07/why-you-cannot-reverse-...

I think you (and the author) are wrong. It is a good list.

Sure Google's algorithm is continually evolving and uses 10,000 ranking factors. But the web is made of links, and links will always be trust factors for a search engine.

Google just wants to make sure they are 'genuine organic' links.

Let me elaborate, then. The list is dubious. There are some solid arguments against making such lists in the article I linked to. The web is made of links, people, companies, bots and more. What's a "genuine organic" link?

> Exact match keywords are an absolute killer when it comes to ranking

Where are these keywords supposed to be? I'm guessing exact-match domains. I don't think they matter much but that's up for debate.

> For the same reason, have no follow backlinks as well.

I think that's supposed to be "no nofollow". I dislike nofollow a lot. Some of your links are bound to end up in places that use nofollow, no need to worry about it.

> Backlinks from .edu/.gov sites

Spammers have been after those since at least 2005. I'm sure Google knows better than to place a lot of weight on those.

> Metatags are still important, even if Google doesn't care about them.

Meta keywords are dead for sure. Meta descriptions are the oldest standard way to provide a summary of your page, so you should use it. There are also many non-standard meta tags that can prove bad for your site (nocache,noindex,etc).

> PubSubHubbub is excellent to get your site noticed/indexed by Google

Dunno about that, is it any better than a regular RSS feed (and maybe a good old blog ping)?

> The number of indexed pages on your site

If you have enough "positive factors" you will have more pages indexed. How can that be a signal in itself?

> Make your content unique.

How do you measure the uniqueness of your content? Besides that, there's many types of content where uniqueness does not matter and is not expected, for example: lyrics, recipes, press releases, syndicated news and more.

(Duplicate) Content can be aggregated, filtered, commented on, mirrored, reorganised, reformatted and so on.

> A public whois record

Why does that matter? Whois obfuscation is offered by most mainstream registrars, lots of people use it. It's an effective tactic against all the unscrupulous people that mine whois records. Also plenty of ccTLDs do not expose anywhere near as much information as .com and the likes.

> A domain registered for more than one year > An aged domain name

That would be a great way to provide stale content to a lot of search queries. The rate of website creation is surely rising, why do we have to trust their content less? And with many people buying old websites in order to exploit this presumed signal, the noise has certainly gone up.

serious question - have you done a lot of linkbuilding? because i've seen probably 80% of these matter.

what does it matter if google makes 400 algorithm changes/year? you seem to be saying that this means there are no fundamentals to SEO and it doesn't follow.

How do you measure what and how much it matters?

I'm saying making lists of ranking factors is not a good idea. Not that all SEO advice is useless.

you didn't answer my question...

Neither did you.

I commented on half the items in the list to show why I disagree with such lists. Yes, I've done plenty of link-building but I don't think it gives me any more authority on the subject than the next guy.

What's a more important question: "How do you measure which ranking factors matter more?" or "Have you done a lot of link-building?"

I used to track SERPs for dozens of keywords daily for all my domains (I had a lot then), however I've stopped doing than now as I wasn't getting much out of it. I may try it again in the future but I'm waiting until I get some better ideas on signals to track.

Reply on news.ycombinator.com