Paid links - no argument, just dealing with it

Matt Cutts is asking for spam reports on link selling, and doesn’t receive a warm welcome from SEO’s regarding Google’s newest economic policy.

Let’s look at what Matt’s actually saying:

1. Google thinks paid links really impact Google
2. Google is testing at least two new algos to spot and devalue paid links
3. Matt wants reports on live examples they can use the new algos on

The debate over paid vs free links is pretty futile here - Google are giving clear warning that they are looking to attack paid link placements more aggressively, and that’s worrying, as a lot of SEO is built on link buying.

Let’s also not presume that Google are going to start automatically banning sites for selling links. I’ve been on the sticky end of Google’s spam control a couple of times - actually more for autogenerated content than anything else, a tactic I’ve long since abandoned anyway - but overall I’ve found Google to be very fair in trying to allow for as much room for error as possible. You have to be pretty aggressive to get a greyed out PR bar or banned domain.

After all, Google’s primary mission is to process the world’s information - not primarily remove it’s sources.

Google have plenty of patents already in play to evaluate links - Hilltop, LocalRank, TrustRank, Topic Sensitive PR, etc. We play along and try to avoid being filtered. The same with any paid links algo.

For those who continue on heedlessly, a Googlequake is coming - sites shaken out from the SERPs and sucked into oblivion.

As SEO’s, however, most of us have learned to be proactive with Google. They’ve given notice that there are a couple of algos that could be used, so let’s look at what may be involved:

1. Algo for filtering by common on-page signals

These signals should be pretty obivous and easy to determine - headings such as “Sponsors”, “Advertisers”, “Paid links” etc followed by a list of links are surprisingly common features of many sites selling links.

All Google has to do is locate any section of code proceeded by these on-page keyword signals, either as plain text or alt text, and devalue the links in the code block - usually either a clear set of td’s or div tags.

The complication isn’t so much recognising these automatically, as much as integrate such recognition with Google’s increasingly complicated core algorthmic processes.

2. Algo to filter by network

We’ve already seen Google attempt to devalue links based on network connections in October 2005, when link exchanges apparently took a dive.

This is probably a favoured method of Google because they are more likely to take a mathematician’s view on how to devalue links, rather than a webmaster’s, and that means applying a little network theory and looking for common quality indicators.

Simple poor quality indicators are going to be link lists which cover a disparity of topics all placed together - and as so many webmasters whore their link sales to any bidder, gambling and pharamaceuticals links should be an immediate flag when mixed with business, technology, and finance topic sites.

Another indicator could be domain history factors in the sites being linked to. Despite sandboxing processes being in play since early 2004 (I eventually got tired of telling people it existed until it found mainstream acceptance in 2006), many webmasters are still clueless about such SEO issues, so they simply buy links on PR factors alone - and for new sites.

So lists of links which are primarily to new sites, and additionally placed sitewide on sites, should be pretty easy to spot with a little application of network theory.

Also, if a number of sites being linked to in a list are affiliate sites not cloaking their affiliate links, there’s a good chance these are paid for.

Overall

Google want sites to rank according to Google’s own criteria, removed from as much outside influence as possible. SEO is built on creating that influence, and Google has long fought it.

Devaluing paid links has already been a part of Google’s perogative, so algorthmic attempts to combat this shouldn’t really surprise anyone.

Rather than banning sites, Google will more likely simply null value the links - this means the sites selling the links lose link equity by linking out, but the sites being linked to gain no value. I’ve seen this happen anyway on some sites, so I’m presuming the trend will continue.

So what sort of links should we seek to buy that may avoid the basic applications of such coming algorithms?

1. On topic

Relevant links on relevant sites - keeping as close to the topic as possible.

2. Avoid FFA’s

Free For All webmasters who’ll sell links to anyone should be avoided - bad neighbourhood.

3. Seek trust

Forget PR, go on age of domain and domain history - older sites as having more potential for leverage via Trust than newer high PR domains.

4. Diversification

Diversify link strategies to cover different link formats - home page links, sitewides, internal page links, presell pages, editorial links, etc.

5. Indirect purchases

Not all links have to be directly paid for - freebies and webmasters incentivised to cover other websites offer a good channel to editorial links if done right.

Additionally, as Google appears to like stepped processes, identifying individual sites for review where new algos may suggest them link buying probably requires multiple criteria to be fulfilled - so as ever, try to avoid any specific link development strategy that wouldn’t survive manual review from being the overall determinator of a site’s link profile.

Hopefully all of this will help. :)

Meanwhile, anyone spending their time arguing for paid links really should consider themselves to have strong enough links profiles before expending their energy in this manner. Personally, I’m not going to take that risk - time to spend my Sunday afternoon and evening looking at additional link development options to help future-proof sites and clients against Googlequakes.