Link Forensics: Finding Shady Links Before Taking on New Clients

We fall on the very search-engine-friendly side of SEO (I hesitate to use the phrase "white hat" because I think all the lines often blur). Despite this, I have a lot of respect for well-executed darker tactics and find it fascinating to watch and deconstruct them.

What I don't want, however, is to take the blame on a client project for shady tactics employed before we even worked on the project. Some time ago, at Distilled, we had an issue with this when a client had a website banned. We weren't to blame, but it was very bad timing - the course of events looked like this:

• Client (a pan-European brand) hires us as "outsourced SEO department" reporting to their chief marketing people
• We make recommendations regarding their main site
• Recommendations are implemented and traffic begins to rise
• Client buys a company with a well-known brand in one European country (SEO obviously falls under our remit since we are their SEO department)
• Shortly afterwards (and before our first recommendations), the website for the new brand gets banned...

What had actually happened was that their past had just caught up with them. The smaller brand had previously engaged a local SEO firm with tactics that were a bit past their sell-by date (mainly consisting of a huge link network of sub-domains off the SEO company's own site and some other properties). There was not only a question of tactics here, but also of plain common sense - not only was part of the link network on the SEO company's site, but so were proud links to "our clients" (all of whom got slapped, we believe).

Now, anyone who has worked with us will know that we typically look to improve the things directly under the client's control first - most decent size websites have link equity they aren't spending well and so technical and structural stuff is normally our first target. Many clients, unfortunately, still think we have magic buttons under our desks (or perhaps they're hoping we'll submit them to a few thousand search engines), and so it was perfectly plausible to them that even though we hadn't yet made any recommendations for their new site, we might have pushed the button under our desk a little hard and got them banned.

It might be how some SEO companies work - get a new client and plug them into your shady link network - but it's not our style. These kind of tactics might have their place, but in my opinion, that place is not when you're playing with brand websites.

So, once we had diagnosed the issue, calmed the client down, bullied the old firm to remove the network, submitted grovelling reinclusion requests etc., we started to think....

The nature of our business is such that many clients have implemented 'SEO' suggestions before (whether in-house or agency) and many of them have pasts that contain the odd closet with perhaps a skeleton or two. How could we avoid this kind of scenario in the future - where we might get the blame for the sins of our predecessors?

The Pre-Sales SEO Due Diligence

Out of this conversation came a concept that we have gone some way towards but not 100% cracked yet. This is the idea that before signing a new contract, we should undertake due diligence regarding previous tactics independently of quizzing our prospective client (not only are clients not always up-front about previous tactics, but personnel could have changed, external agencies could have been responsible without being straight with the in-house team, etc.).

So, we are now starting to look out for a variety of things that signal warnings to investigate more closely before signing with a given client. We are looking for things like:

• Manipulative patterns in their backlinks
• Cloaking
• Doorway pages

Effectively, we are thinking like search engineers and giving them a little bit of a manual review. Now, we don't have the tools that search engineers have at their disposal, but we can cobble together a bit of a toolkit to get some way towards this kind of process (and as I mention below, I'd love to hear how you guys go about this).

Pre-sales is not the only time that you might need to do this - sometimes you are going to be paid to do it. This happens when you are hired to work out what has gone wrong with a site when the owner can't help. They might not be able to help because either the board doesn't know details of what was done in the past and the team has moved on, the company purchased the site as a whole, or they outsourced SEO to a less-than-reputable source. As part of our global associate role with SEOmoz, we answer a lot of Q&A and this kind of diagnosis forms a fair bit of that work.

So, finally, I'm going to get to the point and present my methodology for diagnosing manipulative issues with websites:

Forensic SEO Process

I look for three main things:

1. Generally deceptive on-site practices (keyword stuffing, excessive internal linking, doorway pages)
2. Cloaking or other unusual serving of information
3. Offsite manipulation - strange linking patterns, etc.

1. For the on-site stuff, I typically:

• do a site: query and just see how many pages they have indexed versus the apparent size of the website
• search for some of their keywords across their site
• view the source of a few key pages
• check their internal linking structure

It's pretty much a short, basic site review.

2. To check for cloaking, I typically change my Firefox user-agent to Googlebot, disable javascript and cookies and browse around a bit watching for different site behaviour. This isn't enough to pick up sophisticated cloaking but it gives a good overall impression. The ultimate check is to compare the Google cache of their pages to the originals. This is a bit more time-consuming, but is the only way I know of to pick up all kinds of cloaking.

3. The one that caught us out with our client was deceptive linking practices. In an attempt to spot that, I delve into their backlinks a little. What I'm looking for here is things like:

• Repeated optimised anchor text
• Sitewide links
• Links from many low quality sites (long, hyphenated URLS, blogspot domains, etc.)
• Footer / sponsored links
• Hidden / cloaked links

Essentially, I look for anything that "smells wrong." Rand has written before about experienced SEOs' sixth sense, and this is definitely the place for those skills.

I'm interested to know whether this is part of your process and what tools, tips or tricks you use. Share all in the comments!