Whether you’re an SEO taking a new client or a business that has just been hit by a Penguin (Google link penalty) update you’ll want to tackle your link cleanup quickly and efficiently. This poses an interesting dilemma – if you have a lot of links, it can take a long time to go through these and clean them up. I’m going to cover the manual approach I take to tidying up a site’s link profile, including the techniques I use to ensure the right people do each step, making the process faster and less painful for the business.
It might be tempting to run one of the many paid, automated tools over your link profile and remove as many the bad links it suggests as you can and then disavow the rest and call your work done. I’d argue that since Google clearly can’t spot and ignore all bad links, it’s unlikely any non-Google tool can do the job perfectly either.
With some suggesting that you must wait until the next Penguin refresh to see whether your work has been successful you face a long wait if you get things wrong. Keeping that in mind, I prefer to, when the workload isn’t prohibitive, tackle the clean up manually.
Step 1: Collate Your Backlinks
You’ll be able to grab backlinks from a number of sources. I’d try to mix together as many sources as you can including, but not limited to:
- Backlink tools – Ahrefs, Majestic, Open Site Explorer
- Google Webmaster Tools
- Link reports from your previous SEO firms
If you only sign up to one of the backlink tools I’d suggest you pick AHREFS. Matt Woodward covers the different tools here in more detail for those of you who want to see some quantitative analysis (and some fun responses from the companies involved) but overall I find it tends to have the best data for sites I’ve worked on.
One tip is to only look on a ‘per domain’ basis initially so make sure to tick ‘one link per domain’ on AHREFS and to only look at DOFOLLOW links as the NOFOLLOW links won’t be causing you a penalty (unless you were very aggressive in your abusive comment SPAM which Matt Cutts recently addressed). Remove similar duplicates from your other list.
You can see above that even for a fairly large US government site the number of links to look at become relatively manageable, so most business sites will leave you with a very reasonable list to work through.
Step 2: Categorise the Sites, Links and Pages
This step is where the power of my strategy to save your time really comes in. To do this well, every potentially bad link needs to be examined by someone with enough experience to make sure nothing is missed. However, those people are very expensive (either in time lost or real costs) to almost any business or SEO agency. This step allows an intern or outsourced team to quickly turn your raw list of URLs into useful data that your expert staff can quickly work through looking for the bad links.
I tend to use Google Docs for this stage so that multiple users can work on the list at the same time. For the largest sites, you may need to split your document up so that you don’t hit spreadsheet limits (or more likely the limits at which Google Docs spreadsheets work well).
Step 3: Analysis of The Categorised Sites
The beauty of having a trainee complete step 2 is that after the first couple of times through this process and having the opportunity to see which sites you have selected for removal they will be able to take over this final step.
Good links in general will appear natural, will be on sites that are relevant to the industry in question and won’t have over optimised anchor text. The headings I’ve chosen for the preparation stage give me a powerful way to identify ‘suspicious’ links for further inspection and, more importantly from a time perspective, a lot of links which are either obviously good or obviously bad with no need for me to visit the site.
Page Rank is far from the most important metric of link quality but any links that are now on pr0 or pr? domains are worthy of further investigation. A sudden loss of page rank can be a sign that the site is selling links, penalised, or deindexed.
What is the Domain?
This is included so we can quickly identify the usual ‘problem’ categories such as directories, article submission sites and social bookmarking sites that are all often the warning signs of a bad link.
What is the Site About?
Here your trainee or outsourced worker will record the general theme or topic of the site. I would look at most sites that have a questionable connection to the site you’re working on improving. A link from a music site isn’t always bad if your site sells sports wear but it is more likely to be than a sports site linking to you and is therefore worth taking a closer look at.
What Type of Page is the Link On?
Here we just note the page type – the most common are ‘blog post’, ‘news article’ and ‘links or resources page’ but others include ‘social share’, ‘bookmark’, ‘forum’, ‘user generated content’ etc.
What is the Page About?
Much as with site topic I use this to help me identify any potential problem areas. If your sports site is listed on a links page about Belgian holidays it’s almost certainly setting off some alarm bells.
Details About The Client Link
The final section of the spreadsheet helps you identify potential problems with the client link. A blog comment with exact match anchor text, for example, would be an obvious problem. The same would apply to blogroll links on irrelevant sites etc.
Step 4: Outreach For Removal
Once again, you now step back and hand over to your trainee, intern, our outsourced workers to find the contact details of the sites you’ve marked for removal. It’s important to be thorough here and record all your steps. If at any point in the future you’re unfortunate to experience a manual penalty, having all the details of the efforts you’ve been making a long time before the penalty to remove your bad links can only help to illustrate your new approach to running your site.
You don’t need an e-mail address that matches the site you’re working for – an agency address is fine – if someone wants to check up they can always approach the site to confirm, but you’ll find people rarely do. In the event you can get a company address though, most people find a slightly higher removal rate results so it’s worth doing.
Keep your e-mail direct, to the point and honest. There’s no point trying to bully people or be rude – remember that in some cases a previous employee or agent of the site you’re cleaning up now has placed some of these ‘bad’ links. I find something short like this works well:
I'm working on behalf of stevebrownlie.com to tidy up the link profile after receiving a Google penalty. In order to get back on track I'm trying to remove some links and would like to know if it is possible for you to remove [link] on [page] or nofollow it if you prefer to keep up the mention of my client.
Please let me know either way.
Follow up twice – about a week apart – and then mark the ones that don’t respond and simply include them in a disavow file to upload to Google webmaster tools. If you write sufficiently generic templates your team will only need to make slight changes to take into account the different sites and circumstances in which the links were created.
Over all this process has proved to be very fast, and effective, with a high number of links being taken down and sites seeing improved traffic after a period of time following a Penguin refresh. As always there are no guarantees and you need to be careful not to remove any links that are actually helping you but this process should help you see more clearly where the issues are even in a large link profile.
If you have any questions please shoot me a message on twitter using the reply button below and share this post so we can help clean up the ghosts of linkbuilding past!