Sign up for our daily recaps of the ever-changing search marketing landscape.
How to use AI for link building and improve your search rankings
As artificial intelligence (AI) continues to grow in popularity among consumers, marketers working in search engine optimization must consider how it can make their work easier and help them get better results.
The use of artificial intelligence (AI) has permeated the mainstream, as companies like Apple, Microsoft, Amazon and Google work diligently to make voice commands and voice search integrative parts of their OS and user experience.
AI’s applications in the search engine optimization (SEO) world are continuing to expand to new horizons. Many have likely heard of the Y Combinator-backed RankScience, which uses thousands of A/B tests to determine how best to positively influence search engine rankings. Complete handling of SEO from AI is likely not going to catch on quickly, however, and there isn’t software currently that handles many of the important aspects of SEO.
While no software exists that leverages AI to build links, we can nonetheless use multiple types of software for various stages of the link building process, including:
- data collection, by using natural language processing to determine if the sites are contextually relevant and keyword relevant.
- site analysis, to determine if the site in question will predictably have an impact on ranking.
This means that AI can be used to augment, automate or automatize processes. Link building specifically can’t be a fully autonomous process, but we can leverage AI to augment human processes to increase efficiency of finding bloggers/influencers and improve the quality of sites that we approach for links. Below is the process for leveraging existing AI in link-building campaigns.
1. Prospect new sites
First, you need to look at websites as a whole, which likely have multiple contributors or people on staff. These can be good link-building opportunities through sponsored or contributed content.
This can be done by finding industry publications or other informative sites that appeal to your target audience. Try searching by industry keyword plus terms like “magazine,” “journal,” “digest,” “newsletter” or “information” in Google to find possible fits.
Things to look at include:
- publishing frequency: Are they actively publishing new content on the site?
- last publish time: Have they had new content up in the last month?
- user experience and design: Is the design up-to-date and easy to use? (You don’t want your content or links to be affiliated with a spammy or outdated website.)
If these sites look like they are a good fit, add them to a potential candidate spreadsheet, with notes and contact information. To keep this going on a rolling basis, consider using Google Alerts to be notified when these types of searches have new results.
2. Identify blogs and influencers
The next target for link building is more specific influencers and their blogs, which are usually smaller in manpower than publications but might have larger networks.
To find influencers and blog content that hit your target market, you can use text-processing analytics like Watson Analytics, which can be used to search for your industry terms in the context of natural language (i.e., the way people talk online).
For instance, someone might not always say, “I am interested in polymer manufacturing,” online, but using AI tools that can predict related text patterns and speech, you might be able to find more influencers who haven’t directly used the terms you’re looking for.
Software that makes connections between what you’re searching for and what related people are actually saying online can help you broaden your net for more link-building opportunities.
Things to look at include:
- comments and social shares on posts: Do the posts get a lot of engagement?
- last publish time and frequency: Are they actively publishing new content?
- user experience and design: Is the design up-to-date and easy to use?
- social platform: Do they have a large social media following on the platforms that are used by your industry? This can be beneficial when they share your content or content that has a link to your website.
- reputation: Sometimes, individual influencers or blogs might have a strong opinion about hot topics that you might not want to get involved with (like politics or religion). Make sure you know your targets and their voices before reaching out to them.
Once you’ve analyzed the possibilities and weeded out blogs or influencers that won’t work, add potential blogs and influencers to your potential candidate spreadsheet as mentioned above.
3. Analyzing the sites to determine impact
Now that you’ve compiled a list of publications, influencers and blogs, it’s time for the hard part: determining if they will have an impact on your ranking for your target keywords.
This is the powerful part of AI — the part that can improve the impact of the links. AI can process data from multiple sources to identify likely variables or variable clusters that correlate with ranking in Google.
AI can also identify patterns that you might not have even known existed (e.g., generating links from sites with a DA of 40+ if the site is about shoes, but with a DA of 20-30 if the site is about blue shoes).
You can use software to process this data (I use Rulex). But you’ll need to gather the data for the software to process. Most of this data exists through APIs for sites and tools like these:
- Link metrics: Moz, Majestic, or Ahrefs
- Technical optimization: BrightLocal, DeepCrawl, GTMetrix and the Google PageSpeed tool
Artificial intelligence is a priority for search engines and big technology companies, and voice assistants are getting more and more robust to help users control their online experience.
As AI continues to weave its way into connected apps (through the Internet of Things), smartphones and more, we SEOs need start thinking about how AI can make our work easier to help us get better results for our customers or business.
Opinions expressed in this article are those of the guest author and not necessarily Search Engine Land. Staff authors are listed here.