Powering Modern SEO with Social Media Data
Trust Insights CEO Katie Robbert Shares New Keyword Research Strategy
Traditional SEO research has limitations that require more innovative methods of discovering keywords. One way is by leveraging subject matter expertise and audited social media data to identify which keywords and topics are truly relevant to your audience.
I spoke with Katie Robbert, the CEO of marketing analytics consulting firm Trust Insights. Through her extensive experience in SEO and data analytics, she built and grew multimillion dollar lines of business in the marketing technology, pharmaceutical, and healthcare industries. Katie and her team partnered with Talkwalker, a social media analytics and monitoring platform, to develop a white paper on using social media data to power modern SEO.
According to Katie, many people these days are just “putting in the bucket of basic SEO research” in the form of keyword planners and tools such as Ahrefs and Semrush. This approach, while effective, limits the user to just search terms that their target audience might be using, instead of reflecting the actual language and conversations they’re having online.
Using Talkwalker’s social media export capabilities, the team obtained conversational, qualitative data sets from different social media channels to figure out the common terms and phrases that people were actually saying – and this got them very different results from what their straight keyword research yielded.
Through a machine learning process called topic modeling, Katie and her team processed the conversational data, figuring out commonalities and similar terms people were using for certain topics (for instance, “latte with extra foam” or “espresso” whenever people would talk about coffee). Katie shares that while there’s plenty of open source code for topic modeling, you’d have to do some coding yourself to have a good topic modeling tool – which was what Trust Insights did, when they developed their own proprietary version.
On Reddit Data, Difficult Keywords, and Google Passages
A great example of this approach is processing conversational data from Reddit forums. This is possible through an API, which serves as a tunnel linking the place you want to extract data from to the tool you’re using to extract data. “You can build the API to connect the two systems and say, ‘I want to extract these things,’ explains Katie.
Competition for a particular keyword determines whether it’s a realistic target for your SEO or if it’s too difficult to rank for it. “Drawing the line is really sort of a judgment call in terms of how hard you want to fight for that particular keyword,” says Katie. “And in order to fight for it, you need to be creating a lot of content around that keyword, a lot of authoritative content.”
Her advice? “Look at the organic keywords versus the branded keywords, because you should be ranking for the branded keywords for your own business, but the organic keywords might be things you might not be thinking about.”
Google Passages and Page Quality Guidelines
In 2020, Google announced a new feature called passage indexing. As Katie explains: “If you Google for marketing over coffee co hosts, then you’re going to get a little passage from the Marketing Over Coffee website, which is going to explain who the co-hosts are.” While this can provide users with immediate, contextually relevant answers to their queries, it can also take traffic away from websites, as users would no longer need to navigate to these pages to find what they’re looking for.
Another important consideration for SEO would be Google’s Page Quality Guidelines for Reviewers, a 175-page manual given to Google’s AI for training data. Humans are training the algorithm to look for pages and return results based on scores under specific criteria. “It’s humans who are rating the quality based on the guidelines that Google has given,” says Katie.
To better understand Katie’s findings, listen to our full conversation, which also features many suggestions on how to improve the quality ratings of your pages.
And to support this podcast, rate and review us on Apple Podcasts here.