A POV by Nina Hale, CEO of Nina Hale, Inc. about Natural Language Processing
Last week I wrote a POV on the importance (or lack thereof) of social media on SEO rankings. In it, I mentioned natural language processing, and a lot of people have been asking for more detail.
Natural language processing (NLP) is a 60+ year old field of research that applies computer science, artificial intelligence and linguistics to the human-computer interface, allowing computers to understand human or natural language. Recently, NLP has become vastly improved and more effective due to the application of statistical algorithms and the availability of big data.
There are MANY applications of NLP in digital marketing, the most relevant of which are:
- Question answering.
- Sentiment analysis.
- Discourse and Topic (understanding what type of question has been asked).
- Does it need a yes/no answer, a content element, a fact?
- Are you asking what color the sky is, or what goes into a nuclear bomb?
- With topics, what is the gist?
- The major elements of translating speech and languages (contextual conversation flow; understanding explicit, implicit, and multiple meanings of words and sentences).
Natural language processing is increasingly important to digital marketing and a connected world. With wearable devices on the rise and open answer applications such as Siri or “ok Google”, we’ve already seen an explosion in voice recognition and the ability of search engines to accurately respond to a complex set of questions. Often, those questions occur in a string of refined searches that, more and more, require an understanding of how the previous searches impact subsequent searches.
So What?
Natural language processing isn’t new in SEO — it’s been part of SEO since the beginning because search engines are predicated on answering a query. But in the past, most successful results have mimicked — as closely as possible — the original search query. In 1994, we had debates about whether to write the focus keyword phrase in a title tag as “masters degrees” because that’s how people searched, or as “master’s degrees” (note the apostrophe) to be more grammatically correct. We made big business out of learning the voice of the consumer and finding creative ways to say “used cars” when our clients were selling certified pre-owned vehicles. But now “semantic search” is rising, which means that Google can more clearly determine the meaning (and synonyms) behind a search. Finally, data hoarding all those billions of searches has paid off!
- We’re moving away from the individual keyword and more towards the page content. Search engines are better than ever at understanding the full meaning behind topics, and they understand synonyms and themes too, due in no small part to NLP. So, if you’re still talking about “keyword density” (how many times should the specific keyword be repeated), it’s time to move on.
- Combine page content — within reason. Follow-on searches will expand this year — Google now understands that a second search is related to the first. For example, if you ask “how tall is he?” after asking “how old is Barack Obama?” Google will relate the two searches. It’s hard to find a published example of this type of follow-on that works, but it will expand to the point of being a normal, everyday function. At some point we will be able to naturally combine content in a logical order on the same page.
Not So Fast!
Notice how I mentioned the voice of the consumer above? Well if you’ve been a good little SEO and followed Google’s rules all along, you sing to the hymn of “unique, relevant content.” That means listening to the consumer and speaking their familiar language (rather than a computer’s). You’ll generally get higher conversion rates, as well as better search traffic using the consumer’s language.
Even more importantly, Google and other search engines are able to determine a search query’s semantic meaning by pulling from many sources, thanks to heightened understanding of how the average person would naturally talk about that subject. This still means that a comprehensive content publishing strategy is vital for SEO. You must still follow the same rules when publishing your content and you must still get people to share, link to, and discuss your content on multiple websites.
Authority matters (probably). I’ll likely have to do an entire POV on this as well because it’s currently controversial. Just as there are better links – say a link from the New York Times is better than a link from Nina’s Backyard News — there are bigger, better authorities as well. People listen more closely to important people, and if they have Google+ profiles, and code their content with a digital signature (rel=author schema), their content gets more SEO love.
Next Steps
Do you need help with content identification, publishing strategies, or social authority? Nina Hale, Inc. takes a measured, ROI-driven approach to multi-faceted marketing strategies, helping companies grow valuable leads and eCommerce sales.
Sources:
http://blog.crazyegg.com/2013/12/13/semantic-search-seo/
http://searchenginewatch.com/article/2325063/6-SEO-Predictions-for-2014
