Artificial Intelligence

How content creators can prepare for AI

Thomas Bosshard
  • Thomas Bosshard

Categories

  • Artificial Intelligence

Search engines are becoming increasingly sophisticated. There is also a lot of advanced artificial intelligence (AI) behind personalization, i.e. recommending content to specific users or user groups. Both serve the same purpose: to deliver the right information to the right user at the right time. This goal has also been at the center of many recent changes to search engine algorithms.

Keeping up with major changes to algorithms can be challenging but is also an absolute must if you don’t want to risk losing visibility on the result pages these engines generate. Google changes its search algorithm frequently. To remain one step ahead of spammers, these changes are well-kept secrets. Nevertheless, specialized online agencies such as Moz or serpIQ draw conclusions about ranking changes by systematically studying how content performs. Their research shows that search engines no longer focus on the particular words a user enters but look at the whole query as well as the context of the query such as a user’s location or previous history. Ever since these major changes were introduced to search engine algorithms, optimizing content is no longer about inserting as many key words as possible into a page. Content must match the “meaning” of user queries.

In a nutshell, content creators should know that most modern search or recommendation engines determine:

  • how appropriate a piece of content is for a user’s search query (SEO jargon: “relevance”)
  • how trustworthy that content is (SEO jargon: “authority”)

The following basic rules can help you accommodate these goals in your content.

1. Focus on topics and user intent

An algorithm must decide whether your content meets the needs of a user or not. If it doesn’t, it will leave your website and consult others until it finds one that optimally matches the type of content the user is looking for. An important step is identifying the intent of the query – or the reason why a user is looking for something. Does he want to check some facts? Or compare some offers before purchasing a product?

Based on lots of data from search engines, user intent can be classified into three categories: informational, navigational and transactional.

  • Checking for facts or figures on persons, events or locations shows an informational intent.
  • Searches performed to locate a specific website like an e-banking or official government website are classified as navigational queries.
  • Transactional intent means the user wants to get something done. Trigger words such as brand names or trying to find reviews help the algorithm infer a purchasing intent.

User intent obviously varies from topic to topic; therefore, topics also need to be classified. Considering the amount of content published on the internet, this sounds like an almost impossible feat – at least for a human being. The same applies to users’ search queries which range from common everyday queries to the very specific ones. That’s why automatic language processing is so promising – if the algorithms learn to get it right. Semantic search is still in its early stages and its recommendations still depend heavily on content and any contextual or behavioral user data it can get.

What does this mean for content creators? Firstly, they should decide which topic their online presence should be dedicated to. Then they should cover that topic thoroughly in order for the search bot to consider it a valid resource and categorize it correctly. Secondly, content creators should use “high-intent” keywords (and get rid of ambiguous ones). For this, content creators need to identify user profiles or focus on personas and then determine the reasons why these user groups or customer segments should visit their website or blog. Keywords should be designed to attract this type of visitor to your website. In SEO jargon, this is called “attracting qualified traffic”.

2. If content is king, then context is queen

Everyone leaves some sort of footprint online. Algorithms can interpret this “behavioral data”. For example, previous searches or subscriptions can reveal where some of the user’s interests lie. Location and demographic user data can help the engine rule out irrelevant information. Transactional queries can be narrowed down to a stage on the customer journey. The algorithm then takes all this information into account to provide only content that is only relevant to this particular stage.

These are only a few examples for why context matters. By providing information for different stages of the customer journey – or for people of different ages and from different places – content creators can increase their visibility as well as the number of qualified traffic visiting their online content. It can also help predict what a person will do next.

Context can also prevent your content marketing efforts to be perceived as spam-like “shots in the dark”. Offering personalized, targeted context can lift that darkness and shed a favorable light on your business or brand.

3. Length and updates count

Text length matters. Since the average attention span has never been shorter than in today’s age of social media, this may seem somewhat counterintuitive. But even though not many people read long articles from top to bottom, longer articles are usually also well researched and more effort has gone into writing them as opposed to short ones. Evidence from the company serpIQ which studied of 160,000 web pages – so quite a representative sample – suggests that long articles fare much better in rankings than short ones. In fact, length seemed to almost perfectly correlate with higher SERP ranks. Regarding social shares and backlinks – two important metrics that help establish a site’s “authority” – blogposts that exceed 2100 words were most successful, according to Hubspot. This wordcount corresponds to about 4.5 pages in a Word document. This blogpost would therefore be way too short (it counts only 1300 words).

Search engines also reward “fresh” as opposed to “stale” content. This does not mean that frequent updates to pages makes them automatically more authoritative and they are thus favored. According to Moz, Google’s concept of freshness also takes the number of pages added to websites into account. Also, external factors such as “link growth”, i.e. the number of links pointing to your articles from third-party websites, play a role. However, any attempts at manipulative link building can backfire. Only “natural” link growth indicates that a website or blog provides good and relevant content that matches users’ information needs.

Freshness also depends on the topic. Apparently, Google acknowledges that old content can also be good content. Therefore, it must differentiate between the nature of sites. News sites that constantly churn out new articles are obviously treated differently to websites on topics that are not so quickly out of date.

4. Tone of voice should be more conversational

Choosing a tone of voice obviously has quite an influence on a company’s online brand perception, and finding a unique tone of voice as well as applying this tone consistently across all content channels can be a tall order. Although not every brand is well served with a “quirky” or “passionate” tone of voice, a conversational tone can engage users more easily than content written in formal or even academic language. Formal language as well as jargon act like a deterrent.

Many “intelligent” devices such as Siri already engage users in voice-powered conversations or chats. As these bots are on the rise, so is spoken language and messaging which both tend to be very informal. Examples of chatbot responses used by ecommerce sites show that even high-class brands are becoming very “chatty” indeed as well as capable emoji-users. It also helps a brand seem more human in the eyes of its customers, especially with younger generations that are used to communicating online.

5. Evaluate your efforts

Lots of SEO expertise is self-taught. Although much like a trial and error procedure, content creators can check what effect their improvements have on rankings and adjust their efforts. This can help them learn more about what counts.

Download infographic

Get in contact – our office in Zurich.

Adviscent AG
Binzstrasse 23
8045 Zurich
Switzerland

+41 43 344 91 89
company@adviscent.com

Free
Proof-
of-Value

Schedule a chat

Contact us to discuss and plan a proof-of-value.