Don’t miss out on metadata opportunities

Suresh says: “It’s all about the metadata: the metadata of the website, web pages, and so on. That is where we need to be implementing schema.

Back in 2011, metadata was almost entirely the meta description, but there is a lot more to it these days. People are pumping a lot of data into their web pages. Even in videos and images, we have metadata. An image will have EXIF data and for a video, a lot of data will be there also.

When it comes to the data on the webpage itself, that is where we can use schema (Schema.org) and the JSON-LD code to help search engines to understand exactly what the page is about. That includes the topic we are talking about, the entities that are present, and a lot of other things as well.

What are the key elements of metadata in 2023 that SEOs don’t know enough about?

“When SEOs do focus on the metadata, they often rely on automatic generators to make use of something like JSON-LD. These generators can be wrong, and that means that SEOs are not getting the maximum that they can from their metadata. Once you start learning the JSON-LD code and how it works - and you can expand on what is already there in the code - then you can make the most of it.

I have carried out single variable testing on it, and it works. I am finding a 90% success rate. That’s exactly what I do full-time: I consult on schema, and I go through all of this, along with technical SEO consulting. That means that I see exactly what happens when we deploy custom schema on a client’s website. Within days, or sometimes weeks, we see a lot of big movements in the rankings.”

What makes JSON-LD so important?

“Google uses JSON-LD, and they recommend JSON-LD. It is the easiest way to get the data from a web page. We should keep JSON-LD in our heads, as it’s the simplest code.

Some people will try and hide their JSON-LD code somewhere (like on the footer) because they think it will slow down the speed of the site. It doesn’t work like that. JSON-LD does not affect PageSpeed at all.

By using JSON-LD, we are just saying exactly what different pieces are, like the header, the title, or exactly what topic we are talking about. We can also use entities from places like Wikipedia and Wikidata as a kind of vocabulary.

For example, if you are from London, you can find the entities for London. You can take them from URLs like Wikipedia and Wikidata and put those into that code block. That way, Google can easily understand (for something like a Knowledge Graph) that you live in London. It works in the same way for entities about your job. Search Engine Optimisation is a topic on Wikipedia, so you can use that as well.”

What is the easiest way to implement JSON-LD?

“As of now, if you have no idea about what it is, then the easiest way to implement it is to pick up some generators. Start filling out the details, get the code, and implement it in the site headers.

While implementing, make sure that you implement one code for each page. That’s how it should be. People just put it in the headers, and it will be the same code all over the place. I’ve seen a lot of cases where homepage schema will be present on inner pages. That’s bad because you are then confusing Google about which the homepage is.”

Can this easily be incorporated into most CMSs? Is there a plugin to do this for something like WordPress?

“Yes, you can do that as well. If you have no idea how to implement it, you can always reach out to a freelance platform, like Fiverr, and find someone who can help. You can also find help through support channels.

Something that I wouldn’t recommend is GTM. Google Tag Manager has to load, it has to deploy the JSON-LD code, and it has a very long rendering time. You don’t want that to be happening for Googlebot.”

What are some of the key schema markup opportunities that SEOs are still missing out on?

“One of the most important ones that I have seen recently is the Clip operator for video. Actually, Google has recently created their own documentation on video Clip data. With video, you want to try and find any juicy parts that are particularly helpful. When you have the answer to a question in a YouTube video that you are embedding, you can identify the key clip with a timestamp, and it will link to the exact part of the video that is relevant. We can add this to the schema as well.

The schema type will be the VideoObject, and the Clip will be the structured data that tells Google what key moment to look for within that VideoObject. By adding this Clip information to the schema on your website, you will get more featured snippets.

That same Clip schema can be used for audio clips as well, which is something else that SEOs are not doing. You can implement this for audio-only content, like an mp3 on the page. Most SEOs are not even upgrading their schema codes - they just leave it as it is. We have to keep upgrading and implementing the new things that Google is offering.”

How does schema help users as well as search engines?

“Every time there is an update, users often have a hard time finding relevant results. People have to browse to the second or third page of Google to find the relevant information that they need.

The more information we can give in the metadata, the better Google can figure out and process that information. Then, Google can get the relevant results to the top of the SERP, visitors can click those URLs and come to the website, and we can obtain those high-intent visitors. That’s exactly what we wanted. It’s a win-win scenario. Now, Google gets the high-intent visitors, we get the potential customers or readers, and the user finds what they are looking for.”

Is it easy to measure the amount of increased traffic or brand exposure and ROI that you get as a result of implementing this kind of metadata?

“There was an update to Google Search Console recently, and now this kind of improvement data can be seen there directly.

Once we implement the code for video schema, for example, the thumbnails will appear in the search results, and you can use a third party to find the attribution for that. I like to use HockeyStack to track where visitors have come from so that I can get that attribution when they sign up for something or make a purchase.

You can use Google Analytics, which is also fine, but the data is skewed, which is why I prefer third-party software.

Measuring things like CTR is simple. If you implement schema for something, you can split test between a couple of pages to see if it has an impact or not. After it has been implemented, just make sure that you have reindexed the same page. People will often forget to do that. The schema has to be indexed so that the information can be processed, and you can get the results.”

When it comes to split testing, what are you testing and what are you looking for?

“The first thing we are looking for in split testing is getting more CTR on the search results. The second thing is the Time on Page, which is always a goal. If we have a video that appears on the SERP, people have to visit the page that the video was on so that they get what they want.

It’s a good signal because the more time they spend on the page, the more likely it is that they are a high-intent visitor. It’s also a good signal for the Google algorithm, to say that these visitors have spent this kind of time on these pages.”

Is this something that SEOs have to do after the fact, or can content creators define where to utilise schema within their content?

“We have to do it after the content has been created; It’s our job to work on this. Content creators handle the content part, but they can’t actually work on the code and all of that. The SEO or web developer has to be there to make the most of it, which is something that you can also automate.

In WordPress, there is a plugin called ACF, which is a Custom Fields plugin. We can also use that software for text tags on other kinds of websites. Based on your requirements and whatever you already have, you can still make use of schema.

SEOs should know what they want so that it’s as easy as possible to convey to a developer. You should make a template for pages that will have the same kind of schema so that a developer can easily deploy them using those variables.

SEOs have to be present in every stage of deployment, including the creation of strategy and advising content creators on what to include.”

What shouldn’t SEOs be doing in 2023? What’s seductive in terms of time, but ultimately counterproductive?

“There are too many tools out there and SEOs are often missing out on the Google Search Console data. They will see things like keyword gaps and content gaps, but they don’t see something called “intent gaps”. Each keyword has its own intent gap. You have to see the data from Google Search Console, and you have to implement it, but everyone sticks with their Ahrefs or Semrush data.

I have tested this using different intent keywords on my test sites. It is a game changer for me: the KGR, or Keyword Golden Ratio. Look at the intent and find similar keywords where no one is competing, so that you can make the most of it. This has to be done manually using Search Console and regular expressions.

When you have skin in the game, you will be able to identify that intent. Once you understand the business, you understand the intent far better. You can also serve more than one intent on the same page using subtopics.

SEOs should not be relying so much on their tools and should focus more on Google Search Console data. You have to make the most out of it because you don’t get it anywhere else.

You can take that data and put it into somewhere like Bitquery so that you can understand things like queries, number of clicks, impressions, etc. - or you can just use the simplest, which is Google Search Console.”

Suresh Kumar is Technical SEO Consultant at Gondi Media and you can find him over at sureshkumar.digital.