How to speak intelligently about Knowledge Graphs: a Primer for PPC Professionals

As a paid search professional, I spend a lot of time staring at one SERP or another. Of course, even if I wasn’t in the business, like millions of others, I’d still be looking at a whole lot of SERPs because that’s how we find information these days. Chances are, you do too. And if you do, I’m sure you’ve noticed “rich snippets” popping up here and there when you ask Google common questions. These snippets draw data from something called a Knowledge Graph. What is that? Well, let’s ask Google:

What is Google's Knowledge Graph?

They said it.


That is an instance of Google using a well-defined question with authoritative answers to give me the answer I was looking for instead of making me go look for it with one of the links it provides.
Google also provides a lot of common information about companies and brands. Type in “Verizon” and you’ll see something like this:

From the Google Search Results Page

From the Google Search Results Page

You can break these types of results into several pieces – the map, the customer service and sales phone numbers, general information about the company, and links to various social network presences. For more information on adding Social Network information, see our very own Sean Van Guilder’s great post on the topic. To find this information, Google draws on a variety of “trusted” sources. These include wikipedia, wikidata and up until recently, Freebase. Freebase used to be considered a sure thing to getting information into Google’s Knowledge Graph, as it was owned and operated by Google itself. However, Freebase was recently absorbed into wikidata, which is run separately and has stricter rules on what can be added commercially.

Currently, there are several recommended best practices for getting information into the KG and increasing the likelihood that data will show up on the SERP:

1. Set up a G+ page. Make sure that it has current information about the company and links back to the official company website.
2. Make sure Wikipedia and Wikidata have accurate and current data about the company. Make sure not to include any “salesy” information in the company profile, as it will almost certainly be rejected by Wiki editors. Run afoul of this too many times, and risk losing editing privileges and gaining very watchful eyes to all future edits.
3. With the passing of Freebase, the most important thing you can do now is to make sure that the information you want available is provided on the official company website in a well-structured manner. The best way to do this is with JSON-LD , a simple, popular markup format strongly recommended by Google. By doing this you can provide important company information in a format easily accessible by Google’s robots. It’s also worth reviewing “Schema”, which is a “joint effort, in the spirit of, to improve the web by creating a structured data markup schema supported by major search engines. On-page markup helps search engines understand the information on web pages and provide richer search results.”

For more information on how to add these snippets to your website, please refer to Google’s excellent documentation. Here you can find information on how to customize your knowledge graph, including contact information (phone numbers) and logos. They’ve also recently opened up their tests into providing videos in the SERP.

With all of this information available without ever having to go to the actual websites, it has gotten some SEO types slightly antsy. If Google starts answering questions on it’s SERP without the searcher having to go to their website, then why would they ever go to the website itself?

But for us paid search types? Well, as best as I can tell, this is really, really good news for us. See, I don’t just want clicks. I want clicks that match a specific type of intent, and capturing those people without sucking in a bunch of other traffic – traffic my clients have to pay for – is a balance I’m constantly trying to strike. I see people who search for keywords I’m bidding falling into two categories:


Bad Traffic – “Bad” Intent

Good Traffic

Good Traffic – Desirable Intent

The types of questions that Google is drawing on it’s Knowledge Graph to answer are the types of questions I’m fine with them answering on the SERP. For instance, a common problem paid search strategists encounter are people clicking on ads looking for the number for customer support. Every time someone clicks on your ad looking for a number to pay their bill or make a complaint, that’s money spent on something which will never see a return – and money that can’t go towards “Good” intent. Of course, comprehensive negative coverage is a necessary strategy to keep your ad from showing up in these cases, but you can only go so far with that without missing out on legitimate traffic. If the customer support number is displayed clearly on the SERP, it lessens the chance that person will click on your ad looking for it.

A picture of an existing customer looking for a customer support number via a paid search ad.

An existing customer looking for a customer support number inside a paid search ad.

Consider Rich Snippets drawn from Google’ Knowledge Graph a steel cage which helps keep users with bad intent from clicking on your ad. For help getting information about your company into a format easily digestible by Google’s robots, along with a wide range of other optimization services, you can view more information about our SEO offerings.

Good luck! If you have any questions, please leave a comment below and I’ll do my best to answer them. Until then, may your paid search efforts be excellent and full of desirable intent.

A user with good intent clicking on your paid search ad.

A user with good intent clicking on your paid search ad.

Point It About the author
No Comments

Leave a Comment: