Entity Search: The Past And The Future [An Overview]

Table of contents

1. History overview of entity search
2. Quintillion bytes of data – is data quantity enough to understand us?
3. The (future) savior: knowledge-first approach
4. What the future holds for entity search?

History Overview Of Entity Search

The topic of schema markup is not new. We pioneered these concepts in the SEO industry even before they became popular buzzwords and known tactics that actually bring value to websites’ SEO performance. We have been using structured data in the past 20 years to describe data and make it more comprehensible across different datasets and systems, so that both humans and machines can process it properly and actually make something useful out of it.

Along the way, we made sure that our readers understood the use of entities and entity disambiguation, since both of these concepts form the basis of our core products and we are definitely fond of the idea to be the visionaries in the industry and lead in the era of entity oriented search. We cared about entity identifiers, entity relationships and entity attributes and it seems that as the time passes, all of these ideas, slowly over time, will become more and more adopted among SEOs until they become mainstream. 

From a scientific point of view, the core problem when dealing with entities in the past 10 years was solving for entity ranking for various tasks, especially in the process of entity linking. You would have a search query that is entity rich, you would employ a retrieval method to obtain entities in the process and then perform entity ranking to form the final search results page no matter whether we speak about a platform based search like an e-commerce shop or news website or a big software like Google’s.

In the last decade, search engine optimization managed to progress to semantic search, linked data and question answering over linked data. This eventually led to ontology engineering, taxonomy optimization and knowledge engineering even though there are not many online resources to prove that this is the case. We know from practice that those companies who seriously get SEO, absolutely invest in these approaches (actually, investing more than just 6 figures on a monthly level).

Quintillion Bytes Of Data – Is Data Quantity Enough To Understand Us?

Roughly 2.5 quintillion bytes of data is created every day. When people speak about it, some of them believe that the sheer volume of available data means we have all the resources to provide an answer to a question without delay. If you can’t, they declare, the solution is to get more data. 

The problem with more data, however, is that it does not necessarily mean that the search engines responsible for providing those answers can process them. If the data is not provided in a structured way, it results in a lot of computational effort and cost to understand the context behind certain user queries. What looks like a relatively simple process is actually a solution that is far from trivial and requires some scaffolding and structuring to function optimally.

The (Future) Savior: Knowledge-First Approach

Knowledge-first approach is more user-centric, relationship-oriented and requires more context when solving problems. We ask ourselves the following questions:

1. Who’s the end user who will consume the data?
2. How is this data related to our systems and our people who work on them?
3. Why do we need to take care of data consumption and data activation in order to be more productive?

It is clear that in order to form a successful SEO strategy, your teams need to be versatile both in the data world but have the business acumen as well. That is why we predict that the next 10 years will include more collaboration between the data and the business teams and these hybrid positions are already in place in some of the bigger companies worldwide.

And not just that. Knowledge graph analysis and knowledge graph scraping will become the foundation on creating intelligent long-tail SEO strategies through finding similar entities in the knowledge graphs themselves.

We are proud that we managed to publish more than 12.5 million entities to the Linked Open Data Cloud through our products and are still continuing with our work, making the Internet better, one day at a time.

What The Future Holds For Entity Search?

The core focus in the past 10 years in entity search was focusing on retrieval models and more specifically on entity representations. Richer query annotations were taken as they are and a bit for granted. In the future, we definitely need to focus more on understanding user interactions, information needs, data sources and novel retrieval methods. Those who want to advance, will exploit the power of knowledge bases who enable the modern features of entity search.

That is why the key questions here are:

1. How to enable humans to maintain and expand knowledge bases at scale?
2. How to improve the entity annotation process so that we can increase the process of entity retrieval?
3. How to provide direct entity summaries to properly answer demanding user needs?

This leads us towards a zero-query search world, where search engines will need to be proactive in order to solve for reactive search.

And not just that: we definitely expect to see mashing up personal, enterprise and public data, enabling multiple views of the same data and gaining the ability to actually edit the data that we see.

Entity search in future will be dependent on the idea of being able to “query” the Internet through SPARQL queries which will use HTML document URLs as Data Source Names. Cross-platform data manipulation will also become a thing. In addition, content hubs and content hub SEO of the future that will rely on graph embeddings and clustering when creating content plans. Everything will simply go in the direction of exploiting the power of intelligent knowledge graphs and knowledge graph embeddings.

However, bear in mind that anticipating user needs in a constantly evolving world is not easy. Luckily, we at WordLift have the technology and the scientific know-how to tackle these challenges in a more structured way. The key for you to solve these ever-changing user needs is to power your content creation and maintenance process with a knowledge graph strategy where all your entities can be described and explained to search engines. We are here for what comes next and ready to work together to solve pressing user-needs.

Are you?