On-Topic Search Engines

on-topic searches

Understanding On-Topic Searches

You probably know that

  1. Minerazzi generates on-topic search engines.
  2. on-topic searching is a form of vertical searching.
  3. vertical searching is not the same as horizontal searching.

The purpose of this article is to review the differences between all these technologies.

From Horizontal to Vertical Search Engines

Search engines like Google and Bing index large portions of the Web. Because their indexes include a wide spectrum of knowledge domains, these are known as "horizontal" search engines.

These types of search engines tend to return less precise and off-topic results as users move down their search result pages. However, users can refine their searches by adding new terms or using specific search modes and advanced search options.

By contrast, specialty search engines focus on a particular knowledge domain. Since their index cover specific products, services, or industry sectors, these are called "vertical" search engines. Users can refine their searches through filters and options.

Vertical searching reduces to enterprise searching when the focus of the knowledge domain is an enterprise. Moreover, vertical searching reduces to site searching when the index queried consists entirely of documents from a given web site.

Over the years, horizontal search engines have incorporated vertical search capabilities to their services through specific search commands, by segmenting their indexes, or by focusing their results on a given topic. Google site search command and Google Scholar (or News, Images, Video...) are typical examples of this.

Commercial search engines have also blended horizontal and vertical search results. Google's "Universal Searches" is a good example of this.

From Vertical to On-Topic Search Engines

There is a subclass of vertical search engines, called on-topic search engines, that index documents about a given topic. Because these can be designed to cover several related subtopics, or have query reformulation features used by horizontal search engines, these are viewed as hybrids of the above two search technologies.

Not all vertical search engines are on-topic, though. Let us address this point.

The index of an on-topic search engine is very small, almost spam-free, and consists of documents about a common topic. By "very small", we mean that index collections can be made to consist of less than 1,000, 10,000, or 100,000 records—to mention a number. These are collected from trusted sources and examined to remove off-topic or spam records.

Since all records must be somehow relevant to a common topic, a search in an on-topic index tends to uncovers co-occurring related terms. Such a micro-index does not have the same precision-recall issues present in horizontal or vertical search engines, nor the indexing need to be full-text. This is a unique scenario that is not exactly present in vertical or horizontal search engines.

From Horizontal to Vertical Searchers

Users tend to search with the default search mode. In most horizontal and vertical search engines, this mode is AND. Regardless of this, horizontal and vertical searchers are two different 'animals'.

Horizontal searchers can refine their searches by adding new terms. However, vertical searchers quickly realize that this search strategy frequently causes "not found" responses because of the small index being queried. They can do better by using the filters and options that come with the vertical.

Another characteristic that differentiates both 'animals' is that vertical searchers are more self-qualified users than horizontal searchers. That is, before searching, they already know that their queries must be somehow related with the type of search engine they are querying.

Vertical searchers also tend to search to satisfy their immediate realities or most pressing information needs concerning must-have tangibles, to-do events, to-do tasks, must-knows, and so forth. For instance, users querying a price comparison or travel search engine are fully aware of the intention of their searches. They are not likely to submit off-topic, arbitrary, or irrational search terms.

All of this has a marketing angle: As consumers, vertical searchers must be targeted differently. Users querying a vertical about law are more likely to buy products or services relevant to the legal domain. To target these users, marketing creatives to be displayed in said vertical should be on-topic, too.

From Searching to Mining

Wikipedia has detailed articles about horizontal and vertical search engines (1, 2). At the time of writing, the search experience with these technologies reduces to staring at a list of search results. In a sense, this is a technology waste.

That is why we developed Minerazzi, the first search engine platform that allows users to interact with search result pages. With Minerazzi, users can mine search results, extract information, and do something with that information. Instead of merely searching, they become data miners.


  1. Wikipedia, (Accessed on 9-9-2013). Web search engine.
  2. Wikipedia, (Accessed on 9-9-2013). Vertical search.