The Human Difference: Data and the Power of Semantics

As search becomes increasingly complex and powerful, the fundamental question remains: will an algorithm ever be as good as turning to one’s left and asking a random person the same question? As Google continues to make radical improvements in its search capability, other search engines, computer scientists, and search providers are turning to a more human solution over algorithms.

Natural Language and Search

As technology and the internet have not only become more ever-present in our lives but also more accessible to people of all ability levels and backgrounds, search engines have adapted by investing in understanding something called “semantic search”.

Put plainly, semantic search allows search engines to move away from traditional keyword-centric search algorithms, and instead focus on how people naturally speak and/or ask each other questions. It’s pretty complex, but the results have been the building blocks of what has made Google the dominant search provider online for the past decade. Moreover, it was one of the focuses of the Hummingbird update.

What does this mean, in practical terms? Ten or fifteen years ago, to find out how tall a celebrity was, the query might have looked like angelina jolie height. By contrast, now you can type how tall is angelina jolie and not only will you get the same if not better answers, but also Google may provide an answer in-search. In fact, not only may Google provide an inline answer 25% of the time or more, but it’s often correct. So correct, sometimes, that it also includes a sly joke in the answer it provides.

Where People Fit In

Though technology has improved considerably, there’s still no algorithmic match for what a person can do.

In a long guest post for TechCrunch, Brainly CEO Michal Borkowski makes the argument that people are integral to information discovery, in large part because of the biases of algorithms. Search algorithms, in particular, are weighted towards “freshness,” with new content being favored (fairly or unfairly) over older and/or more established sources.

Moreover, Borkowski argues that natural language search isn’t at the point where it replaces good Q&A between actual human beings. More fully, he is saying that in a world where people are overwhelmed by data, bringing humans back into the search and discovery process reintroduces information into search. Where data is numbers and results without context, information is data with knowledge — context, meaning, etc. — applied to it, making it real and actionable.

It’s not just Q&A that’s using real, actual human beings to make better decisions about discovery. Facebook uses human testers who identify news feed items that are interesting or that are not. Their behaviors — which are closely monitored and questioned — are folded into the algorithm in increasingly complex ways. It’s proof that pattern detection on a programmatic level simply isn’t enough; a human touch is key.

Tying It Back to Listings

Much like simple search and algorithmic discovery, local listings are better when an actual human being is managing them. The key is understanding that Yext, which merely disseminates the correct, human-touched information to publishers in a controlled manner, differs from the algorithmic approach that drives a purely aggregator strategy.

We’ve blogged before about how the local data ecosystem moves data through the environment. Every publisher — aggregators included — scrape, crawl, gather, solicit, purchase, and acquire information on businesses by any means necessary. These then enter a “pure-math” filtering process that creates winners and losers. Winners become the information displayed on the directory or publisher — at least, until the filtering begins anew.

Yext removes your business listing from that algorithm, and instead allows you to exert a human (and, thus, arguably superior) touch on what is otherwise an purely mechanical process. The only programmatic element to Yext is the method by which it takes over a listing; the information that gets sent and is guaranteed to display is only the information entered by a real person — that is, by a professional marketer or the business owner themselves.

A human solution is increasingly being sought as the answer to complex technical questions. Listings and digital presence management are a perfect space where leveraging technical solutions with a human touch yields the best results. We’ve seen it before with Google Now and the Facebook News Feed algorithm tweaks, and we shall see it again as more and more local marketers turn to Yext for their digital presence management needs. As a Yext Certified Partner, you’re ahead of the curve — probably because you’re smarter than the algorithm, search or otherwise.