It often feels impossible to keep up with the swift pace of the digital world. From social media to the basic tenets of web design. best practices are constantly evolving. During this time, however, one constant has remained: a ‘typical’ Google search looks a lot like it did ten — even twenty — years ago.
Some things have changed, of course. Google’s algorithm, for example, no longer rewards keyword stuffing and instead focuses on content quality, and increasingly, user experience. Callout extensions and snippets enhance the search process, making it easier for users to quickly find the information they need — and allowing businesses to streamline the process to boost conversions.
Still, there’s no denying that the basic process of searching for keywords and browsing the search engine results page (SERP) resembles the familiar steps we took to find web content when Google was still new.
That’s all about to change. There is no such thing as permanence in cyberspace — and Google is certainly no exception. The very concept could soon undergo a huge transformation that will influence every aspect of our digital lives. We delve into the potential changes below, so keep reading to learn what you can expect from the future of search:
The New Proposal From Google Researchers: What You Need to Know
When Google first made its mark as a search engine, its success was made possible by a then-revolutionary strategy known as PageRank. Under this system, search results were obtained based on their relevance to user queries.
Now, PageRank is poised to give way to a whole new system of searching. In a recent proposal published by top researchers at Google, experts shared a radical idea, that, upon closer inspection, makes a lot of sense: instead of relying on keywords, searchers can ask questions. From there, an advanced language model will ensure that their queries are answered directly. These answers could be accompanied by citations, so searchers know where to seek additional information.
This could revolutionize every aspect of search. Consider the current Google experience: you type in a keyword and receive a list of resources that may potentially lead you to an answer for the question underscoring your search term. This may be helpful, but it still means that you need to complete the bulk of the research on your own.
The standard approach is especially problematic when your question can only be answered by drawing information from multiple resources. Unfortunately, the very nature of PageRank is to provide surface-level solutions, rather than diving in with a deeper analysis.
The answer may lie in new versions of a language model known as Generative Pre-trained Transformer 3 (GPT-3). This advanced solution is already able to draw information from a variety of resources. It harnesses the power of deep learning to create shockingly human-like text. In fact, many people struggle to distinguish between text produced by GPT-3 and phrasing drafted by talented human writers.
The potential role of GPT-3 in search is examined in depth in the recent report “Rethinking Search: Making Experts out of Dilettantes.” Published by prominent Google researchers such as Donald Metzler, this paper references the concept of the dilettante — someone who shows interest in a particular topic or activity, but only at the surface level. Google’s current indexing looks much like the classic dilettante in many respects, as it only provides the basics in response to user searches. With the right technology, the search engine could provide the depth that users crave.
In the future, the indexing aspect of search engines could be replaced by a practice known as information retrieval (IR). The paper proposes that future systems operate by “collapsing the indexing, retrieval, and ranking components of traditional IR systems into a single unified model.”
What Needs to Happen for Google’s New Search Engine System to Work?
While the proposal highlighted above is intriguing, it’s far from ready for implementation. At this point, it’s not a realistic option simply because too many issues accompany existing language models.
Resources (or a lack thereof) may be the top source of trouble, as new models will extend far past traditional indexing in terms of usage. As such, the University of Washington’s natural language processing expert tells MIT Technology Review, “Large language models…require large memory and computational resources.” She adds, “I don’t think they would replace indexing.” Despite this, she still believes that such models will play an important role in the future of search. We just might not see these changes right away.
In addition to extensive resource consumption, the proposed search system would be problematic from an accuracy standpoint if implemented as is. Unfortunately, GPT-3 is not yet able to keep track of sources or offer evidence to back up the answers it provides. As such, it’s possible for this language model to parrot back incorrect information.
Google’s Donald Metzler believes this problem can be addressed by simply training future GPT-3 solutions to track where their information originates. This might not be so simple, however, as the end of indexing as we know it will mean that URLs cannot be simply removed when a source is discredited.
Metzler’s ultimate goal: to combine the best of new and old systems to deliver far better results than are possible under the current indexing model. He believes that “classical information retrieval and large pre-trained language models can be synthesized and evolved into systems that truly deliver on the promise of expert advice.”
Major changes are already in the works for Google. Some of these will likely involve the search engine’s increasing reliance on artificial intelligence and machine learning. Other shifts may be driven by trends in voice search and other changes in how, exactly, people access or interact with search engines.
Change is tough in the digital marketing world, but it’s also inevitable. Instead of hoping that Google maintains the status quo, it’s time to embrace the opportunities that new search methodologies will bring. Take this time to determine how you can stand out in an evolving digital sphere. Continue to focus on providing high-quality content and an easy-to-navigate page, and you’ll have no trouble weathering whatever is in store for the future of search.
Embracing New SEO Opportunities with Vinci Digital
At Vinci Digital, we make a point of always remaining at the forefront of all things digital marketing. This means keeping a close eye on trends in search engine optimization. We’re prepared to guide you through the transition as Google adjusts its search algorithm or seeks even bigger changes with the help from advanced language models.
PS: How do you plan to address future changes in search engine functionality? Do you think proposed changes will enhance digital marketing or create new challenges? Feel free to share a comment below to let us know your thoughts and plans.