Digital Transformation: The Evolution of Search

If there’s anything in enterprise that’s changed due to this pandemic, it’s that we’ve seen Digital Transformation become a “thing” again. It was a hot buzz phrase that was overused and downright waning by the early 2010s.

However, just one layer below that transformation, exposed in plain sight, a different shift has transpired that only a handful of companies have fully realized, much less taken full advantage of for their own unique set of incredibly valuable data that lies within their own organization.

Google - Consumer Standard and Educator

There are effectively two entire generations of employees and customers that have been taught by Google (not social media) to start their transactional journey by typing a four, five or more word question into a search box—or by directly speaking their question aloud to an AI-powered device to initiate the Google search for them.

For employees, it's asking questions about their job... how to do it, where to find enterprise information on how to do it better, or where to get a new job entirely. And for consumers, it's the same ballpark number of words typed or spoken about the product or service they desire.

In the scenario for employees, checking several different information silos is likely going to be a time suck. In the likely event that they don’t locate the answer, they will then turn to a colleague or to Google and click away—resulting in more wasted time.

As to the case of the consumer, once they’ve come to your site and found the navigation nothing like that of Amazon or Google, they will end up making your bounce rate stats even worse than previous (adding insult to injury) by heading back over to Google and costing you potentially more money (remember, you already got them to your site via some channel you previously invested in).

Once there, they will often click one of your ads—ugh! But now, as they are back into the wilds of Google, they could very well simply choose to head off to one of your competitors. The most expensive click is the one from your site back to Google.

In an ever increasing number of instances, Google is providing the answer to the searcher directly in the customer example. This “answer” is most often not coming from your website, but likely another 3rd-party source that you’ve not even sanctioned. And worse yet, it could be wrong.

As an executive or team player in enterprise transformation, innovation, or experience, you might ask yourself, do we have the right search and discovery priorities and correctly corresponding initiatives and projects (in partnership with IT—not something vacuous that will never launch) underway? Something that will actually position your firm to not only answer the aforementioned questions, but have the process set up for them in a way that they will actually initiate the journey and continue until a satisfactory end in their (employee or customer) mind?

Enterprise Search Vendors on the Treadmill

I began my technology in the mid-90s, jumping fully into Search platforms in 1998 (the same year Google launched), with the last decade focused on enterprise search for both corporate intranets in banking as well as large-scale ecommerce for client-facing implementations.

Over this time, I’ve heard a spectrum of pros and cons on the search and discovery solutions that executives considered, and either rejected or implemented. As with many platform technology projects of the time, a high percentage were considered failed upon arrival, but they’d gone too far down the road to reconsider and so to save face in their organization, they trudged on.

But it really wasn’t totally their fault or the fault of the product and service partners they’d brought in to present and execute a “paradigm shift on discoverability or findability of information for the company.”

It was actually Google’s fault. Why? Because for all the meetings I attended with hundreds of companies since 2010, in at least every other meeting I heard, “Why can’t the results be more like what I see from Google?”

You see, Sergey and Larry, over the first decade of the millennium had educated the planet on what the search experience should be—far surpassing Yahoo, Ask, Bing, Excite, Lycos, etc., effectively educating the first of two generations. They then continued to add features and capabilities (at breakneck speed) that most all the 3rd-party enterprise search engines had to emulate and integrate to keep up. Why? Because enterprises saw it and wanted it—right then and there. But Google wasn’t done…

All the while Google had everyone else on their heels, and indeed a treadmill of feature creation and implementation—yet they were actually up to something even larger.

Enter the Knowledge Graph

Of course, leave it to Google to zig when everyone else was zagging. In 2012, Google owned 60% of the public-facing search market and was selling its own search appliance* into commercial concerns during the same period.

More specifically, since its founding, Google had compiled a tremendous amount of data based upon the billions of searches it had been processing. More than just searches—it had been getting asked long form questions in natural language. This was no longer about the 1-2 keywords people entered. It was more complex.

As the questions were answered, a positive feedback loop was created as searchers began to enter even more complex questions—as is human nature. As a result, Google started segmenting much of this data into a structured knowledge graph of people, places, things like phone numbers and addresses.

Plus, it also had semi-structured data in the form of FAQs, and indeed many organizations had indexed unstructured, long-form content into white papers, research, stories, Google Scholar, Books, etc. The result was the implementation of a multi-algorithmic strategy whereby results are provided in an experience that a user can rationalize and consume logically.

By the way, since educating and setting the expectation for the second of two generations by implementing a knowledge graph (featured snippets, knowledge panels, FAQs, etc.), Google now controls nearly 93% of worldwide search traffic.

And this is not limited to just Google any longer. Knowledge graphs are the de facto standard for the highest valued technology companies of this past decade. Every time you ask Siri or Cortana a question, you interact with a knowledge graph.

It’s not a mistake that these massive tech brands—including FAAMG (Facebook, Apple, Amazon Alexa, Microsoft, Google)—each owe much of their growth, dominance, and mastery over each of their own sources of truth, both inside and outside their organizations, to the implementation of knowledge graphs as the core methodology of discoverability by Google.

They have fundamentally transformed themselves with better search by leveraging knowledge graphs and multiple algorithms that interact according to the types of requests being made.

Competing with Tech from the Last Millennium

To borrow from a famous Seinfeld episode, these companies are the masters of their search domain. They have and will continue to win because of it, as the ultimate source of knowledge about a company is the company itself.

There are many recent business examples that have either been lapped up or filed for bankruptcy (Hertz, GNC, Mallinckrodt Pharma, Chesapeake Energy, Brooks Bros.) due to not managing their digital and/or search transformation. Now think about all the ones that haven’t had this happen to them yet.

The even more frightening reality of this is that virtually every enterprise is leveraging search technology (keyword index based, aka “insight engines” per Gartner) from the Seinfeld era. Their search transformation is at high risk, and so is their business.

Four Components to a Modern Enterprise Search Transformation Initiative

This article references a fair amount of enterprise terminology that may have not been very prevalent on executives’ radar in the past. Terms such as natural language technology, knowledge graphs, AI Search, multi-algorithmic federation, etc., and their scarcity can be linked to the fact that these technologies have been primarily controlled by FAAMG for the last decade. But all that has changed.

For instance, natural language technologies are an incredibly high growth, emerging industry that will drastically impact the ways in which we interact with computers in the coming decade. This, as well as other technologies have been made more widely available and innovative platform companies are moving to develop search transformation solutions specifically related to the commercial data that you own and control.

Natural language understanding, knowledge graphs, support for multi-algorithms, and a fourth element—the data from your enterprise (the ultimate authority)—are the requirements that can unlock the search transformation experience for your clients and employees.

Whether you’re guiding a customer to self-serve a support issue, a service agent to resolve a complaint more expediently, facilitating a transaction on your ecommerce site, showing a financial advisor how to open a 529 college savings account, or revealing the contraindications for adverse drug reactions to a physician assistant, today’s search applications require modern technology components from today—not from the 1990s.

Search is no longer simply a utility whereby you plug and play a search platform vendor based upon a cookie cutter RFP document that was written in 2012.

In the enterprise—whether outward facing for support or ecommerce, or harvesting critical information for knowledge workers to be more productive—an AI Search initiative—indeed a Search Transformation—can play a pivotal role in an overall enterprise digital transformation. One that provides the direct answers, creating the experience that is expected by customers, prospects, and employees alike.

#datascience #ml #machinelearning #ai #mlengineer

*Google Search Appliance—End of life as of April 2019