Recruit Holdings: Owner of Indeed and Glassdoor

Anyone that has applied for a job knows about Indeed, but probably never knew Indeed was a fully owned subsidiary of Recruit Holdings (RH), a Japanese holding company founded in 1960 and based in Tokyo. RH started as media company, and throughout the 60’s distributed free guides to university students filled with paid content from companies looking to hire new talent. The offerings expanded to paid real estate listings with more accurate calculations of walk time to nearby train stations, travel brochures, and myriad other offerings in the job placement space.

Recruit Holdings in its modern iteration was born only 8 years ago, when RH began an aggressive series of acquisitions, including the 14th largest U.S. staffing firm Staffmark for $300M in 2011, the 11th largest global staffing firm Advantage Resourcing for $410M in 2012, and the #1 job website at the time, Indeed, for $1B, also in 2012. In 2014 the company listed on the Tokyo Stock Exchange. In May, 2018 RH completed its latest major acquisition, purchasing Glassdoor for $1.2B in cash. Today, RH owns 357 individual companies.

Recruit Holdings is structured as a holding company, with 3 subsidiaries: HR Technology, Media & Solutions, and Staffing.

  • HR Technology (Indeed & Glassdoor): This business unit is synonymous with Indeed, and following the recent Glassdoor acquisition, is definitively the world’s largest job site. The combined Indeed / Glassdoor web properties will have ~260 million unique monthly visitors. Prior to acquiring Glassdoor, Indeed had ~$780M in revenue in FY 2017.
  • Media & Solutions: Offers marketing and recruiting support to multiple business segments in Japan, such as wedding venues, restaurants, and beauty salons, with ~$6B in revenue in FY 2017.
  • Staffing: Provides temporary staffing services in Japan and internationally, with ~$12B in revenue in FY 2017.

Indeed itself has a fascinating history from a startup perspective. The company was founded in 2005 by Rony Kahan and Paul Forster, who raised one and only one VC investment, led by Union Square Ventures for $5M. Rony and Paul had previously founded Jobsinthemoney.com, a job site for financing professionals that they exited to Financial News in 2003. The company was inspired by Google’s pay-per-click business model, Google itself having just gone public in 2004. Google inadvertently contributed to Indeed’s success, by indexing jobs listed on the site and oftentimes enabling Indeed’s job listings to be a top search result. In other words, Google became one of Indeeds largest source of organic (i.e. unpaid) web traffic.

Moving forward, I will be writing more about Indeed / Glassdoor, which today is one of four of the major players in the job listing / job search space. The other three, include Indeed’s oldest competitor, LinkedIn (now owned by Microsoft), and new (but formidable) entrants: Facebook Jobs and Google Jobs. In effect, while much larger today in scale than 10-15 years ago, the world of job listings is effectively dominated by tech companies that began as venture backed startups.

Have We Invented “Artificial Intelligence” or “Decision Machines”?

A brand is a set of strong associations, positive or negative, with a certain word or phrase. These associations may be logical and straightforward or unreasonable, primal and emotional.

The word “artificial” has a logical association of “not natural” or “machinelike”. The word “intelligence” has a logical association of “intelligent, but not human” or “developed consciousness”. These associations are all fine and good from an analytic point of view. However, humans are not only an analytic species, we are also an emotional species.

At this emotional level, the word “artificial” feels “inhuman” or “cold”. At this same level, the word “intelligence” feels “alien” or “godlike”. Combined into “artificial intelligence”, these feelings create a potent emotional sense of unease and fear. We imagine cold, godlike machines that will soon treat humans as lesser beings. Strange then that we are seeking to create an entire industry organized around a brand that feels frightening to many humans.

What’s worse, until the (theoretic) advent of the “singularity”, these emotions are completely baseless. We are not even close to developing “general intelligence”. Thus far, every algorithm programmed into a computer has failed to demonstrate human “intelligence” whatsoever.

Intelligent humans possess: intent, independence of thought, self-actualization, and will. On the other hand, computer programs lack all of these characteristics innately.

  • There is no “intent” behind the actions of a computer except to serve the function for which it was programmed by the humans that designed it (e.g. win at chess).
  • There is no “independence” of thought, only outcomes at the intersection of different algorithms.
  • There is no “action” without a human to push the “start” button.
  • There is no “will” to persist aside from our allowing the computer to remain plugged into a source of electricity.

We should stop the self-destructive branding of a technology that serves a much less amazing, but no less useful feat, than demonstrating some new form of “intelligence”. The feat that is actually worth celebrating is that we are developing computer programs that can find solutions to complex decisions.

We have not created “Artificial Intelligence”.

We have invented “Decision Machines”.

Just like silicon machines before and metal machines before that, decision machines are a powerful invention. But, like all inventions, they are tools in the service of the humans that created them. In the case of decision machines, their “job” is to help humans make complex decisions, rather than calculate numbers (silicon) or shape physical objects (metal).

Today, decision machines are helping doctors to examine medical images to ensure human experts don’t miss critical information. Tomorrow, decision machines may help us drive cars, by making decisions about the speed and distance our vehicle should travel in relation to other cars and pedestrians.

We should all want a future where we possess a powerful tool to help us make better decisions. One way to get there much faster, is to stop calling these tools by something other than what they actually are: decision machines.

Let’s Speak the Same Language to Solve the “Skills Gap”

We have a problem in the United States. Many employers struggle to find employees that have the capabilities they need to do certain jobs. This problem is called the “skills gap”. Whether or not it is actually mismatch of skills, is subject to debate. However, we can’t even have a debate because we aren’t speaking the same language. We use different words to mean the same thing. They’re not.

Let’s start first with the common misuse of the words “knowledge” and “skill”. They are not the same thing.

Knowledge is abstract. It is a concept that exists in your mind that you either read in a book or learned from a teacher. The capital of Mali (Bamako!), the State flower of Montana (bitterroot), and the Pythagorean theorem (a^2 + b^2 = c^2) are all forms of knowledge.

A “skill” is applied knowledge. It is knowledge with a purpose. Skill is superior to knowledge, because it exists in the real world. Skills include: speaking French to call a cab in Bamako, making tea out of bitterroot, and using the Pythagorean theorem to make sure your ladder can reach the gutter of your 12’ roof.

We can play word games if we want and force knowledge to equal skill. For example, the “skill” of spelling, or the “skill” of addition. However, these “skills” are only useful in an artificial environment like a classroom or spelling bee…not the “real world”. The real world is where skills exist and thrive.

However, “skills” are often necessary but insufficient to perform a given job. Often, to apply your skill, you need a “tool”. If you are skilled at painting murals, you’ll need the right kind of paint brushes. If you’re trying to build a financial model, you’ll need to be able to use Microsoft Excel. You may have the knowledge, and the skill, but if you don’t have the right tool for the job, you won’t be able to do it.

It may seem painful to separate tools from skills. After all, you usually can’t be skilled at something without knowing how to use a tool to perform that skill. However, we need this distinction because there are oftentimes more than one tool to do the same job! This is particularly applicable in the world of computers and software. You may have knowledge of accounting, and the skill to create a financial statement, but employers may also care if you can use the tool of NetSuite or the tool of Quickbooks.

The final variable is experience, which defines the context where you have used your knowledge, skills, and tools in the real world. The same combination of knowledge, skills, and tools may be applied in a variety of industries, a range of company sizes, and for short or long periods of time.

Knowledge. Skills. Tools. Experience.

With these categories in mind, we can ask more precise questions about the “skills gap” in the United States, particularly at it relates to higher education:

  • Do we have a “skills gap”, or gap of “knowledge”, “tools”, “experience, or some other combination of all three variables?
  • Are high schools, colleges and universities producing graduates with “knowledge”, “skills”, and “tools”? If so, what combination of attributes will a college graduate in finance have if they graduate from one school versus another?
  • What “knowledge”, “skills”, and “tools” do employers need today? Which of these qualities will employers need more in the future? Do high schools, colleges, and universities understand these dynamics in order to tailor their future classes accordingly?

We can’t answer any of these questions if we blur the distinctions between knowledge, skills and tools. If we do this, it is quite possible that every component of the U.S. labor market will be doing the wrong thing.

Imagine these worst case (and frighteningly plausible) scenarios:

Colleges churn out graduates with 90% knowledge, 10% skills, and only a few tools, while employers are screaming for 80% skills, 20% knowledge, and an overflowing toolbox.

Workers who want better jobs, but don’t know whether an online course, community college, or graduate degree will provide them with the “right stuff”.

Employers writing job descriptions that don’t mention specific tools, and then wondering why the hiring manager isn’t hiring any of the applicants.

The future of American students, workers, and businesses is too important to risk these worst case scenarios. We must actually speak the same language, and it starts with using the right words to describe the world of work:

Knowledge. Skills. Tools. Experience.

Once we do this, we can try and understand what problems exist, and how to solve them. Best of all, we’ll be speaking the same language to get the job done.