I’m sure many of you heard Google’s CEO Sundar Pichai’s recent announcement of Google for Jobs. He stated that it will make it much easier to find a job. If you would like a detailed rundown on the new initiative I would encourage you to read Josh Bersin’s article in Forbes, “Google for Jobs: Potential to Disrupt the $200 Billion Recruiting Industry“. I highly respect Josh for his perspective on HR, Recruiting, Technology and everything in between. He knows what he’s talking about and I find his perspective is balanced. I don’t always agree, but the point of this post is to identify the shortfalls that I see with Google for Jobs.
Before I tell you what they are I want to describe to you what my understanding is of the new platform. Here it goes.
The first thing to know is that Google has been building what is called a “job family taxonomy”. This is nothing new but we know that titles are used so differently throughout every single industry. Take the example of HR Manager. An HR Manager could be a Director, Lead, Senior Consultant, Senior Manager, Head, and the list goes on and on. Google is basically pulling these different titles together so that when you search for a specific job title the result will pull other titles that do the exact same, or similar, type of work. To be honest, I have no idea how they will achieve this, but if they do it will be a miracle.
The platform takes it one step further. Instead of just searching job titles it also assesses the job posting and the content contained within it. It combs every single word and through its Natural Language Processing (NLP) algorithms it will further improve the results of the search.
The final aspect is its filter capabilities. Not only will it allow users to filter by traditional means (e.g. location, experience level, date posted, etc…) it will include many other new filters. For example, you will be able to search jobs by commute times by leveraging Google Maps.
So on the surface I like the idea that this platform will help solve the problem of job seekers not currently getting great search results. I would see it as a success if they improve the process by just 1%. I have no idea how this would be calculated, but the point is any positive outcome is better than none.
Here are the Problems that Will Continue
The first problem is regarding the balance between active and passive job seekers. Bersin states that 20-24% of Americans change jobs every year, which means more than 41 million people are searching for and being recruited into jobs every single year (US only). I would like to see of the 41 million, which ones were actively seeking new opportunities versus not actively seeking new opportunities. Why does this matter? Isn’t the point of Google for Jobs to make it easier for seekers to find jobs? If for example, 75% of the 41 million people who changed jobs last year were passive seekers then I don’t really see how Google for Jobs will help them? They’re not actively looking trolling the inter-webs. The only potential way is if someone sees a job and then sends it to someone in their network. But that’s not what Google for Jobs is intended for. At least this is not what I think it’s intended for.
The second problem is what I call “garbage in, garbage out”. Job postings are brutal. We know this. They’re horribly written, not enticing, lack any kind of branding whatsoever and lack clarity on what the core of the job is. So if Google for Jobs is a machine learning play how in the world can it work if the data it is evaluating is brutal at best? There is no one standard. Each organization, recruiter, hiring manager, etc… write differently. I am truly hoping that I am wrong but you can’t be that effective if you’re using crappy information as your starting point.
The third problem is that there is a ton of available work that is not posted anywhere online. Some sources say that the percentage of unadvertised represents 70-80%. This is a massive hidden market. How would Google for Jobs match seekers to the hidden job market? It simply can’t. This is a huge problem because this is where seekers need to be spending more of their time, cracking the hidden job market open.
While I am always interested in anything that comes from Google I am not overly optimistic about how Google for Jobs will solve the key challenges facing talent and organizations today. However, if it does what it says it will do, which is help improve the ability of active seekers to find jobs then I think it’s a huge win.
Your point about the lack of clarity in job postings, let alone job descriptions, is very good. I am curious to see how googles machine language manages this challenge
I suspect, like you have already concluded, garbage in = garbage out.
Bill, thanks for weighing in. Only time will tell what will happen. I’m not sure that the most complex technologies will ever be able to fully figure this out. You have to deal with different styles, methods of writing, uses in terminology and the list goes on. For example, I recently discovered a new job title that is being used, “Chief Encouragement Officer”. What exactly is that and how would any machine learning platform be able to analyze this? Or maybe it can. Time will tell if this problem will be solved.