In 2010 I founded an enterprise software platform called WorkMarket to enable companies to manage their on-demand labor.  At the time on-demand labor accounted for about 25% of the labor force.  The general consensus among labor “experts” was that by 2020, 50% of the labor force would be on-demand labor.  Given what we do at WorkMarket, this was a very beneficial forecast!

However, this prediction, like many predictions about the future of work, never had the slightest chance of coming true, and the fact that so many people quoted it as gospel was frustrating to those of us driven by data.

Too many predictions about the future of work are overly simplistic, causing many business leaders to make decisions based on false expectations.  Sensational predictions are certainly good for headlines, but they don’t actually help leaders make informed decisions as to how companies and workers come together to produce goods and services.  The public policy decisions we make, from regulations to tax and trade policies, are largely based on where the world of work is going.  Companies, families, and society need accurate and thoughtful predictions to effectively plan for an unknown future.

This is why I wrote “The End of Jobs.”  I wanted to bring some clarity to the discussion of the future workforce based on real evidence.  In my opinion, labor market predictions need to be based on evidence including: history, data trends, and corporate decision making.

History tends to rhyme.  Things change and evolve, but as we enter what many are calling the fourth industrial revolution (robots and artificial intelligence), we need to study how workers, companies, and society reacted to the first three industrial revolutions (mechanization, electrification, and computerization).  In each previous change, a new technology massively shifted the balance of power to companies.  Understanding how companies and workers adjusted to a changing power dynamic in the past is informative for unions, policy makers, managers, and shareholders today. 

Now onto the data.  Those who don’t understand data have zero business predicting anything, let alone predicting the future of work.  We can simply look back on the prediction that on-demand labor would be 50% of the market by 2020 — an erroneous prediction based on merely one year’s worth of data (the change from 2009 to 2010).  Data needs to be observed over long periods of time so that predictions can be made within their historical context. Thus my book includes a discussion of data and how labor statistics change over time.

Our last pillar of thoughtful and meaningful predictions regarding the future of work is corporate decision making, which is vital to understanding how the world of work will evolve.  Many assume companies will always choose the lowest cost worker at all points.  This is categorically untrue.  Labor resource planning involves many factors (in my book, I refer to it as the Labor Equation) such as duration of the work, intellectual property involved, customer touch points, touch points to other stakeholders, technology, and regulation.  Cost is, of course, an important factor, but it is only one of many.  We need to study how companies engage workers today and how they have done so throughout history to get a firm foundation in predictions of how they may engage workers in the future.

History, data, and a nuanced understanding of how companies make decisions will help us predict the future of work as we stand on the precipice of the fourth industrial revolution (which has become even more paradigm-shifting in light of the pandemic).  One cautionary tale is that of the Automated Teller Machine.

When the ATM achieved near universal bank branch penetration in the mid-1990’s, there were 500,000 bank tellers in the United States.  And what prediction did so many people make about the bank teller job?  With the proliferation of the ATM, many said that there would be no need for tellers in 20 years. 

Others looked at the ATM and saw an example of co-bots (robots working alongside humans and taking away the mundane tasks).  The mundane task of accepting and dispensing cash moved mostly from the teller to the automated teller. Many tellers were moved from behind the glass to focus on customer service and upselling mortgages and investment products.  So the ATM reduced the need for tellers, but also allowed for an expansion and evolution of a teller’s responsibilities.

Well, here we are 25 years later, and there were 600,000 bank tellers employed in the US last year.

It turns out the biggest variable on teller employment in the US was actually banking deregulation.  The number of bank branches in the US approximately doubled, and the average number of tellers per branch decreased from 21 to 13.  As banks competed more directly, they had to  improve the customer experience to attract new customers.  Automation led to fewer tellers per branch (as did mobile banking), but other factors led to an increase in branches.  

The ATM example illustrates why a simple conclusion about jobs belies the complexity of how companies and workers interact.  Just because technology X exists, that does not mean all the jobs in function Y will be eliminated.  Is it possible? Sure.  But history, data, and the way companies actually make decisions tell us it is highly unlikely.

So be wary of any predictions on the future of work that do not take the above factors into account.  Be wary when someone predicts huge movements in labor statistics, because history shows they rarely happen.  Be wary when a prediction is made on limited data.  Be wary when someone makes simplistic statements on how companies engage workers.  Study the past, analyze the data, and talk to those who are actually doing the long-term planning on labor resourcing, and you will have a more accurate view of the future.