The ongoing debate about winners and losers in the near future labor market continues apace. Some believe we are facing an armageddon of sorts while others see great hope. I see a confusing picture where the near term seems fairly well set while the further out future is anything but clear.
This is a topic I've been batting back and forth with Phil Fersht, CEO Horses for Sources (HfS.) This is my take, which, based upon those same conversations, is not so far adrift from where Fersht sees the action. Let's level set this debate so that the context is clear.
Most recently, the prophets of doom over at Oxford University as reported by Axios, doubled down on their prediction that there will be catastrophic job losses as automation in its various guises picks up and takes hold:
...automation threatens 47% of American jobs. And he belittles critics who say technological upheaval seems scary but always generates sufficient new work...
Despite appearances, brick-and-mortar still accounts for the overwhelming percentage of sales — 92%, according to Citi. But e-commerce is up from a 2% share a decade ago, and the number is rising by 20% a year.
The cost of automation is falling, making retailers more willing to use it to lay off more costly workers.
80% of the jobs in transportation, warehousing and logistics are vulnerable to automation.
63% of sales jobs are vulnerable.
The average number of robots in fulfillment centers like Amazon warehouses was 461 in 2013. But now it is 3,200
"The disruption is very much at its infancy," Martin Wilkie, the lead author of the report, tells Axios. He said that, in the coming decades, "the premise is that stores are bypassed in their entirety."
The authors acknowledge that real world job losses are patchy but that doesn't prevent them from pointing to trends and projecting forwards in an aggressive manner. Still others think that the application of robotic automation, machine learning and other forms of applied artificial intelligence will impact almost every job sector within a short period of time.
The financial rewards are viewed as too great to ignore. That thinking is pushing some companies to aggressively adopt advanced technologies at a furious pace. Some data I have seen suggests that automation (in the broadest sense) adds 3-5% to the bottom line today with much more expected in the future.
Optimists on the other hand, draw upon history to assert that while legacy work disappears in an automated world, new types of job emerge that ensure people have work. I'm not so sure.
Regardless of your position, all agree there may be some short term disruption in the market place, although some point to the fact that in the U.S. at least, any disruption is not showing up in the labor statistics with historically low levels of unemployment. The same is true in the UK where unemployment rates are at an historic low. Both those statistics can be disputed for a variety of reasons but let's hold that thought.
A jobs crunch and an expansion
For its part, HfS sees significant impact in IT/BPO services across multiple geographies although the overall impact is considered manageable:
...the global IT and BPO services industry employs 16 million workers today. By 2022, our industry will employ 14.8 million - a likely decrease of 7.5%* in total workers (see our research methodology below). This isn't devastating news - we'll lose this many people through natural attrition, but what this data signifies is this industry is now delivering more for less because of advantages in automation and artificial intelligence.
Where the analysis gets interesting is in the composition of those jobs over time. Here, my concern lays in HfS's contention about filling the demand for high and medium skilled jobs it is predicting through 2022. See below for high skilled jobs as an example where there is a lot of potential good news:
The position is not quite so rosy for medium skilled jobs:
At this point I'm not going to argue HfS figures. Neither am I going to argue Fersht's assertion that things will be sort of OK the next five years.
Retail as the first target
However, I sense that the pace of change will be disproportionately high in some sectors, and especially among retail and retail support. I believe that changes in those segments will move much more quickly than we currently imagine. I take that view from reviewing the reported progress and numerous examples we are tracking through our retail case studies and ongoing analysis of firms like Walmart, Target and Amazon. Those are big employment segments.
Amazon is an interesting case in its own right. In conversation with Fersht, he and I agreed that savvy buyers are defaulting to Amazon for a whole range of goods.
I'd argue that has extended to companies like Wayfair, Groupon and others where the ability to get simple return, refund and/or replace beats the heck out of traipsing out to a store or, worse still, dangling in call center hell with the companies who make and market the goods that can be also found on Amazon.
Granted, Amazon has yet to come close to automating pick and pack - that will come - but they've smoothed out the customer experience to the point where they're crushing it in some high value categories and among a whole slew of commodities.
Service oriented business?
Putting retail aside, my next point is to ask what happens among customer service oriented businesses like banking, insurance and telecommunications? Here the picture gets murky very quickly. Much of the line one banking support among the major institutions has been outsourced to India. If Fersht is correct that self service (as an example) will accelerate, then those low level inquiry type jobs disappear. But it is what happens when you need to escalate that concerns me.
Both Fersht and I agree that escalation queries today - exception handling if you will - is poorly served. Those line 2 and 3 jobs would fulfill HfS medium and high level jobs for which there is anticipated demand. But where are those people and what do you need to do in order to train for those situations which often threaten the entire customer relationship?
I argue that what's needed is an intimate understanding of both the customer and the manner in which the institution chooses to operate. But already we see botched examples of how this doesn't work in the real world.
HSBC in the crosshairs
HSBC has recently become the subject of unwelcome attention in the media with its 'Safeguarding' inquiries that have flooded the mail boxes of UK based businesses. Couching their inquiries in language designed to make you believe that their highly detailed business questions are to help ensure your banking safety, HSBC has managed the double whammy of not only annoying customers but summarily denying access to some customer accounts:
An avocado importer, an e-cigarette seller and a toilet-cleaning gun maker are among British companies that have had accounts closed or frozen by HSBC in the last two months, unintended casualties of a crackdown by the bank on illicit money flows.
Dozens of companies have been affected, with some unable to pay staff and suppliers, and others suffering financial losses they fear could force them to close.
The number is a tiny percentage of all business accounts at HSBC, but those affected are in the small business sector which the government has said it wants to encourage and hopes will thrive after Britain leaves the European Union.
One distressed casualty took to Medium to offer a warning about HSBC - the short version:
I cannot wait for this nightmare to be over. Once it is, I cannot wait to transfer my business away from HSBC. Assuming I still have one left to transfer.
The author finally got his situation resolved but, it seems, only because the article went viral and a customer service exec picked up the phone to clear the problem.
Startups are particularly vulnerable in these circumstances because, as the author says, they rely on multiple services to keep going. Once those services start failing through denied payment, the business stops existing.
Just to cap this off, HSBC has misled customers. The intrusive level of detail the bank requires cannot possibly be related to that required under normal banking conditions. Our distressed author has the answer:
Safeguarding is a process that HSBC is taking all of its accounts through in order to better understand how they are used. It’s the fallout of HSBC receiving a record $1.9 billion fine as a result of a US Senate ruling where they were found guilty of allowing money laundering to “drug kingpins and rogue nations”.
Would any legitimate business want to be associated with THAT?
The more serious aspect of this though lies in the fact that HSBC are so messed up that, apart from summary action denying service in some cases, it is taking otherwise perfectly legitimate customers weeks to get access back to their accounts. That's a failure of huge proportions in the bank's core compliance processes that will be hard to quickly fix given the scale at which they are attempting to operate the compliance process itself.
More to the point, it shows a real failure on the part of this bank (and I suspect many others) to have much clue about what's going on inside their customer accounts, despite all their talk of customer care and understanding.
Accounts payable as an example
Research on deep process areas like accounts payable, which is covered in the Financial Leaders Benchmarking Report (PDF) suggests there is a long way to go:
The average AP organization processes only 64% of its invoices electronically. Upper-quartile AP organizations, though, managed to process 93% of their invoices electronically, a significantly better performance. One organization even stated that they processed 100% of invoices electronically.
Next, how many are handled touchlessly—that is, without human intervention or error correction? In the average AP organization, the figure is 45%. Upper quartile performers managed a significantly better level of performance, touchlessly processing 81% of invoices. Best-in-class AP organizations, though, raised the bar significantly, processing 99.8% of invoices touchlessly.
Once processed, invoices must be paid in order to close the transaction. How many AP organizations manage this in the most efficient way possible—namely electronically? Once again, we see significant variations in performance. On average, AP organizations pay 67% of invoices electronically, while upper quartile performers manage significantly better than this performance, paying 90% of invoices electronically.
In the U.S., the three part check is alive and well - need I say more?
In other areas like document signing, there is a huge gulf between those firms that 'get' the process improvements brought by electronic signing and those that don't. Here I suspect that a laggardly legal profession, keen to keep paper flowing, is a main culprit.
Where to next?
Fersht has some additional research which has yet to be published but to which he gave me exclusive access in advance of an event the company is having, designed to discuss the problems which I am highlighting. Bear in mind that I am only scratching the surface and the body of available research is slim. Here are some highlights from the research I can share.
- Firms say that they are far more focused on opportunities than threats by a factor of more than 2:1
- Firms are more concerned about data than people and very few are concerned about adapting legacy work culture to become more real-time
- Back office is not keeping pace with the amount of change in the front office - finance ad accounting are particularly vulnerable
- While the back office is not seen as a priority, there has to be a question mark over whether accounting and HR can support the required changes to meet stated objectives
- Automation is seen as very much a back office requirement rather than front office
Fersht has consistently said that he sees plenty of acceptance of the need to modernize at the highest levels but less commensurate appetite at the operational levels and especially among the BPOs. Given what we know about the impact on low level jobs and the empires they have spawned, is it any surprise? Even so, modernization continues apace in many parts of the business.
I suspect both Fersht and I are in violent agreement that while the near term future course is reasonably easy to see as settled, there is not only a LOT to do, but it is hard to know where the skills are going to come from that will deal with anything beyond simple exception handling.
I'd go further and argue that while HfS research findings are optimistic in terms of outlook, the focus among researched customers is too skewed towards data and not enough towards modernizing the support functions like HR and ERP while the people elements are definitely taking a back seat. And while firms say they want to work with ecosystem partners in enabling business, I question whether those skills exist among many of the current crop of BPOs.
It is all very well for optimists about the jobs market saying that structural problems can be solved through a greater emphasis on STEM education. That takes decades to feed into the system. Such education always needs augmenting with experience. Yet it seems that ageism is alive and well with positive perceptions skewed towards the younger person as more useful than those who, through their experience, understand the creative workarounds for that crappy customer process.
That is the nub of the problems both Fersht and I are currently seeing in our different ways.