** My attempt to write more consistently about things that I find interesting or spend time thinking about. I’m hoping to write 10–12 pieces this year.**
Even in my first few years working, I’ve felt gradual shifts away from the center (and we’re not talking politics here). Uber drivers now commute to work in San Francisco from Sacramento, the newest closest bastion of affordable housing. Friends have moved out of the city in search of fresh air, new opportunities, and better access to schooling and housing. The predominant undercurrent here is that new economic opportunities are displacing old ones. There are a number of ramifications here, but one of the most exciting and worrying aspects is the rate of change appears to be accelerating
What do these shifts at the macro and meso-levels tell us about how companies and workers might react or adapt?¹ Looking at the next five years, I think we’ll be seeing the increased impact of new tools reflecting increasingly distributed workforces, further emphasis on data collection and automation, more integrated services, and changes in which core tasks require human input.
Work has traditionally been categorized into four buckets based on skillset: unskilled, semi-skilled, skilled, and highly skilled.²
However, I think that work is better understood as a set of commonly conducted core tasks. These core tasks can increasingly be classified as “already automated”, “enhanced by automation”, or “cannot be automated yet”. A 2017 McKinsey report found that 60% of jobs have at least 30% of tasks that can be automated using current technology. The jobs impacted by automation will be not limited to just unskilled and low-wage ones. In fact, we already see advances in radiology, traditionally one of the highest paying jobs, where robust data sets are being used to achieve similar or superior results to human review.
Automation has and will continue to be a theme as long as the cost of human capital continues to exceed the perceived price of replacement technology. That means that relatively complex low-wage tasks will continue to thrive due to the relatively high cost to develop automated solutions, e.g. landscapers who trim hedges and plant flowers. On the other hand, tasks involving creativity, complexity, and dexterity have traditionally proved to be resilient to automation will fall, though maybe not in the same time frame. We will see traditionally safe white collar jobs be “enhanced by automation”, especially in healthcare. Historically, we’ve often seen a lag of 2–5 years between hype and adoption, so I’m very excited for companies like Project Ronin, Canvas Medical, and Epharmix see further penetration into oncology, primary care, and patient outreach.
On the other end of the spectrum, the tasks that might prove most resistant to automation will be ones where people have a learned preference for human interaction. While I would rather talk to a nurse than a machine, but even those preferences might change over time; transitions in population and advancements in technology might see to that.
New Skills, New Tools
As more roles are merged and eliminated through changes in the workplace, the demand for new skills will emerge. Both companies and workers will be incentivized to gain additional skills. Historically this may have been filled by vocational schools and more recently by online courses and coding bootcamps. I think we see strong continued demand for new services teaching people new skillsets.
And as workers upskill, we should see changes to the tools used in the workplace both in hardware and software. In hardware, I would expect to see increased usage of hands-free AR devices that will help professionals who regularly use their hands for their tasks. In software, we will see increased usage around flexible, integrated, and more technical interfaces, as workers retool to become more technical and analytical in their roles. We should also see better user awareness of the benefits of automation tools that remove rote tasks from their plates. To this end, I’m excited to see tools like Retool, Notion, Airtable, Zapier, and X.ai become more widely adopted across organizations.
The ability to acquire massive data sets while being one of the main drivers of how work is conducted in the next five years. The influx of data will come from digitizing decades of data, concentrated efforts at manual collections, and sensors. Sensors will be everywhere, in homes, factories, transporters, stores, warehouses, streets, and everywhere in between. I think new technologies will be appearing in the next five years that utilize machine learning to harness these dedicated reams of data. We’ll see this being applied in all industries, but I think we’ll see the biggest changes in manufacturing, health, retail, and transportation.
Focusing on manufacturing for a second, I’m expecting to see an acceleration of the iteration speed. In the fashion industry, we’ve seen optimizations in the supply chain have allowed a shorter time between idea to store. Other industries stand to benefit from optimizations such as localized additive manufacturing. Companies can eliminate weeks from their product development cycles by relying on providers like Voodoo Manufacturing and Origin. Instead of having to send your designs out for test runs half-way across the globe and then get it shipped back, companies can build and iterate on samples in-house/near-house, reducing lead time from weeks to hours.
We’re currently starting to see a new school of thought emerges with regards to data aggregation. Previously, the major trend was data aggregation into individual data silos. Segment has helped companies connect their data across services internally. Google and Alibaba (and their portfolio companies) have touch points to collect data on each online and offline interaction. Now, we’re seeing this counter-movement towards the sharing of data across different entities. I don’t expect to see this universally, as there are too many incentives in place for many existing companies to relinquish their valuable data, but I do see opportunities for financial, insurance, and medical providers to benefit from shared data to reduce (compliance) costs and increase (transaction) speeds. We first have to see sustained success from POCs such as Japan’s Payment Card Consortium, which will share data between processors to reduce instances of fraud, before seeing broader adoption from lagging adopters in the 5-year time frame.
The New Work Environment
Buoyed by secular shifts in the increased flexibility of work, I think we will see people work increasingly as individuals in remote settings. More people will take on-demand roles, take advantage of new individualized business models e.g. social media, and trade-off centralized commercial amenities for more flexibility. This will result in a demand for new tools and organizational processes for collaboration, creation, management, distribution, and monetization. Furthermore, any future recession will accelerate this shift as companies look to reduce costs, leading to more people siloed to the home and looking for alternative means of income.³
Commute Time: 0
I’m really excited by the potential of an autonomous commute. According to the most recent US census, an average person spends about an hour on their commute each day. That’s 3.6 billion workdays spent in traffic each year. Instead of sitting behind the wheel, imagine being able to spend those hours taking care of other tasks. Companies will also be able to monetize these rides in a number of different ways. I’m particularly interested in the potential of advertising and tailored mobile workspaces.
One might think that there are tensions between the trends of increasingly autonomous transportation and remote work. Will more remote workers mean fewer people commuting or will remote workers dampen the use of autonomous transportation? I actually think the two trends share a common theme around the break-up of the traditional 5-day / 8-hour work week in the name of increased flexibility. There’s already been plenty of research indicating how current open offices are inhibitors to productivity, making centralized locations ideal for collaborative work, but perhaps not for individual work. It’s easy to imagine a world where employees have the flexibility to choose the best work environment for their given tasks.
On the benefits side of work, I believe we’ll see employers increasingly provide more integrated services in a bid to attract and retain talent. The largest companies already offer some combination of financial, recreational, childcare, and healthcare services to their employees. Integrated healthcare is a no-brainer.
Healthcare is the second largest expense after employee wages and continues to be the fastest growing operating expense for employers.
I think we’ll see concentrated efforts from larger companies to provide more integrated health services such as on-site clinics, telehealth visits for remote workers, and utilization of predictive analytics and wearable data in a bid to improve margins. WeWork is heading down this path for their members while companies like Collective Health are helping smaller companies make the transition. I’m also expecting to see additional resources spent on mental and emotional health, an area that continues to fight stigma and staffing shortages.
Automation, tooling, and distribution are recurring themes that we’ve seen play out through many iterations in the past. For me, I think it’s a useful exercise to re-examine these themes every year or two both as a way of generating new product ideas and of understanding of areas where upskilling may be valuable.
What do you think? Do you disagree or think there are more important themes that will define work in the next half decade?
¹ My thoughts here mostly apply to the US, where I have the most access and experience. Though after having spent 20% of my time in China in 2018, I might spend some time thinking about how work in China is changing on an adjacent but somewhat overlapping path.
² Examples of unskilled workers are cashiers, cleaners, construction workers, and farm laborers. Semi-skilled workers complete repetitive tasks but require some training and monitoring. Examples include truck drivers, restaurant waiters and cooks, and customer service representatives. Skilled workers include teachers, electricians, sales representatives, and law enforcement officers. Tasks require decision making but training can be done in a relatively short time frame. Highly skilled workers include doctors, lawyers, architects, scientists, engineers, and financial managers, who often must make decisions or generate new ideas in a broad range of uncertain situations.
³ On a related note, I don’t quite yet have a clear thesis on the impact of a widening gap in income distribution.