January 29, 2024
Consider a handful of headlines from the last two years:
– ‘Global Cargo Trade Runs on Paper Slips’ (Bloomberg, October 2023)
– ‘Southwest Airlines Meltdown Caused by Faulty Tech’ (Bloomberg, March 2023)
– ‘Logistics runs on MS Excel at many companies’ (Forbes, Apr 2022).
Stranded airplanes, packages arriving months late, organizations caught in endless spin unable to make decisions. Stories like these, usually the quiet frustrations of everyday life, have come to the fore since the COVID-19 pandemic, as companies went remote, labor markets have tightened, and the gap between legacy industries and ever-improving technology has grown. These problems mostly cannot be easily solved by throwing more people at them. They expose broken systems and paradigms that have failed to scale and reveal a quiet truth: the world is not as digitized as it seems.
In practice, digitization for most industries has meant Microsoft Excel and Email. These tools are powerful: a glance at an office of the 1960s shows vast secretarial and courier pools no longer needed with instant, electronic communications. Modern spreadsheet tools enable the average analyst to work far faster than entire teams two decades ago1. Low-code integrations on top continue to bring improvements to productivity.
Yet, for the most critical digital needs, these tools are far from sufficient. Spreadsheets and email are no way to direct airplanes across the globe, track cargo ships on the high seas, or even share scientific information to get to the next critical experiment. Fundamental business processes require systems of record with a single source of truth, interfaces tightly tailored to user roles, and bespoke analytical operations reflecting industry-specific processes. Excel is too often the opposite design paradigm from what is needed: powerful and solipsistic, not regimented and universal.
The clearest example of the potential of true digitization comes from finance, arguably the only industry besides telecom/tech that has fully converted2. Funds flow seamlessly through the banking system, across firms and continents3. Equity trading is free and can be done from an app on your phone. Continually fresh webpages show us up-to-the-minute details on our financial affairs and the state of the market. Financial systems rely on centralized, digital clearinghouses, exchanges, and information brokers, enabling evermore efficient operations as new sinews tie together what was once siloed.
By contrast, the few financial transactions that remain on paper, such as mortgages and 401(k) rollovers, feel archaic. Mortgages are still mostly contracted on paper forms. Without automated clearinghouse systems, notifications come only by mail when they are traded, and billing needs to be reconfigured by hand. 401(k) plans are rolled over in checks sent by mail, requiring all their component funds to be sold, even when they are immediately repurchased in a new account.
Digitizing (most of) finance was not an overnight process: two decades of innovation have spawned the $245 billion fintech industry. Connections have been built to enable rapid, automated transaction processing, funds transfer, equities and derivatives trading, etc. Mortgages and 401(k)’s represent two of the last vestiges of old, manual finance—and their lack of digitization serves to make clear the stark difference when an industry actually commits fully to the process of switching over.
The power of a mobile banking and brokerage app is inconceivable in most other industries, reflecting just how rarely true digitization has been accomplished. The internet and telecoms revolution did not end in the ‘90s or the ‘00s. Rather, we are just scratching the surface of the benefits of digital technology. There will be enormous gains from effective systems tailored to industry processes. There will be entire new ecosystems built digitizing the economy.
Consider Logistics. At the core of trade, manufacturing, transportation, and more is the ability to move things from one place to another efficiently. Underlying these processes, sit numerous optimization, forecasting, inventory management, and product tracking problems. A disparate software ecosystem provides costly solutions to individual tasks, while much of core operations remains in Excel. Optimization, Planning/Forecasting, Inventory Management, and Product Tracking all occur in separate tools, if systems for them exist at all.
Lacking integrations between tools, teams cut and paste assumptions, actuals, and forecasts between tools. Updating projections is slow and costly, requiring manual labor, creating multiple points of failure, and cascading errors. Gaps like these underlie logistics failures at Southwest Airlines in December 2022 and March 2023 and FedEx in the early stages of the COVID-19 pandemic4.
Much as Fintech revolutionized finance, a new generation of digital tools for logistics have begun to attract serious investment, such as Samsara and Flexport. These tools recognize the fundamental gap in being able to consistently track and monitor goods as they move around the world and to integrate actuals with planning, forecasts, and optimization. With a stronger digital foundation, it will be easier to knit together end-to-end solutions for modern, efficient logistics.
Biopharma, notably, uses very few of these systems and has especially ad hoc supply chains, despite the very high value of goods, limited manufacturing capability, complexity of distribution, and low tolerance for error. Just-in-time manufacturing and similar innovations are rare at best. In a previous role, my team saved millions of dollars for late-stage clinical trials by building a simple drug tracking and depot-to-site distribution app for drug supply. We allowed the clinical supply chain team to precisely assign drugs to trial sites, while accounting for drug expiry and local regulations. These tools reduced the overall drug needed for Ph2/3 oncology trials by a factor of three, saving millions of dollars in drug costs. Solutions like these require digitization, and show its enormous untapped potential especially in biopharma.
I used to believe that biopharma was uniquely and suboptimally not digitized. I no longer believe that is the case. Major industries, such as healthcare, are even less digital. Our pain points in biopharma are not unique: they are symptoms of the larger problem of incomplete digitization. Most every industry needs its own fintech revolution: companies building the tools, primitives, and paradigms to enable the enormous productivity gains from digitization.
Amidst today’s artificial intelligence (AI) boom, we cannot lose sight of the primal importance of digitization. Without it, we are not ready for AI:
• AI requires digitization: it demands large, curated datasets and careful application.
• AI is not sui generis: it can speed the curation and management of data, but only once digital-native processes and tools are built.
• The same infrastructure that underlies the current AI boom makes creation of new digital systems dramatically faster and more feasible: ubiquitous, cheap, secure cloud computation + data storage, and advances in developer productivity tools and infrastructure.
• Digitization is a first step. The data we organize will train the right AI/ML models. The processes we digitize will guide the application of AI/ML.
We have far from exhausted human intelligence: our use of artificial intelligence will be all the stronger by understanding where it can best augment digitally enabled human processes.
1 There’s an excellent description of how accounting, finance, and strategic planning worked in the immediate postwar era at Ford in David Halberstam’s The Reckoning. The objectives are completely recognizable. The time, methods, and resources it took to achieve them are not. Similarly, searching Hacker News for Excel gives a fascinating look at its utility.
2 This chart from McKinsey Global Institute is overly simplistic, but directionally correct.
3 Modulo slight speedbumps, improving constantly.
4 There is an interesting connection between unionized labor and digital systems. Southwest’s union had long been a major proponent of fixing digital infrastructure as smoother operations lead to better work conditions. Unionized UPS has more sophisticated systems than its competitors, which were a component of last year’s labor dispute.
To subscribe to Engineering Biology by Jacob Oppenheim, and receive newly published articles via email, please enter your email address below.