Micro-Trends Affecting The Next Wave of B2B Tech

By Esteban Reyes, Founding Partner at Las Olas VC (a.k.a. LOVC)

We spent the last 6 months researching early-stage, investment opportunities in B2B software. This work has become foundational to our investment thesis at Las Olas VC (LOVC), which so far is yielding promising results. We are sharing our findings in this blog post series, which hopefully is helpful to other founders and investors exploring similar opportunities.

In previous posts I explained the Intelligent Enterprise investment thesis and unpacked the macro trends driving the opportunity.

I this post I unpack the micro-trends.

Please reach out to me directly with any feedback, comments, or ideas to er@lasolasvc.com

Micro-trend 1: The digital transformation fallacy

According to Gartner in 2018 the global enterprise IT budget rose at an annual rate of 6%, nearing $4 trillion, and enterprise software spend at a rate of 11% — the highest growth rate since 2007. However, a survey conducted across 679 executives — in which 94% of respondents confirmed they’re increasing their focus on digital initiatives — suggest that not everyone is happy with current growth rates and productivity improvements are marginal.. Further, the World Economic Forum reported that in 2019 over $1.2 trillion will be spent by companies worldwide on their digital transformation efforts, and yet analysis suggests that only 1% of these efforts will actually achieve or exceed their expectations.

In summary, productivity continues to decline despite technological advances and enterprises increasingly investing in digitization.

We believe this dichotomy is driven by three underlying factors:

1. Digital transformation is slow, costly, and risky for large corporate stakeholders. New technology adoption represents organizational change, and in many cases cannibalization of existing revenue streams.
2. Software development and technology integration are still inefficient, requiring costly talent and complex tooling.
3. The utility of modern technology hasn’t been applied deeply enough to effectively solve industry specific problems.

Bain & Company’s research findings state that “most of the digital transformation that’s taking place is experimental.” Enterprises have become adept at generating ideas and testing them. The struggle begins with the next step — scaling the best ideas. Specifically, it means integrating new technology across the organization at a size big enough to make a true difference, and then embracing the new ways of working required to maintain the momentum. Generally speaking enterprises deeply understand the challenges and opportunities within their industries, but they generally lack the ability to modernize and effectively drive productivity.

Micro-trend 2: The machine-augmented knowledge worker

Knowledge workers are people who reason, create, decide, and apply insight in non-routine cognitive processes. There are 230 million knowledge workers globally, and it’s the fastest growth worker classification, representing about 48% of the workforce in the U.S. While humans are the most advanced at general purpose intelligence, creativity, intuition, empathy, and contextual awareness, we lack many abilities that machines can augment.

Machines are devices created to perform specific, repetitive tasks, in a fraction of the time a human can and with equal or better quality. They don’t get tired, can process vast amounts of data, and narrow down to the optimal choice given defined conditions. In software, “machines” are statistical models devised to solve prediction problems, which are typically used to automate tasks and/or improve the quality of decisions.

These facts make it obvious that human and machine collaboration can unlock unprecedented productivity and improve people’s lives. To take full advantage of the possibilities presented by this collaboration, companies need to understand how to make machines intelligent and redesign existing knowledge-work processes and jobs.

Example of human + machine collaboration in medical diagnosis:

General view of fundamental requirements for advancing the utility of machine learning (ML):

• Lots of data to build a new model, or little data applied to an existing model that’s offered as a service
• Subject matter expertise in the particular problem to which ML is being applied to in order to identify features that are persistent and predictive
• Better and faster computing, infrastructure, and tooling that makes it easier to process data, and host and run algorithms

General challenges with ML:

• ML is good at saying what’s going to happen, but not good at telling us why, what’s best, or what’s morally right — only tells us what can be inferred from the past.
• High cost of inaccurate predictions as users have low tolerance and can lose trust in the system (i.e. Clara.ai, x.ai, etc.). This leads to users to ignoring the recommendations made by the machine and breaking the feedback loop.

Importance of domain focused machine-augmentation:

• Easier to capture data and train domain-specific models.
• With domain focus you get optimized data, which leads to better predictions.
• Things like context and structure make a big difference in ML performance.
• What works well in one domain may not work elsewhere.

Up Next — b2b tech investment opportunities in legacy industries: logistics, manufacturing, professional services, and insurance.