#44 - Companies as Experiments Amid Bubbles & Golden Ages

Sharing how I'm thinking through some of the big changes happening in our world + giving some practical guidance on how to apply the thinking itself

Over the last several months, there have been a wave of layoffs hitting across the tech industry (including most recently at my employer, Microsoft). These layoffs have happened amidst a potentially-Cambrian moment for artificial intelligence technology in particular. 2023 is turning out to be the year of ChatGPT (or maybe just the Q1 of GPT…) and also the year of tech “right-sizing” as many pundits are calling it.

This simultaneity can be odd to process. You have:

1. A new technological paradigm beginning to install itself in our markets and culture

2. Macroeconomic tumult amidst global wars which are leading to an era of RIFs (reductions in force) and

3. Societal challenges (i.e., climate, housing affordability / COL, aging populations, educational quality) that could undermine the long-term growth prospects of the United States (and many other countries contending with the same challenges) [will cover in future letter]

How do we think through this? Let me start with a lightweight framework from Carlota Perez, a deeply-respected academic who wrote this book I’m currently reading called Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages. In the book, Carlota posits this way to think about new technologies:

It’s usually not possible to really define what was a “big bang” until much time has passed, allowing you to properly weight the significance of an innovation. But I think the rise of generative pre-trained transformer technology (a.k.a. GPT, made almost-household famous by ChatGPT). It’s quite possible that we are in the early inning of an installation period right now.

As installation transitions into full-on deployment, many news jobs will be created. They will be staffed up by many people who’s prior jobs were eliminated or rendered irrelevant by the rise of GPT technology (and other types of artificial intelligence tech).

These times are almost always scary for people. Even for people (like me!) who tend to view new technological paradigms as times of great opportunity, we have to acknowledge that disruption is painful. A golden age for some is a dark period for others.

But let’s go back to that part about “times of great opportunity”

Many people who have more experience in the world than I do seem to think that this is a time of great opportunity, particularly for innovation. Out of the many hundreds of companies that have conducted large reductions in force, many very capable people are suddenly on the job market. Many of these people harbored or nurtured entrepreneurial dreams, and as the search for a traditional job yields sub-par options or no options at all, they’ll consider building. They’ll dream of creating something of their own.

But dreams are not a strategy.

To dare to dream is not in itself sufficient to lead to results. You also need to dare to experiment, constantly. There’s a joy in viewing things are experiments, because it (at least in my experience) helps to you to remain more objective.

As a personal note, I tend to view learning new skills as constant experimenting, because it keeps me focused on improving + significantly reduces how shitty it feels to “lose” constantly.

I’ve long done this at work, and I think experimentation should flow from the Person all the way to a company. And the companies that constantly experiment are the companies that will discover the future, and they just might own it, too.

Companies as experiments

Whether you are a startup or a mature company, it’s helpful to think about much of what you do as an experiment. From my work with many organizations whose products we use every day, many of them think in bets and monitor bets via experiments.

• And good experiments have a timeline…

• They have variables you’re controlling for…

• They have review periods throughout to monitor the pace of change for select variables…

• They expect degrees of failure and focus on what was learned from failures

• They quickly design + run additional experiments to keep the Rate of Learning (RoL) high

It is no accident than many of the most beloved companies in the world are constantly A/B testing different aspects of their experience. It’s because a wise manager never assumes that the current state is good enough; you should always be mercilessly charging forward to extract even 1% improvements to improve your customer experience (and yes, that also means internal customers a.k.a employees).

Examples of company-level experiments

1. This one is controversial but common: allowing multiple teams to work on the same opportunity space, to have multiple horses in the race, so to speak.

   1. There’s a reason many sci-fi movies about humans escaping earth and searching for a new home almost always feature MULTIPLE space ships…

2. Rapidly standing-up a new, MVP product and doing an “early release” program with a set of very early customers who want to be on the cutting edge

   1. And then getting their feedback constantly to refine the experience, including biz model

   2. Note: this is a great way to test if you want to make a change across the board of your product while managing risk

3. Forging a strategic joint venture (see letter #34 on joint ventures and channel partnerships) as a way to quickly add a capability to your offerings that you can take to market

   1. An on-theme-with-this-letter example is the Microsoft + OpenAI joint venture / exclusive partnershpi (btw, I swear this isn’t a Microsoft blog; no one from our marketing team approves / really cares about my writing as of now :D)

4. Conducting minor (read: not major) re-orgs to give talented people an opportunity to thrive in a different context. Just because someone was hired into one role doesn’t mean it’s still the optimal role. Sometimes the “company-level” experiment is moving a senior leader and their team to a different part of the company, and seeing what happens. Yes, this causes a lot of disruption, and has a surprisingly low success rate (for more, read HBR’s 2016 article Getting Reorgs Right. I remember reading that on a plane…whilst going through a reorg).

Experiments, like innovation, can be incremental or transformational

As you think about the set of experiments a company is making, you’ll want to make sure you’re covering the range of possibility. In other words, your experiment portfolio should be testing incremental changes as well as transformational ones. This, btw, is one of the bedrocks of innovation consulting — categorizing “bets” into categories.

A good example of incremental innovation is: Spotify making their Concerts feature easier to discover (Spotify team, this feature is still so buried! It’s probably for a reason, but come onnnn)

An example of transformational innovation is: Netflix completely pivoting from sending DVDs in the mail to standing up a world-class cloud infrastructure + licensing and content creation capability to own an entirely different market (streaming).

Some would argue Netflix isn’t transformational innovation. To placate those people, I will give yet another example: SpaceX. SpaceX has innovated in countless categories, including especially rocket construction, and their ability to have rocket boosters land again (to drive re-usability) is a classic example of a transformational innovation. No one else on the market (except for maybe those aliens we haven’t yet discovered near Alpha Centauri ;-) ) had gotten even close to landing (ehh…Ehhhh!) that maneuver. SpaceX stacked innovation on top of innovation to create something both functional and inspiring.

Winning companies “stack” innovations to grow their competitive moats

The companies we rely on today largely have stacked many successful experiments (“innovations”) on top of one another to create entirely new types of experience. Uber “stacked” innovations like ordering a car from your phone, leveraging GPS for both the driver and rider, assigning a rating in this particular type of two-sided marketplace, all while building and improving its matching algorithm. Any one of these innovations could have been its own product, but it was when they were all stacked together that it turned into an unstoppable beast of a company. … And when you stack innovations together in a user-friendly way, you earn the right and trust of your customers to give you a shot on your next set of innovations.

• Apple taking the trust and design love from Mac to iPhone. iPhone product trust helped lead to a successful introduction of the Apple Watch… and Air Pods…and now probably their XR headset in late ‘23 or ‘24… And before you know it a company owns the entire “product ecosystem” around a use case…

• Note: I learned about product ecosystems when I worked at Doblin, Monitor Deloitte’s innovation / design consultancy

It’s the companies that make bets…experiment…uncover innovations and then stack them that will find value across the life cycle of technological revolutions…

From the installation all the way through the deployment phase (see graphic above from Carlota Perez’s book). And for those who are trying to uncover the next big thing…the next company to invest in…the next company to go and work for, it can be helpful to think about:

1. Is this company working on an incremental or transformational innovation?

2. How mature is this market? Is it still in the installation period (characterized by market-making and many small or medium-sized competitors), or is it in the deployment period (characterized by market-taking and fierce competition among established players)

3. How susceptible is what I’m doing to technological disruption?

   1. If you’re working in a highly creative, strategic, revenue-generating job, you’re going to be harder to displace

   2. If you’re working in a job that is repetitive in nature and/or relatively low in complexity, bear in the mind the pressure to automate away the job itself

4. How can my current company (or companies I’m working with) take advantage of the market changes that are happening?

   1. …Maybe you can be the one to explore this, and run the experiments to find out how to take advantage.