Investing Lessons from Mary Meeker's '95 Internet Trends Report

And Follow Ups on Spotify

“Yes, we are excited about Internet investment opportunities and the performance of the stocks, and yes, we are nervous about the valuation levels. In order to mitigate valuation/execution risk, investors must take a long-term, portfolio approach, and we recommend building positions over time. While in the short term, making money in Internet-related companies has been like shooting fish in a barrel, over the long run, selectivity and patience will be key as companies/technologies/stocks come and go” — Morgan Stanley, 1995

In February 1995, Morgan Stanley’s Mary Meeker and her team published their seminal report on the internet wave and the opportunities that will arise from the Internet “Big Bang”. The report is a tour de force and worth revisiting for tech investors and aspiring financial analysts. Not only does ~300-page report provide valuable insight into the frameworks of one of the best internet investors/thinkers of the last ~40 years but the quality of writing is impeccable.

The full report can be found here: Internet Trends 1995 and I summarized a few of my Lindy Investing Lessons and Takeaways from the report below.

Lindy Lessons and Takeaways Mary Meeker’s 1995 Report

Don Valentine Checklist for New Markets & Innovator’s Dilemma

When looking for investment ideas in new markets, we default to our favorite maxims from Don Valentine of Sequoia Capital, who is known as one of the toughest and smartest technology venture capitalists in Silicon Valley. Don follows several simple rules in choosing early-stage tech investments:

  1. Find “monster” markets that can be really big, like the Internet

  2. Find good technology and good technologists who can stay ahead of competitive threats

  3. Find outstanding leaders/management teams that can drive the technologies and markets forward

  4. Buy companies, not products, and try to find companies that have achieved critical mass with their products — or can achieve it, and can create some form of “barriers to entry”

None of the values of the checklist are new to investors but an aspect that investors consistently seem to take for granted is the third point. In large and rapidly growing markets, if the business has a strong economic model (i.e., on a unit economics basis they can generate cash flows), management’s sole responsibility is to allocate excess cash flows in areas of high returns on capital (S&M expansion, R&D, buybacks, acquisitions, etc). Companies that cannot identify areas to create or build upon on their existing advantages will likely have a terminal value that is likely far less in value than investors expect, this is especially tech companies.

The challenge is that while traditional capital allocation levers (buybacks, M&A, dividends, etc.) are easy to quantify and appreciate in financial statements, the capital allocation decisions made by growth/tech companies and their management teams (R&D, S&M build-out) are far more challenging to understand due to the limitations of GAAP accounting. For this reason, managerial incentives and shareholder expectation alignment are crucial for driving long-term shareholder value in technology companies.

Over-weighting one part of the equation over the other will result in the destruction of shareholder value. Microsoft is a great example of a company that has found the right balance between incentivizing management to invest in the product/platform and traditional financial metrics (revenue growth, margin).

Meeker has an apt case study of what happens in the technology industry when the balance is not struck appropriately.

With each step in the evolution of computing (Figure 1.2) from mainframe to minicomputer to PC or from centralized to decentralized to networked computing more and more people gained control of their computing capabilities. Smaller, cheaper computing systems also made this control more economical. From a business perspective, the overriding theme was that no company that dominated one generation of computing managed to dominate the next; each became wedded to its legacy systems and cash flow

If you want a more recent example of this cycle of companies creating excess profits only to be later disrupted by their inability to keep pace with the new industry technology/standards, look no further than IBM and Intel.

Framework: Latent Demand Unlocked by New Technology & Standards

One of the key investment frameworks that can be abstracted from Meeker’s analysis is: incremental advancements in technology over a relatively short period of time can unlock massive latent consumer demand. When there is the right confluence of technology changes and consumer behavior, the growth and adoption of the technology/business model can be asymptotic.

What’s going on? We call it “The Great Communications Backfill Opportunity.” Over the last ten years, Bill Gates of Microsoft has been the most vocal repeater of the mantra “A PC on Every Desktop.” Now we are getting to the point where PCs, while not on every desktop (or every lap)…Spreadsheets, word processors, and games are cool, useful, and important, but there ain’t nothing like communicating with a PC (whether via e-mail; transferring a file; interacting with someone who shares your interest in migration patterns of Canadian Geese; or obtaining information on anything from the weather to the best beer halls in Munich). Or yes, conducting good old-fashioned commerce in a new way

In the dot-com boom, latent consumer demand to send/receive information in significantly less frictionless was unlocked by a series of events:

  1. Personal Computer: the decreasing cost of computing led ~150mm consumers adopting PCs over nine years

  2. Growth of commercial ISPs and infrastructure (over) investment

  3. Development of browsers (Mosaic)

  4. Adoption of internet standards (WWW, HTTP, etc.)

(This is my horribly short summary for summary purposes. I’ve linked to more complete sources at the end of the post)

The non-linear nature of this framework is important to internalize. There is no one singular event or company that serves as the catalyst for an inflection but the change is akin to watching a dam break. No single droplet of water can be blamed but they are all responsible for the forceful wave that is released.

This framework bore out in a nearly identical fashion ~15 years later with the explosion of the ride-sharing. Here is Floodgate’s Mike Maples discussing their investment in Lyft.

Technology inflections involve exponential improvements in the price/performance of technologies like computation, sensor accuracy, bandwidth, and the like.

When Floodgate invested in Lyft (Zimride at the time), one of the inflections was that GPS locators in smartphones were becoming orders of magnitude more accurate, which meant they would be able to locate riders and drivers much more precisely.

Adoption inflections involve nonlinear changes in the adoption rate of a technology.

Returning to the Lyft example, in 2010 we believed it was likely smartphone adoption would increase exponentially. This meant we could imagine enough drivers and riders would have smartphones to make it possible to build a scalable transportation network in the future — Mike Maples

Floodgate’s insights around consumer adoption of the smartphone and technology advancements in the devices led to unlocking massive consumer demand for a new and better mode of transportation (ridesharing is one of many business models that benefitted from the inflections that Maples identified).

Risk/Return: The Power-Rule of Growth Investing

The power rule of venture investing is just as prevalent in public markets, especially industries where the rate of technological change is rapid. Here is Meeker in 1995 cautioning investors about the dispersion of outcomes and power-law of tech investing:

Consider the last big IPO boom, which was related to the development of the PC industry. In those 15 years, 581 tech companies went public and created more than $240B in net market capitalization. At the end of 1994, 45% of the companies traded at prices below their IPO price, and only 16 ten- bagger stocks had been created (or 3% of the companies that went public in the 15-year time period). In addition, only 17 of the top 100 PC software companies in 1981 were still on the list in 1995.

The same trend bore out when the dot-com boom inevitably burst. Here is Meeker in 2001 retro’ing the dot-com bubble.

Investing regardless of stage usually follows a power law, where a small fraction of the companies creates an outsized portion of the value in the sector over the long-term.

Picking the right industry or trend is not good enough to drive outsized returns, never has been and never will be. Being judicious about the companies you pick to ride the wave is far more important than identifying the wave before others. With the technology boom-and-bust cycle, the lesson always seems the same: invest in good economic models with strong management teams that are capable of riding the trend for 10 years+.

For the losers, the reasons for poor performance were typically related to competitive pressures and low barriers to entry, overhyped expectations, technology obsolescence, and poor management execution.

Follow-Up on Spotify

Taking Share vs. Gross Minutes

As a response to my previous post, one reader provided the following feedback/question.

Right. But if they don’t reduce music streaming time - but rather just add podcast listening time - they won’t reduce costs and may even increase them depending on how much they have to pay for big name talent.

Their point highlights an implicit assumption I made in the previous analysis, the gross minutes a user spends on audio stays fairly flat and Spotify just continues to take a share of a fairly fixed pie. The best available data suggests that the growth in total ecosystem listening is growing ~3% annually — but this provides no signal really in terms of what usage on Spotify looks like…so more work to do be done here.

Lydia Polgreen on Spotify’s Podcasting Opportunity

Peter Kafka hosted Spotify’s Head of Content for Gimlet recently, where they discussed a wide range of topics. My highlights from the conversation are below

Appendix: History of the Internet

Ian Peter's History of the Internet

The History of the Internet in a Nutshell

Internet History Podcast