What Innovation Means


In starting a new company, you have too many options.

I recently started thinking about whether to build an internet based company or not.  By that, I mean a company whose product or service is primarily delivered online - as opposed to by phone or in person such as a restaurant, grocery store, neighborhood bookstore, etc.

So much attention in business and society is focused, rightly, on the social, political, economic and ecological implications of connecting the entire world, for the first time, in a single instantaneous communication network.  We are thirty years into the commercial internet, thirteen into the mobile web, and not yet to the coming global frictionless payment networks. The world is still new and changing.


But, is it? I can’t help but look around and see there isn’t much new in online business models today, compared to what is yet possible in a more fully digital world.

Oversimplifying, there are three ways large internet companies make money online today:

  • Version A: Aggregate supply and demand

The internet is coming! and it enables anyone to connect to anyone else without meaningful marginal costs. Early online companies could meet core customer preferences better and cheaper than when they needed to have a physical location (1M books in a virtual warehouse!) or place an ad in the yellow pages to find customers.  Most importantly, geographical constraints were largely dropped as it didn’t matter whether the video rental store was down the block or across the country if the ordering was done online and the DVD arrived in the mail (even before online streaming).

This business model led to our biggest internet based companies, starting with basic consumer product categories and expanding in the last two decades to many smaller services1 (dating!) and products (homemade crafts!)

The thesis is still alive and well, and growing especially in fintech and edtech.  My kids love Outschool, bringing a globally available audience of school age children (demand) to niche educational suppliers such as the wonderful teacher from Calgary who recently taught eight students from four continents (including my eight year old) about ancient ciphers.

Arguably the model hasn’t really even started yet for University-level education and healthcare, which still require high cost in person service delivery for a small portion of the potential customer base (early models like Cousera are working to change this).  

Canonical Examples

  • Amazon (e-commerce more generally)
  • Netflix (over the top content more generally)
  • Dollar Shave Club  (direct-to-consumer more generally)
  • Apple’s App Store

What’s most interesting about this business model is that it is the most obvious transition point for companies that existed in the pre-internet world. If you are a successful local/regional operator whether you are a bank, brewery, or bakery, you can now aggregate larger demand for your products.  While many of these businesses can’t serve national or global customers as effortlessly as a digital product (it still costs money to ship cookies or lobsters across the country) the opportunity set has expanded for ambitious entrepreneurs.  

  • Version B - More data, better decisions

The second business model is more novel in that many applications are only possible when commercial interactions move from atoms to bytes.  Moving online, it turns out, creates a massive amount of structured data, which can be captured, analyzed and used to improve decision making.

The native business model here sells advertising in exchange for user attention.  Facebook (and all other social media companies) aren’t replacing advertisers such as newspapers, magazines, the TV networks and billboards because their content is better, they simply know more about each customer and can serve better targeted ads (and provide better performance metrics back to their customers, the advertisers).  Online behavior creates data, data enables targeting, targeting increases ad efficacy, improved efficiency shifts spending from larges advertisers.  That is the equation for trillions of dollars of wealth creation - from Google (90s) to Facebook (2000s) to Snapchat (2010s) to TikTok (2020s).

New applications of the model came when we moved to mobile phones, with business models like ridesharing.  Uber and Lyft are better than taxis because they have better data. The location data your phone and in the car, allow for better utilization of driver’s time, lower prices and better service.  

Canonical Examples

  • Facebook / Instagram (and other social media)
  • Affirm, Sofi, Opendoor (and other data driven online lenders)
  • Instacart (and other online delivery)

Where this goes from here becomes more interesting. The principle is widely acknowledged, and more data sets are created every day through mobile phones and the internet of things. Yet, consumers and regulators are increasingly concerned about privacy and the disparate impact of decisions made from data.  My sense: regulation will be a growing force in many consumer focused sectors but the commercial incentives for better decision-making will overcome the constraint (not always a good thing).

  • Version C  -  Centralize a cost center, deliver it more better/cheap through the internet

As data became more ubiquitous, a new services model developed native to the internet. We start entering the ‘software is eating the world’ thesis from Marc Andreessen. Enterprise companies noticed that millions of small businesses were replicating the same tasks, while incurring higher expenses than needed because they were sub-scale.  

What were once fixed costs such as racks of servers held in your building, moved to variable costs delivered over the internet (AWS).  Companies like Salesforce and, later, Twilio recognized that as the knowledge work increased more companies would want to off-load basic engineering tasks such as CRM development and customer communication tools.  Independent companies can invest in scaled solutions, then allow companies around the globe (even direct competitors) to tap into the same digitally delivered service..

Canonical Examples

  • Salesforce (and many SaaS models generally)
  • AWS/Azure (and cloud models generally)
  • Asana
  • Twilio

This version seems the most recent of the three, as bandwidth and trust are catching up to previous opportunities. Companies are moving more and more mission critical systems to the cloud to lower costs, sharpen their focus and reduce the internal complexity of their own business.

What’s Changing

Loosely speaking, those are the three models for making money on the internet.  Sometimes things get interesting in the combination of the three. Spotify and Netflix aggregate global demand for their content, but also a lot of data to then both make better decisions and lower costs for the products. Amazon now has business units across all three of these categories with their core marketplace, their advertising business, and AWS.

While all of the above create value, none strike me as profoundly innovative as business models. No disrespect to the leaders that have created incredible companies and to the technology breakthroughs that have enabled, for the first time, many of these opportunities.  But, it seems to me most of what we have see today is the online application of traditional business models using powerful new digital tools.

As a thought experiment, would Mr. Sears and Mr. Walton have a hard time wrapping their heads around the Amazon marketplace business?  They may be envious of the scale, of course, but the model of lowering costs by aggregating supply and demand would not be foreign to them.

More recently, enterprises as varied as Capital One and the Oakland A’s are famous for finding and analyzing new data sets to make better decisions before the internet came along.  

Over one hundred years ago, Mr. Edison developed the first centralized electricity plant (fun novel on the story here) to power buildings in New York, removing the localized cost center that was borne by each building, much the same as AWS does with servers today.  

Betting on One Big Change

Jeff Bezos has a famous quote about betting on things that will remain the same.  In twenty years, what are the chances will customers care about quick delivery? Very high, so any investment in that capability from Amazon is likely to be a good one.


I love that framework as a way of focusing on what Clayton Christensen would call sustaining innovations. It is a great way to improve an existing product or service.


But what about a big bet?  Amazon has also made those as well.  The initial online marketplace could be viewed as a long term bet that bandwidth will increase, and more people will spend more time online.  It seems obvious now, but when the company started in 1994 we had dial-up service and low online penetration. Bezos had conviction in one big idea - the internet would be big (great interview here from 1997 showing this).

And that is what stands out for the modern winners of the internet.  It isn't what they do, or even how the do it.  It is when they made their one big bet.  Go back through the list above considering the time when companies first got started.  

Was it obvious in 1999 that companies would trust online software for critical customer relationship management function? Here’s a CNN article from that same year announcing AOL’s proposal to move its 15M customers past dial-in connection speeds for the first time. Imagine the foresight to start Salesforce at that time.  Or, creating Netflix two years before then.  

Apple opened the App Store in July 2008, when there were less than 25M smartphones in the world.  

Successful internet companies are successful not because they break new ground in what the business does, but in when they make their big move.  It is the timing of the bet that matters.

And, it isn’t just limited to online businesses.  Look at some other examples of a focused long term bet on a trend that unlocks customer value thesis.

  • Tesla - The company (from the early days) was premised on the thesis that battery costs would drop to enable electric cars to be cheaper than gas. The founders believed customers want products that move us from a carbon based economy, way ahead of where we are today in the climate change debate.
  • Whole Foods - Organic was niche and expensive when founded in 1980. The founders bet on the long term movement toward healthier eating.  They plugged away for two decades, then all the sudden Walmart moved organic, the company created billions of dollars in market cap and was bought by Amazon.
  • Walmart -  founded in 1962, only six years after the Federal Highway Act of 1956 created the modern interstate highway system we have today.  Publicly, Low prices have always been a cornerstone of the company philosophy, but logistics mastery is what enables those prices.  The foresight to see how 20th century infrastructure would open up a national retailer was a big bet in the early 60s.
  • Microsoft -  in 1975, when the company started, it was far from clear software would be valuable to the average American. This was two years before the founder of DEC famously said “there is no reason anyone would want a computer in their home.”

Iconic businesses make one big bet on a long term market shift. It could be the progress of an enabling technology, the change in price for a product input, or the opening up of new and sustained customer demand for a ‘niche’ product.

Companies succeed once they make that large bet, build a brand and organizational capabilities that are inextricably linked to the market outcome they anticipate and then...either hope it plays out the way they expected or, better, bring about the change they anticipated in the first place.

My takeaway: develop a thesis on the world that, if correct, would result in a big opportunity.   What 25-year trend do you have conviction in that opens up new opportunities?

My answer is here