TensorFlow and Monetizing Intellectual Property

Posted on

Ten years ago Bill Gates suggested that open source software was the province of “modern-day sort of communists” whose views on intellectual property were hopelessly outdated:

The idea that the United States has led in creating companies, creating jobs, because we’ve had the best intellectual property system — there’s no doubt about that in my mind, and when people say they want to be the most competitive economy, they’ve got to have the incentive system. Intellectual property is the incentive system for the products of the future.

Gates’ perspective was understandable: he had built Microsoft into the biggest company in technology and one of the biggest in the world by, for all intents and purposes, selling licenses to text. Sure, that’s a dramatic over-simplification of Windows and the other software Microsoft sold, but that didn’t change what a seachange the Redmond-based company seemed to represent: one where the pure expression of ideas could make you the richest person in the world. Yet those antediluvian open-source zealots wanted to simply give it all away.

The Open-Sourcing of TensorFlow

Microsoft is still a big company — their market cap was $427 billion at yesterday’s market close — but an even bigger company today is Alphabet ($506 billion), which has a decidedly different approach:1 earlier this week its Google subsidiary announced it was open-sourcing TensorFlow, its formerly proprietary machine learning system. From the official Google blog:

Just a couple of years ago, you couldn’t talk to the Google app through the noise of a city sidewalk, or read a sign in Russian using Google Translate, or instantly find pictures of your Labradoodle in Google Photos. Our apps just weren’t smart enough. But in a short amount of time they’ve gotten much, much smarter. Now, thanks to machine learning, you can do all those things pretty easily, and a lot more. But even with all the progress we’ve made with machine learning, it could still work much better.

So we’ve built an entirely new machine learning system, which we call “TensorFlow.” TensorFlow is faster, smarter, and more flexible than our old system, so it can be adapted much more easily to new products and research. It’s a highly scalable machine learning system — it can run on a single smartphone or across thousands of computers in datacenters. We use TensorFlow for everything from speech recognition in the Google app, to Smart Reply in Inbox, to search in Google Photos. It allows us to build and train neural nets up to five times faster than our first-generation system, so we can use it to improve our products much more quickly.

We’ve seen firsthand what TensorFlow can do, and we think it could make an even bigger impact outside Google. So today we’re also open-sourcing TensorFlow. We hope this will let the machine learning community — everyone from academic researchers, to engineers, to hobbyists — exchange ideas much more quickly, through working code rather than just research papers. And that, in turn, will accelerate research on machine learning, in the end making technology work better for everyone.

Machine learning is super important to Google; just a couple of weeks ago, on Alphabet’s Q3 earnings call, Google CEO Sundar Pichai stated in his opening remarks, “I also want to point out that our investments in machine learning and artificial intelligence are a priority for us”, and followed that up with a series of examples where machine learning was serving as a differentiator for Google. Pichai later added, in response to a question:

Machine learning is a core transformative way by which we are rethinking everything we are doing. We’ve been investing in this area for a while. We believe we are state-of-the-art here. And the progress particularly in the last two years has been pretty dramatic. And so we are thoughtfully applying it across all our products, be it search, be it ads, be it YouTube and Play et cetera. And we are in early days, but you will see us in a systematic manner, think about how we can apply machine learning to all these areas.

At a superficial level, this doesn’t make sense: if machine learning is core to Google’s future, then what is the point of giving it away? Does the company not care about making money? Are they — gasp — communists, or more charitably, as former head of Google’s Webspam team Matt Cutts put it, “releasing technology so that the entire world can benefit, not just Google”?

To be sure, there is a lovely PR angle to this news, but I think Google’s thinking is a lot more strategic than that. Open-sourcing TensorFlow makes a ton of sense, and the lessons as to why are broadly applicable.

Differentiating Software

Think carefully about what differentiates today’s technology companies. To take a few of the most prominent examples:

  • Apple’s devices are differentiated first and foremost by their software, but the company makes money by selling hardware. Or, to be more precise, they make money by selling devices at scale that integrate software, hardware, and services, and to do that requires not simply an operating system but also a world-class industrial design team and an all-but-impossible-to-replicate supply chain that stretches from well over 200 suppliers all over the world to almost 500 Apple retailer stores and tens of thousands of resellers
  • Amazon isn’t simply a website but also a massive logistics network that connects tens of thousands of vendors to customers via nearly 100 distribution and sortation centers in North America alone, while AWS probably has nearly another 100 data centers with millions of servers and an increasingly rich ecosystem of partners dedicated to getting companies onto AWS’ cloud
  • Facebook’s value comes not from its software but from the fact that the social network has over 1.5 billion monthly active users, over a billion of whom use the service daily, plus three other services (WhatsApp, Messenger, and Instagram) with active user bases numbering in the (high) hundreds of millions; all of those users are connected to each other

The examples go on-and-on: companies that are built on software but differentiated by a difficult-to-replicate complement to said software. And this, I think, is the way to understand Google’s decision to open-source TensorFlow.

Google’s Machine Learning Advantage

I’m hardly qualified to judge the technical worth of TensorFlow, but I feel pretty safe in assuming that it is excellent and likely far beyond what any other company could produce. Machine learning, though, is about a whole lot more than a software system: specifically, it’s about a whole lot of data, and an infrastructure that can process that data. And, unsurprisingly, those are two areas where Google has a dominant position.

Indeed, as good as TensorFlow might be, I bet it’s the weakest of these three pieces Google needs to truly apply machine learning to all its various business, both those of today and those of the future. Why not, then, leverage the collective knowledge of machine learning experts all over the world to make TensorFlow better? Why not make a move to ensure the machine learning experts of the future grow up with TensorFlow as the default? And why not ensure that the industry’s default machine learning system utilizes standards set in place by Google itself, with a design already suited for Google’s infrastructure?

After all, contra Gates’ 2005 claim, it turns out the value of pure intellectual property is not derived from government-enforced exclusivity, but rather from the complementary pieces that surround that intellectual property which are far more difficult to replicate. Google is betting that its lead in both data and infrastructure are significant and growing, and that’s a far better bet in my mind than an all-too-often futile attempt to derive value from an asset that by its very nature can be replicated endlessly.

The Android Example

Indeed, Google has already demonstrated that this approach can be devastatingly effective. Gates was right to fear the open-source threat to Windows: in the smartphone era Google took Microsoft’s former position as the default operating system for the masses by open-sourcing Android. I absolutely believe that Android would not have achieved the dominant position that it has without that step, and Google was able to do so because its business model of advertising was a complement to its software, not because it sold software itself.

It’s fair to object that open-sourcing Android ensured it would never be a real money-maker for Google; I’ve made that case myself. But remember, Android was originally intended as a defensive measure for Google’s search business. And, from that perspective, it wildly succeeded.

I suspect open-sourcing TensorFlow will have a far more positive effect on Google’s bottom-line. Google is approaching machine learning from a position of strength: the company already has the most data and the most imposing infrastructure, and as noted open-sourcing TensorFlow accelerates the removal of the primary limitation to leveraging that advantage: the quality of the system itself.

Broader Lessons

There’s a parallel to be drawn to my piece last week about Grantland and the (Surprising) Future of Publishing. The fundamental nature of the Internet makes monetizing infinitely reproducible intellectual property akin to selling ice to an Eskimo: it can be done, but it better be some really darn incredible ice, and even then the market is limited. A far more attainable and sustainable strategy is to instead focus on monetizing complements to said intellectual property, resulting in an outcome where everyone wins: intellectual property consumers, intellectual property copiers, and above all intellectual property creators.

  1. Microsoft has since embraced open source ↩

Grantland and the (Surprising) Future of Publishing

Posted on

It’s dangerous, I suspect, to draw too many lessons from the ignominious end of Grantland, the high-brow sports and culture site that ESPN shuttered this weekend, several months after parting ways with Bill Simmons, the site’s founder and editor-in-chief. Much of what transpired seems to have clearly been personal, not only the rancorous way in which it ended but also how it began: was ESPN every truly committed to a brand-building endeavor that didn’t even have the ESPN name, or was Grantland a pet project meant to keep Simmons, the most influential and famous sportswriter of his generation, in the fold?

Conventional wisdom is that both are worse off for having split: ESPN for having lost said writer and, in the end, a truly remarkable online magazine, while Simmons lost access to ESPN’s massive audience. Still, there is no question that the numbers didn’t add up: Deadspin reported in May that Grantland had 6 million unique visitors in March, a relatively meager sum that in no way shape or form could support a team of over 50 writers, editors, and back-end staff, even if Grantland had been much more aggressive in monetizing through advertisements (the site usually carried at most one banner ad plus a “You might be interested in…” block at the end of articles). And, so, as Chris Connelly, who replaced Simmons as interim editor-in-chief, noted in an interview with Sports Illustrated:

When you are doing a site that you understand is not making money, you kind of understand when times get challenging or there is a new economic climate, you will be scrutinized very closely. I think the site continued to do fantastic editorial, for which I want to be sure not to take credit. That was the product of the editors and writers who were there every day of the week. But in this economic climate you will be very closely scrutinized if you are not a money-making operation.

Indeed, Grantland was not the only cut at ESPN: the network also recently laid off around 300 people in the face of rising costs and declining subscriber numbers; there’s not much room for brand-building when you’re already showing valuable employees the door.

Grantland’s Value

There’s no question that Grantland was an editorial success, at least in terms of quality. The site garnered three National Magazine Award nominations, won a primetime Emmy, and was nominated for and won a number of Sports Emmys, Webby Awards, Eppy Awards, and more.1 Speaking personally, the site was one of only a handful I visited directly daily; if an article was on Grantland I presumed it was worth reading.

To put it another way, Grantland was a “Destination Site.” I first defined this term in an article this spring about Facebook’s Instant Articles, and noted:

It’s really hard to become a destination: you need compelling content of consistently high quality. Notice, though, that that is precisely the opposite of what most online publications have focused on: in their race for ever more content and ever more clicks most publications have lowered their quality bar.

That right there is the rub, as Todd VanDerWerff lamented on Vox:

The problem is scale. A larger, general-interest site can’t be built purely atop longform, because longform takes time — both for writers to produce and readers to read. Therefore, as both Buzzfeed and Gawker realized early on, well-done longform could be the steak, but it couldn’t be the meal. (Grantland perhaps realized this too late.)

The non-steak portion of the “meal” that VanDerWerff refers to are those quick-hitting posts that, to use his employer (Vox.com) as an example, embed and summarize an interview with a pop singer, do the same embed and summarize routine about a $1500 sandwich, and rewrite a news story about Jon Stewart going to HBO that countless other sites covered as well. And that’s just (a small selection from) Tuesday! In fact, the majority of stories — including the most popular ones of the day — could have been written by anyone anywhere.

As a reader this is frustrating: Vox stripped down to the political and policy coverage that is its raison d’être would almost certainly rate as a destination site for me; instead, faced with a deluge of rewrites and summarized videos, I miss most of the good stuff save for what I stumble across on social media. I’m certainly not going to waste my time wading through the filler on the homepage.

Grantland’s Worth

Still, the fact remains that Vox is by all accounts thriving while Grantland, which eschewed filler, is dead. Moreover, the only destination sites that really seem to be working either have massive brand equity that is being leveraged into subscriptions (The New York Times, The Wall Street Journal, The Financial Times), or are tiny one-person operations that leverage the Internet to keep costs sustainably low while monetizing through small-scale native advertising (e.g. Daring Fireball and the now-retired Dooce.com, for example) or subscriptions (e.g. The Timmerman Report and the site you’re reading). It certainly seems that the lesson of Grantland is that there is no room in the middle: not enough scale for advertising, and costs that are far too high for a viable subscription business.2

The Publishing Curve Revisited

But remember my initial warning: it’s possible to read too much into Grantland. Last year I wrote about Publishers and the Smiling Curve, characterizing publishers as being akin to original equipment manufacturers assembling computers and phones at cost, while profits flowed to aggregators on one side (Facebook, Google, etc.) and to “stars and focused publications” on the other.

Publishers and the Smiling Curve
Publishers and the Smiling Curve

I certainly think I got the aggregator part right (several months later I generalized the concept to Aggregation Theory), but it wasn’t clear in the article — or ultimately to me — how exactly the “stars and focused publications” would profit outside of one-man operations.

Simmons, though, just as he pioneered online sports writing, may be leading the way once again in demonstrating how writing can be monetized: by not even trying.

Writing As Lead Generation

Simmons is pursuing two endeavors post-ESPN: he’s creating a television show for HBO Now (although it will reportedly air on HBO proper as well) and he’s recording The Bill Simmons Podcast. The business models — and thus the incentives — for both couldn’t be more different than those for successful online publishers:

  • HBO Now, like all subscription services, has to surmount the far stiffer challenge of convincing customers to get out their credit cards, not just click a link (or, in television terms, watch an ad-supported show). This changes the focus from common-denominator low-cost fare like reality TV or game shows (or video summaries or news rewrites) to focused, expensive “destination shows” that fewer people watch, but those who do care intensely.

    One show in one niche isn’t enough though; the best balance between revenue and costs is found through bundling: people may be willing to pay a lot for one show and a little for several others, which means a well appointed bundle can ultimately get more revenue per customer from a wider base of customers. In the case of Simmons, he brings a younger male audience that cares about sports, and who may also be interested in, say, Game of Thrones or Hard Knocks, or perhaps a bit of comedy from, say, Jon Stewart; it’s the collection that makes said audience willing to pay $15/month, even though they may have been willing to pay Simmons alone less.

    Why, though, does Simmons have those fans in the first place? Because of his writing. The flipside of writing being hard to monetize is that it is the most digestible and sharable medium, allowing folks to accrue large audiences that they can leverage elsewhere.

  • Podcasts run on an advertising model, but said advertising is far more valuable than display ads in particular. They are truly native: Simmons, like most podcast hosts, reads the ads himself in the middle of his podcast, meaning they are more likely to be heard and are more engaging to boot.

    The trouble with podcasts is that they are difficult to grow: while text can be shared and consumed quickly, a podcast requires a commitment (which again, is why advertising in them is so valuable). Simmons, though, by virtue of his previous writing, is already averaging over 400,000 downloads per episode.3 Podcast rates are hard to come by, but I’m aware of a few podcasts a quarter the size that are earning somewhere in excess of $10,000/episode; presuming proportionally similar rates (which may be unrealistic, given the broader audience) The Bill Simmons Podcast, which publishes three times a week, could be on a >$6 million run rate, which, per my envelope math in the footnote above, could nearly pay for a 50-person staff a la Grantland.

To be sure, as I noted above, Grantland was certainly much more expensive, and it’s not clear just how far podcasting can scale (but I think that’s only a matter of time), but Grantland actually had an entire stable of podcasts, several of which were quite popular because they featured writers whose writing was popular.4 Just as with HBO, it turns out writing is great lead generation for an actual monetizable business.

The Grantland Lesson

To be sure, it’s tempting to pull a “That’s Fine for Bill”; the guy has been writing online for eighteen years (and, technically, he’s not writing now).5 It’s a fair point but I think there’s room for another, equally compelling one: too much of the debate about monetization and the future of publishing in particular has artificially restricted itself to monetizing text. That constraint made sense in a physical world: a business that invested heavily in printing presses and delivery trucks didn’t really have a choice but to stick the product and the business model together, but now that everything — text, video, audio files, you name it — is 1’s and 0’s, what is the point in limiting one’s thinking to a particular configuration of those 1’s and 0’s?

In fact, it’s more than possible that in the long-run the current state of publishing — massive scale driven by advertising on one hand,6 and one-person shops with low revenue numbers and even lower costs on the other — will end up being an aberration. Focused, quality-obsessed publications will take advantage of bundle economics to collect “stars” and monetize them through some combination of subscriptions (less likely) or alternate media forms. Said media forms, like podcasts, are tough to grow on their own, but again, that is what makes them such a great match for writing, which is perfect for growth but terrible for monetization.

That’s why the lesson to be learned from Grantland may be the exact opposite of what it seems: the problem isn’t that ESPN spent too much money on a web site that couldn’t monetize, it’s that the web site should only have been step one to a multi-media endeavor that converted visitors to fans willing to invest time in formats that can actually pay the bills.7

  1. Via the afore-linked Deadspin article ↩

  2. Some back of the envelope figures:

    Presuming a cost per employee of $100,000 (which, including benefits, is probably on the low side):

    • A one person site needs $100,000 in revenue, which means 1,000 subscribers at $100/year
    • A two person site needs $200,000 in revenue, which means 2,000 subscribers at $100/year
    • A 50 person site (ignoring Simmons’ reported $5 million salary, plus the fact the site included former Pulitzer Prize winners) needs $5 million in revenue, which means 50,000 subscribers at $100/year. Add in Grantland’s other expenses and the number is likely double that

    Keep that $5 million figure in mind ↩

  3. I’ve heard the number now is actually closer to 500,000 ↩

  4. The podcast point is the most legitimate of all of Simmons’ complaints about ESPN: there was clearly minimal effort made to monetize the B.S. Report or the other Grantland podcasts ↩

  5. This is also the part where I note that the reason I and so many other online writers revere him, despite his warts, is that he was a true pioneer: no one had made a career on the Internet before him, and now lots of us have ↩

  6. Indeed, I’ve not only written repeatedly that a traditional advertising model only works at scale, but also argued that Facebook and Google are on pace to capture ever more of it: the two Internet giants have better user data, better tracking, and better inventory, and relatively unlimited inventory (at least compared to a medium like TV, which certainly took a chunk of advertising revenue from traditional publishers, but only a chunk). ↩

  7. One more thing: I don’t think linear TV is that format. Fans want to read/see/listen to stars, but time is a constraint; on-demand a la HBO Now or podcasts is a much better fit ↩

Exponent Podcast: The User Experience

Posted on

On Exponent, the weekly podcast I host with James Allworth, we discuss the importance of the user experience in not only avoiding disruption but also in building dominant companies.

Listen to it here.

In Defense of The New York Times

Posted on

Something rather amazing happened this week.

On Monday, two months after The New York Times wrote a brutal exposé on Amazon’s workplace culture, Jay Carney, former White House press secretary for President Barack Obama and current Senior Vice President for Global Corporate Affairs at Amazon, wrote a blistering piece on Medium entitled What The New York Times Didn’t Tell You:

When the story came out, we knew it misrepresented Amazon. Once we could look into the most sensational anecdotes, we realized why. We presented the Times with our findings several weeks ago, hoping they might take action to correct the record. They haven’t, which is why we decided to write about it ourselves.

The Times got attention for their story, but in the process they did a disservice to readers, who deserve better. The next time you see a sensationalistic quote in the Times like “nearly every person I worked with, I saw cry at their desk”, you might wonder whether there’s a crucial piece of context or backstory missing — like admission of fraud — and whether the Times somehow decided it just wasn’t important to check.

It was really something: Carney’s accusations were strong, disturbingly detailed (Amazon presented presumably confidential employee performance data), and perhaps most curiously of all, out of nowhere.1 The fact the response was on Medium was interesting as well: the placement helped guarantee attention from a certain segment of the public without tying it too explicitly to Amazon.

Then, a few hours later, something even more surprising happened: the Executive Editor of The New York Times, Dean Baquet, responded, also on Medium:2

In response to your posting on Medium this morning, I want to reiterate my support for our story about Amazon’s culture. In your posting — as well as in a series of recent email exchanges with me — you contested the article’s assertion that many employees found Amazon a tough place to work.

As the story noted, our reporters spoke to more than a hundred current and former employees, at various levels and divisions, over many months. Many, including most of those you cited, talked about how they admired Amazon’s ambitions and urgency even as they described aspects of the workplace as troubling…and any reading of the responses leaves no doubt that this was an accurate portrait.

Both pieces were remarkable for reasons that had little to do with their content.3

Amazon Goes Public

The importance of Amazon’s response is obvious: unlike days of old, when corporations or individuals in the news had to resort to letters to the editor (which may or may not have been printed) and angry calls to the editor-in-chief, Amazon can go straight to the public with their complaints; it may sound cliché to say that “everyone is a publisher” but for the fact it’s true. Moreover, like anything else on the Internet, Amazon’s response was immediately available to everyone in the world: we take that for granted today, but compared to not that long ago when distribution required printing presses and delivery trucks this is truly an astounding development.

The New York Times Response

Even more important, though, was the fact that Baquet responded, and on the same (small-m) medium as Amazon to boot. An unfortunate side effect of owning said printed presses and delivery trucks was that newspapers held themselves as the oracles of truth, none moreso than The New York Times. Consider the motto printed on every paper: All the News That’s Fit to Print; I criticized the mindset behind this motto in Why BuzzFeed Is the Most Important News Organization in the World:

It’s important to appreciate that this was more than just a slogan and [the front-page meeting was more than just a] meeting; there are important assumptions underlying this conceit:

  • The first assumption is that there is a limited amount of space, which in the case of a physical product is quite obviously true. Sure, newspapers could and did change the length of their daily editions, but the line had to be drawn somewhere
  • The second assumption is that journalists, by choosing what to write about, are the arbiters of what is “news”
  • The third assumption is that the front page is an essential signal as to what news is important; more broadly, it’s an assumption that editors matter

My point then was that none of these assumptions held on the Internet: there is an unlimited amount of space, news can come from anywhere or anybody, and that the front page is a lot less important in the age of social media. And, I noted, as long as The New York Times held to these assumptions, they would slowly but surely fall behind.

This is why Baquet’s response is so significant:

  • His response was public, not private, and why not: an extra web page is free
  • His response was on the same medium as Carney’s post, which is fine, because Medium is just as accessible and potentially newsworthy as nytimes.com
  • His response was a part of a conversation, not a pronouncement

Baquet actually made this conversation point himself in an interview with Kara Swisher and Peter Kafka at Recode’s Code/Media conference in September:

The construct was, “It was true, it was important, we made the case there was something anomalous about Amazon.” And most importantly, and this to me is what the best journalism does, it sparked vibrant debate about the workplace.

A vibrant debate about the workplace? That is journalism? Not printing the truth?

The Nichification of the New York Times

In fact, this was one of a whole host of very interesting things Baquet said in the interview. Several minutes prior he had attacked another journalism shibboleth, that of the necessity of a “wall” between the newsroom and the business side of the paper.

I think our relationship with people outside the newsroom is different…in the world I grew up in, and in the world that created The New York Times…I think that that rule that there was a big fat wall between the news room and everybody else doesn’t make sense anymore in the modern era…I think that we now understand that that’s sort of nuts…I think that that was a comfortable position when we had a 30% profit margin, but it went on too long…

I think of myself as primarily the executive editor whose job it is to ensure the quality and the integrity of the report. But I also think of myself as somebody whose job it is to preserve The New York Times which means I do think about advertising, I do think about The New York Times as a business. That does not mean that I drop the wall and sell ads. But it does mean that I think about the whole of the enterprise.

Regular readers will know just how important I think this is. I wrote last month in Popping the Publishing Bubble:

Publishers going forward need to have the exact opposite attitude from publishers in the past: instead of focusing on journalism and getting the business model for free, publishers need to start with a sustainable business model and focus on journalism that works hand-in-hand with the business model they have chosen. First and foremost that means publishers need to answer the most fundamental question required of any enterprise: are they a niche or scale business?

What is exciting about the Amazon story is that, at least according to Baquet, it came from embracing the nichification of The New York Times.

I think that people know that the Amazon story came from The New York Times. I think my job is to ensure that the percentage of stories we do is very different. My job is to do as many Amazon stories as possible and to do fewer and fewer of the traditional stories that don’t work as well as the bundle disintegrates. My job is to produce a lot of Amazons.

In other words, the job of The New York Times is no longer to produce “All the News That’s Fit to Print”; rather, it is to invest in stories that make a difference — stories that start a conversation — and trust that readers will be willing to pay for quality. The content follows from the business model.

Will the Niche Model Work?

Of course, while this all sounds good on paper, the proof is in the numbers. And, it turns out, the numbers are pretty encouraging. Baquet wrote earlier this month:

We recently passed one million digital-only subscribers, giving us far more than any other news organization in the world. We have another 1.1 million print-and-digital subscribers, so that in total, we have more subscribers than at any time in our 164-year history. Many news organizations, facing competition from digital outlets, have sharply reduced the size of their newsrooms and their investment in news gathering. But The New York Times has not.

In the piece Baquet lists the qualifications of the New York Times’ reporters: a Yale-educated lawyer covering the Supreme Court, a former soldier covering abandoned chemical weapons in Iraq, a former Federal Reserve employee who wrote about income inequality. Sure, he’s almost certainly cherry-picking, but the broader point about a focus on quality and impact stories supported by readers directly is very much spot-on: it’s the exact approach niche publications need to pursue.

To be clear, I don’t use the word “niche” as an insult, and it would be absurd to do so: the New York Times remains the most influential publication in the world. Rather, I’m referring to the choice all publications must make: to go broad and cross-platform with a goal of maximizing readership and monetizing through advertising, or to instead focus on maximizing revenue from the customers who actually care about your brand. To be niche.

Encouragingly, there is more evidence beyond this interview that The New York Times has embraced this approach: the company released a strategy memo earlier this month that made clear the company’s goal was to double its digital revenue (from $400 million to $800 million) primarily through a niche strategy:

“Many of our competitors focus primarily on attracting as many uniques as they can with a view to building an advertising-only business,” the memo said, referring to unique visitors to websites. “We see our business as a subscription service first, which requires us to offer journalism and products worth paying for.” That engagement, it said, will also help attract advertisers.

Twelve percent of Times readers, the memo said, deliver 90 percent of its digital revenue. “To double our digital revenue, we need to more than double the number of these most loyal readers,” it said. “We will need to develop them increasingly from younger demographics and international audiences.”

It’s not certain this strategy will succeed, to be sure, but it is a strategy that is at least coherent, and one that I celebrate.

Journalism and the Search for Truth

The fact of the matter is that The New York Times almost certainly got various details of the Amazon story wrong. The mistake most critics made, though, was in assuming that any publication ever got everything completely correct. Baquet’s insistence that good journalism starts a debate may seem like a cop-out, but it’s actually a far healthier approach than the old assumption that any one publication or writer or editor was ever in a position to know “All the News That’s Fit to Print.”

I’d go further: I think we as a society are in a far stronger place when it comes to knowing the truth than we have ever been previously, and that is thanks to the Internet. It’s a good thing that Amazon can post to Medium, and it’s healthy that Baquet responded. My alma mater the University of Wisconsin declared back in 1894:

Whatever may be the limitations which trammel inquiry elsewhere we believe the great state University of Wisconsin should ever encourage that continual and fearless sifting and winnowing by which alone the truth can be found.

The New York Times doesn’t have the truth, but then again, neither do I, and neither does Amazon. Amazon, though, along with the other platforms that, as described by Aggregation Theory, are increasingly coming to dominate the consumer experience, are increasingly powerful, even more powerful than governments.4 It is a great relief that the same Internet that makes said companies so powerful is architected such that challenges to that power can never be fully repressed,5 and I for one hope that The New York Times realizes its goal of actually making sustainable revenue in the process of doing said challenging.

So You Want to Change the World

Note that I haven’t said much about the Amazon article in question; in fact, as I wrote at the time, the article bugged me quite a bit not because of its description of the environment (which, according to both my friends who work there and the company’s general reputation in the Seattle area, was broadly correct) but rather the article’s dismissive tone towards what Amazon has accomplished: the company is fundamentally changing enterprise IT, with all the knock-on effects that entails from disrupting tech companies to real estate to venture capital; changing commerce; and even changing how we consume ideas. There is an argument to be made that this sort of impact doesn’t happen if you only work 9-5.

That’s why I’m an optimist though: all kinds of people did make that argument — this site, former employees, even Amazon itself — and the net result is that we are collectively closer to the truth than we were before that article. So it was good journalism, and given the increasing importance of technology, we as an industry should embrace it: you can’t claim you want to change the world and not appreciate that the more ambitious your goals, the more necessary the challenges to exactly what you’re trying to change, and how you’re trying to change it.

  1. Why is Amazon resuscitating this story? Is it hurting recruiting? Is there another story coming from The New York Times soon (my guess)? Do they simply want to send a message to journalists generally? ↩

  2. Unfortunately, Carney disabled the display of responses, which is pretty weak  ↩

  3. Carney responded to Baquet’s response here  ↩

  4. I still worry more about government’s collecting data though: they are the only institution capable of throwing you in prison ↩

  5. This is why my number one concern is about regulating access to the Internet; the Great Firewall is the most frightening thing in the world ↩

Venture Capital and the Internet’s Impact

Posted on

Much has been written of the difficulty in building “another Silicon Valley.” To be sure, many countries and regions have tried, seeking to assemble the perfect mix of willing investors, eager entrepreneurs, and ready-made markets that will produce the sort of self-perpetuating ecosystem that will lift the region, country, nay, the world to a new level of prosperity and modernity.1 The problem is that like most real-life systems Silicon Valley is non-linear: it is impossible to break it down into component parts that can be reproduced, and no one can know for sure what a small change in inputs will mean for the outputs.

Moreover, one of the most important stories of the last several years is how the structure of Silicon Valley itself is changing, particularly when it comes to funding. Instead of traditional venture capital firms investing in startups from PowerPoint to IPO, there are angel investors and seed rounds on one end and traditional public market investors investing in private unicorn rounds on the other, with venture capital firms somewhere in the middle. And no company is more responsible for this radical transformation than Amazon: the company changed the inputs, and the butterfly effect is upending the entire system.

Venture Capital as Arbitrage

There’s a tendency in tech journalism to view venture capitalists as the moneymen (I always try to use gender-neutral terms on Stratechery, but it would be dishonest to even make an attempt here given the pathetic fact that only 4% of partner-level venture capitalists are women). In truth, though, middlemen is just as appropriate: the actual money comes from limited partners like family trusts, university endowments, pension funds, sovereign wealth funds, massively wealthy individuals, etc. Limited partners have highly diversified portfolios of which venture capital is only one part — the high-risk high-return part — and the reason they “hire” venture capitalists is for their skill in identifying and investing in new companies about which LPs have neither the expertise, time, or knowledge to invest in by themselves. Moreover, they pay handsomely for the help: venture capitalists usually charge around 2% of the fund per year2 in fees and keep about 20% of profits (fees are often but not always subtracted from the final payout; however, if the fund loses money the fees aren’t repaid).

I point this out to highlight the fact that at a basic level venture capitalists are arbitrageurs: they have access to more information than those with the capital, and access to more capital than those with information, and they profit by exploiting the mismatch.3 And to be clear, this is not a bad thing! Our entire economy is predicated on middlemen: no one grows their own food, to take an extreme example; rather, we depend on an entire supply chain of middlemen that results in $4 toast from wheat that costs $4/bushel.

In the case of startups, during the 45 years after Arthur Rock founded the first venture capital partnership in 1961, the vast majority of new firms needed significant funding from day one. Hardware startups of course needed specialized equipment, the funds to make prototypes, and then to set up actual manufacturing lines, but software startups, particularly those with any sort of online component, also needed to make significant hardware investments into servers, software that ran on said servers, and a staff to manage them. This was where the venture capitalists’ unique skill-set came into play: they identified the startups worthy of funding through little more than a PowerPoint and a person, and brought to bear the level of upfront capital necessary to make that startup a reality.

Amazon Web Services and the Angels

In 2006, though, something changed, and that something was the launch of Amazon Web Services.4 Because a company pays for AWS resources as they use them, it is possible to create an entirely new app for basically $0 in your spare time. Or, alternately, if you want to make a real go of it, a founder’s only costs are his or her forgone salary and the cost of hiring whomever he or she deems necessary to get a minimum viable product out the door. In dollar terms that means the cost of building a new idea has plummeted from the millions to the (low) hundreds of thousands.

In turn this has led to an entirely new class of investor: angels. There are a lot of people in the San Francisco Bay Area especially who have millions in the bank — enough to live comfortably and take some chances, but nowhere close to the amount needed to be a traditional limited partner in a venture capital firm. On the flipside, though, these folks have a huge information advantage: they are still a part of the startup scene, both socially and professionally; they don’t need someone to make deals for them.

Previously these individuals would have probably tried to join a VC firm and chip in some of their own money to a fund alongside traditional limited partners. However, thanks to AWS (and open-source software) and the fact starting companies no longer needs millions, these angels are able to compete for the opportunity to fund companies at the earliest — and thus, most potentially profitable — stage of investing.

In fact, angels have nearly completely replaced venture capital at the seed stage, which means they are the first to form critical relationships with founders. True, this has led to an explosion in new companies far beyond the levels seen previously, which is entirely expected — lower barriers to entry to any market means more total entries — but this has actually made it even more difficult for venture capitalists to invest in seed rounds: most aren’t capable of writing massive numbers of seed checks; the amounts are just too small to justify the effort.

Instead, venture capitalists have gone up-market: firms may claim they invest in Series’ A and B, but those come well after one or possibly two rounds of seed investment; in other words, today’s Series A is yesteryear’s Series C. This, by the way, is the key to understanding the so-called “Series A crunch”: it used to be that Series C was the make-or-break funding round, and in fact it still is — it just has a different name now. Moreover, the fact more companies can get started doesn’t mean that more companies will succeed; venture capitalists just have more companies to choose from.

Indeed, one can absolutely make the argument that the advent of angels has been good for venture capitalists: now, instead of investing in little more than a Powerpoint and a person, firms can invest in real products that have demonstrated traction in the market. And to be sure, startups still need the money: it may be easy to get off the ground, but that means it’s just as easy for potential competitors. The new competition amongst startups is about scaling and marketing and sales, all of which are expensive and require expenditures months or years ahead of expected profits, which is exactly what venture capital makes possible.

The Disruption of Venture Capital

If you’ll forgive a brief digression, one topic I cover quite frequently is publishing. My reasons, though, go beyond the fact that’s the business I myself am in; publishing is ultimately about text, and text, by its very nature, translates perfectly from analog to digital. And so, from the very first days of the Internet, the publishing industry has been like a canary in the digital coal mine: whatever befell it is likely to portend what might befall other industries once some essential part of their business is impacted by the Internet.

In the case of publishing, what happened is that the Internet was, at least at first, a huge boon: suddenly newspapers were reaching millions of people all over the world that they had previously had no access to. That breaking down of geographic barriers, though, ultimately undermined an entire business model predicated on arbitrage between readers seeking information and advertisers seeking attention.

I think there are parallels to be drawn to venture capital: sure, it’s nice to be able to invest in products instead of PowerPoints, but the tradeoff is the loss of proprietary knowledge about which startups have outsized potential and which don’t, and the influence on founders when it comes to everything from hiring to follow-on funding to when is the right time to go public. That influence is now increasingly gained by those investing in the seed stage, whether it be angels or incubators like Y Combinator.

Moreover, just this week came two pieces of evidence that some of these early stage investors are interested in encroaching further on venture capitalist turf:

  • First is AngelList, which just raised $400 million from CSC Venture Capital, the U.S. arm of China Science & Merchants Investment Management Group.

    AngelList is the most systematic effort to date to give structure to the world of angel investing. Angels who source a deal can form “syndicates” in which other angels invest in the sourced deal for a share of the investment’s returns commensurate with their investment. AngelList’s new fund aims to make this even easier: qualified investors can make firm offers knowing that AngelList will fill in the funding gap between sourcing a deal and recruiting other angels to join a syndicate. And, more importantly, AngelList can partner with syndicates to fund follow-on rounds in the best companies. In other words, Series A and beyond.

  • Second is the aforementioned Y Combinator, the incubator that has seed funded startups worth a combined $30 billion, including Airbnb, Dropbox, Stripe, and a whole host of other companies you’re probably familiar with. Just yesterday, Y Combinator reportedly led a Series B round in Checkr, which automates background checks. The funds were from Y Combinator’s new Continuity Fund, which supposedly would be making pro rata investments at <$250 million valuations in all of Y Combinator’s startups gaining additional funding, but the question as to whether or not Y Combinator has reversed its previously stated policy for the fund is less interesting than the fact the firm is also moving up market.

It is, in some respects, a classic disruption story: angels and incubators were happy to get down in the mud with the huge number of new startups enabled by Amazon Web Services and open source software; meanwhile, said startups’ low up-front costs didn’t provide an adequate return on a venture capitalist’s time (or check). Instead venture capitalists fled up-market, only to find the folks they were so happy to benefit from moving on up into their space.

The Venture Capital Squeeze

The story doesn’t end there: the trouble for venture capitalists is that they are getting squeezed from the top of the funding hierarchy as well: a new class of growth investors, many of them made up of traditional limited partners like Fidelity and T. Rowe Price, are approaching unicorn companies on a portfolio basis. I wrote in a Daily Update last June:

If you wait to invest until startups are already unicorns, or nearly so, you can be invested in a portfolio of unicorns! Just look at the portfolios of some well-known late-stage investors (all data from Crunchbase):

  • T. Rowe Price has invested in 16 unicorns, including 3 of the top 10, and 7 of the top 25
  • Fidelity has invested in 10 unicorns, including 5 of the top 10, and 8 of the top 25
  • Tiger Global has invested in 13 unicorns, including 1 of the top 10, and 3 of the top 25
  • DST Global, who in my opinion have the most responsibility for starting this trend, has invested in 10 unicorns, including 3 of the top 10 (and 5 of the top 12)…

You could make the analogy about all of these growth investors to venture capitalists: they are investing relatively speaking small amounts of money into a portfolio of unicorns, and all they need is for one or two to make it to a major liquidity event to profit.

Sure, this is relatively dumb money, but that’s where those angel and incubator relationships come in: if startups increasingly feel they have the relationships and advice they need, then growth funding is basically a commodity, so why not take dumb cheap money sooner rather than later?

The Internet Impact

Interestingly, just as in every other commodity market, the greatest defense for venture capitalists turns out to be brand: firms like Benchmark, Sequoia, or Andreessen Horowitz can buy into firms at superior prices because it matters to the startup to have them on their cap table.5 Moreover, Andreessen Horowitz in particular has been very open about their goal to offer startups far more than money, including dedicated recruiting teams, marketing teams, and probably most usefully an active business development team. Expect the venture capitalist return power curve to grow even steeper.

The more important takeaway, though, is that this upheaval is happening at all: even a seemingly impenetrable clubby human interaction-driven industry like venture capital is susceptible to change that, in retrospect, is really quite radical. You see it in industry after industry: hotels presumed that people wouldn’t stay in strangers’ homes, television networks presumed that programming schedules were constrained by time, and, speaking of Amazon Web Services, enterprise technology companies presumed that servers and software would live on corporate premises. Once that premise is removed, though — ratings commoditized trust, streaming commoditized time, scale commoditized data centers — everything else that you didn’t think mattered does. Airbnb has better selection and often cheaper prices, Netflix is cheaper and has a broader selection, Amazon offer customizability and flexibility.

So it is with venture capital: once startup funding requirements were reduced, the superior information and the willingness to hustle of angels and incubators earned the trust of the big companies of tomorrow, reducing more and more venture capitalists to dumb money hardly worth the 20% premium. The inputs to the Silicon Valley system have been changed, and we’re only now seeing the effects, and that should be a cautionary tale for just about everyone who thinks they and their industry are safe from the Internet’s impact.

  1. Or, as critics may counter, a new level of commercialism and intrusiveness. But I’m an optimist — and a realist ↩

  2. Usually for 10 years, the traditional life of a fund ↩

  3. To be sure, the best sort of VCs do more than generate “deal flow”, as it’s called: they offer advice, help with hiring, make connections, find additional investing partners, and perhaps most importantly, at least for the most well-known firms which capture an outsized share of venture capital returns, validate the startups they invest in with potential employees, customers, and partners ↩

  4. Obviously AWS in 2006 — which was just the S3 storage service — wasn’t capable of supporting a startup; it took several years to add the necessary services. Moreover, I am unfortunately giving short-shrift to the role of open-source software, which is the left hand to Amazon’s right ↩

  5. Semil Shah, who provided feedback on an early draft of this article, wrote about this in August ↩

Twitter’s Moment

Posted on

It’s Always Darkest Before the Dawn

I try to save the most over-used of clichés for special moments, and that’s exactly what this week feels like for Twitter. You may disagree, of course — Wall Street does, having driven the stock down yesterday to just a dollar above its IPO price (and 38% down from its first day close) — but that’s why the cliché works: things may seem dark, but I’m optimistic that the horizon has just the slightest glimmer of light.

Long time readers know that while I love and value the product, I’m no Twitter fanboy. The company’s user retention issues were apparent well before the IPO, and the company had a clear product problem that, ultimately and correctly, cost CEO Dick Costolo his job. Tack on a messy — and thus, unsurprising — CEO search that culminated in a part-time CEO in direct contradiction to the Board of Director’s public statements to the contrary,1 and no wonder the stock has been down-and-to-the-right, with all of the problems (especially around retention) that that entails. Again, this was predictable.

I think, though, it’s time for a new prediction: that the summer of 2015 will be seen as the low point for Twitter, and that this week in particular will mark the start of something new and valuable. Crucially, the reasons why are directly related to why I was bearish for so long: the product, the CEO, and the stock.

The Build-up

It was a bit disconcerting when, during the conference call to announce the appointment of Jack Dorsey as CEO, Lead Independent Director Peter Currie, Dorsey, and newly promoted COO Adam Bain brought up Project Lightning, an internal project that was bizarrely revealed to BuzzFeed by Costolo just days before his departure; at the time it was hard to see the pre-announcement as anything other than a last ditch attempt to save his job, and one wondered if the mentions on the conference call had a similar motivation: give Wall Street something, anything, to hold onto, even if jacking up expectations would hurt the new product when it launched.

Well, the product launched…and it’s fantastic. Moreover, it’s not only that it’s fantastic from a product perspective — actually, there is a lot to nitpick — but that it is fantastic from a strategic perspective.

The Product

Moments has three components:

  • When you first tap the Moments tab at the bottom of the Twitter app2 you’re dropped into the ‘Today’ view that lists a mishmash of stories that, well, happened today.

    IMG_0153

    Touch any of the stories to get a curated list of tweets that tell the story in question through videos, images, and sometimes just text. It’s a really great experience, and I found the sports stories with their combination of highlights and tweeted reactions particularly enjoyable3

  • For any Moment in progress, you can tap a button to add tweets about that Moment to your main timeline. Crucially, though, those tweets only persist for the duration of the event in question; the ‘Unfollow’, which is the most essential action when it comes to building a Twitter feed you actually read, is done for you

  • Finally, in what was probably the biggest surprise in the product, there is a carousel at the top leading to more focused categories:

    carousel

    Each of these categories includes not only ‘News’ or ‘Entertainment’ Moments that just happened, but also more timeless content, particularly in ‘Fun.’ Look carefully at those category titles, though — they sure look familiar:

    shutterstock_2569368

    That’s right, Twitter just reinvented the newspaper. It’s not just any newspaper though — it has the potential to be the best newspaper in the world.

The Strategy

This newspaper angle touches on the strategic aspect of Moments. The demise of newspapers is the most obvious example of Aggregation Theory and what happens when distribution becomes free. You go from a situation where geographic bundles are the only way to reach consumers to one where consumers can access any story from any publication — it’s those that control discovery that have all the power:

aggregationpublishing

Google was the first and biggest beneficiary of this shift, but, as I noted last week, Google has always been more about intentional discovery as opposed to directed browsing. On the flipside, Facebook excels at serendipitous discovery, but if you actually want to get informed about what is happening, without knowing what you want to know, then it’s not clear what is the best answer.

True, you can go to a specific publication like the New York Times or USA Today, but you’re not getting the best possible content for every story, and besides, the fundamental format of their stories was designed for paper, not mobile. The NYT Now and BuzzFeed News apps do well to bring in stories from other apps, but the mixture is a bit haphazard and again, long-form text is good for some things but not all things.4 Snapchat, meanwhile, has Discovery — from which Twitter clearly took product inspiration, and for good reason5 — but that too is organized by publication, a decision that makes sense for business development reasons but not necessarily user experience ones. In other words, I think Twitter Moments’ organization is superior.

What’s most interesting, though, and most exciting, is understanding how it is that Twitter can pull this off: the company doesn’t need stories from publications because it has nearly all of the originators of those stories already on its service. In other words, if the Internet broke down newspapers to their component stories, Twitter breaks down stories to their component moments,6 and those moments are chronicled not only by normal people on the ground but by the best news-gatherers on the planet.

The fact that news is reported first via tweet has always been true of Twitter, and it’s a huge reason why so many are so devoted to the service. However, to become an effective aggregator you have to aggregate not just the source material but also the audience, and the problem for Twitter is that learning the product has simply been far too difficult for most new users. There is nothing difficult about Moments, however, and Twitter should make it the default screen for their app.7

What is exciting is that Moments isn’t close to fulfilling its potential: imagine a tweet-based newspaper drawn not only from the best sources in a mobile-friendly format, but one perfectly customized to you. This is what Twitter is already like for power users, but again, getting to that state is simply too difficult. Figuring out how to do this systematically on users’ behalf should be Twitter’s chief aim.

“Should” is probably a bit superfluous: the incentives for Twitter to focus on this type of customization are massive. Twitter is uniquely positioned to understand what its users are interested in, something that at least theoretically rivals Facebook’s imposing demographic information, SnapChat’s youth advantage, or Pinterest’s grasp of my aspirations. The reason customized Moments matter is because there are two payoffs: the user experience is better, and the advertising that will undoubtedly be sold in Moments will be better targeted and more effective.

I do think the ads will be great, regardless. The placement opportunity within stories is obvious, and like the best sort of in-stream native advertising the brand in question will take over the entire screen, if only for a moment. And, like the other networks I just listed, Moment ads will be both unblockable and located where users live, not where they visit.

The Content

Certainly, as any number of articles have already breathlessly noted, Twitter Moments is built for basic users. It’s essential, though, that Twitter not forget about those living mostly happily with the company’s core product — they’re the ones that create Moments’ content (for free, I’d add). This is where Dorsey’s return as CEO has the potential to have the most impact, because the core product has stagnated horribly, particularly in regards to the 140-character limit and conversations.

Specifically:

  • It remains inexplicable that tweet payloads, such as links, pictures, and videos, count against the 140-character limit. Any and all attachments should not count against the limit
  • There is one medium missing from the previous bullet point: text. Twitter cards can already handle the aforementioned videos or photos (or link previews), so why can’t they handle text? I’m not advocating for the removal of the 140-character limit for public tweets,8 simply a way for people to attach more fully-formed thoughts to the existing stream without needing to rely on obnoxious tweet storms. Links are great, but following a link is expensive for the user in time, bandwidth, and mental burden, and there’s no reason Twitter shouldn’t make it easy to keep ideas in its app
  • Twitter conversations, which often contain the very best content on Twitter, are horribly broken and nearly undiscoverable. First, it’s again ridiculous that @-names count against the 140-character limit, making conversations with more than one or two people untenable. Second, it is far too difficult to follow a conversation — Twitter should investigate folded conversations like on Sina Weibo, the old Friendster, or Facebook. Third, Twitter should make it easy to temporarily follow a conversation, just like a moment, or, on the flipside, easily drop out

As far as I can tell, the primary reason none of this has been implemented is that no one at Twitter has had the authority to tamper with the sacred 140-character limit. This quote from a Recode story about potential changes to the limit was damning:

“People have been very precious at Twitter about what Twitter can be and how much it can be evolved,” said one current senior employee. “Having Jack come in and say it’s okay makes all the difference in the world.”

The fact of the matter is that 140 characters is an implementation detail; Twitter’s core value is in its interest graph, and it’s gotten to the point where fealty to that detail has had a material effect on said graph. If Dorsey really is able to just walk in and get people in line — even if only part-time — then I feel all the better about my endorsement of him for CEO:

Twitter has long been captive to its best users [and apparently employees] who rail against any change on the margins, much less even a rumor of changes to the core product; I suspect this hesitancy has been in large part driven by the fact that everyone in Twitter’s leadership was ultimately a hired gun. Dorsey, though, is a founder, and however controversial his first stint at the company may have been, there is no denying the authority this fact gives him when it comes to making changes.

Beyond Moments — indeed, for Moments to be successful — the core product has to evolve, and that’s why Dorsey is so important.

The Company

To be sure, most of this article has dealt with what Moments specifically and Twitter broadly can be, not what it is. The truth is Twitter still faces the challenges I and many others have documented endlessly:

  • The company faces the far more difficult task of convincing users who have tried and abandoned the product to return; getting new users, which is usually thought of as the most difficult problem in consumer tech, is far easier in comparison
  • The company faces far more competition when it comes to media consumption than it did a few years ago. Facebook is the 800-lb gorilla, but Snapchat, Instagram, Pinterest, and a whole host of other sites and apps are making the same bet Twitter is
  • That diminished stock price undoes rather nicely the golden handcuffs that are vesting stock grants, particularly when a seemingly endless supply of unicorns are flush with private money and, as is ever the case in Silicon Valley, circling Twitter looking to poach the employees most capable of turning Twitter around

Per the first point, I absolutely think Moments is the right product to win users back, and I’m encouraged that Dorsey said on the investor call that the company would be launching an integrated marketing campaign around the product. Twitter should spare no expense.

Secondly, the fact remains that Twitter has the best content in the world already in a format perfectly suited to mobile. For a long time the lack of product vision and execution has obscured that content from most users, but, in a fair fight, I like Twitter’s chances.

Finally, the upside of a depressed stock is that, unlike most post-IPO companies, Twitter can actually offer new employees in particular the potential for real wealth gains from their stock grants. Anyone entering Twitter now is basically on the same footing as a pre-IPO employee at a unicorn; sure, the growth may never materialize, but there is a lot of upside and Twitter should leverage that. Moreover, there’s hope that Dorsey’s return and a renewed product vision could keep a lot of important folks in the fold.9

Much has been made of the comparison between Dorsey and Steve Jobs; certainly his return to the company he helped found and was later banished from fits the mold. It’s easy to mock this comparison, particularly given the fact that Dorsey has at times seemed so eager to invite it, but in fact we should all hope the comparison holds. There’s just something different about Apple, a company that seems so full of contradictions yet one that has continued to lead the industry both financially and in key innovations. I’d argue the same about Twitter: it doesn’t make sense, hasn’t really ever made sense, and perhaps that’s the reason it, and the irreplaceable ideas it contains, are so important. I hope its moment — its dawn — has arrived.

  1. Hello shareholder lawsuit  ↩

  2. Moments is not yet available worldwide; I got it to work by connecting to the U.S. via VPN ↩

  3. I thought this moment about the Seattle-Detroit American football game was particularly good, in part because Twitter has secured rights to extended video highlights from the NFL ↩

  4. In fact, BuzzFeed News does incorporate tweets and GIFs; it’s one of the reasons I prefer it ↩

  5. I’ve said this before and I’ll say it again: I believe that good ideas are discovered more than invented ↩

  6. I like Nuzzel for discovering long-form pieces, but half the allure is seeing a story’s associated tweets, and it doesn’t really help with anything live ↩

  7. The company can use a setting to let power users stick with the timeline ↩

  8. Thank goodness the company finally removed the limit on Direct Messages, which served little purpose beyond driving people and relationship information off of the platform ↩

  9. For reasons beyond me I original referred to repricing options in this paragraph; like nearly every other company in tech, Twitter now offers Restricted Stock Units, not options. I apologize for the mistake ↩