Please Tell Me He Didn’t Actually Say That

This from Paul Graham in a recent Inc. interview when asked for a specific “tell” for poor likelihood of success when assessing YC applicants:

One quality that’s a really bad indication is a CEO with a strong foreign accent. I’m not sure why. It could be that there are a bunch of subtle things entrepreneurs have to communicate and can’t if you have a strong accent. Or, it could be that anyone with half a brain would realize you’re going to be more successful if you speak idiomatic English, so they must just be clueless if they haven’t gotten rid of their strong accent. I just know it’s a strong pattern we’ve seen.

At best, an irresponsibly vague and unsupported generalization. He is “not sure why” a foreign accent would impede a CEO. He doesn’t present any kind of evidence (statistically significant sample?) on which to base his conclusion. And what exactly is a “strong foreign accent”?

Perhaps more importantly, even if there is a big sample, who’s to say that this is causation and not just correlation? There might be another factor, like prejudice, that makes this bit of “wisdom” a self fulfilling prophecy. Of course, the true causal factors might be more benign, but that doesn’t change the underlying point.

Leaders should be working to make our communities more inclusive, not less so. Sure, a strong accent might make things harder. But let’s try to be more open minded and fight the instinct to judge someone’s competence or likelihood of success based on their proficiency in a language that isn’t their first, and might be their 3rd or 4th. Don’t rule them out with a lazy generalization. Find ways to compensate. Help them. If VCs in general aren’t making the effort to see beyond the accent, set the example and encourage others to follow.

This isn’t charity. That person with a strong foreign accent probably achieved a small miracle by getting from wherever that accent comes from to the point where they are making a YC pitch in the Valley… in English.

Courage and tenacity — probably important qualities in a startup CEO.

Paul Graham is very highly regarded in the startup and tech community. People look up to him and take his lead. If the quote above is really his, I wish he would elaborate. If it isn’t, I hope he disowns it and explains what he was really getting at.

This post has been updated from the original, more emotional version.

A Tale of Two Tiers: The End of Premium Display Advertising as We Know It

For a while now advertisers and publishers have thought of display ads in two main buckets: premium and remnant. Premium campaigns were sold to the advertiser, either directly or through a network, by a salesperson. The advertiser would be sold a specific number of impressions during a specific time period and they would theoretically have some control over where the ads would appear — on which websites and sometimes even alongside which content.

But publishers were always left with some unsold impressions and would typically sell them to a network like Google AdSense as remnant inventory. On a good month the impressions sold as remnants would be single digit percentages, but on a bad month they could be a much larger share of the overall number of impressions.

The canary in the premium advertising coal mine was the massive difference between the CPM (cost per thousand impression) rates that a publisher could charge for their premium and remnant inventory. In my experience the difference could be an order of magnitude.

This screamed two things:

  1. Benefits that premium ads offer to advertisers would have to be big enough to support this level of price discrimination between premium and remnant inventory.
  2. Remnant inventory represented a big opportunity for arbitrage, so it was inevitable that new players would enter the market with ad products that packaged so-called remnant inventory in interesting ways.

When I started seeing ads from our premium advertisers also appearing on via AdSense, I realized that #1, the price discrimination, was getting harder to justify. Identical creative, appearing in the exact same ad position on the same day. Just 10x cheaper.

Then came #2 — the rise of the machines — ad exchanges that to a large extent automate the matching of ad inventory supply and demand. Although the remnant inventory for an individual publisher was often a small percentage of their overall inventory, the aggregate remnant inventory across thousands of sites represented a lot of eyeballs. And at a much lower cost per eyeball.

Not only that, but purchasing ads from an exchange involves no sales people and no traditional sales cycle. Want a few million impressions for your ad during the first week of May? Bam. You got it. Sure, you don’t know exactly where those ads are going to appear, but the buy took you 5 minutes and cost a fraction of the price.

Of course, the more overall inventory that is sold programmatically, the better the average “quality” of those impressions and therefore the better the value proposition that programmatic buying offers. So we can expect ad exchanges, demand side buying and their ilk to accelerate their inroads into the premium market. In fact, this process is already quite far along. Case in point is a company like Federated Media, which was one of the pioneers of high touch premium campaigns and conversational media, but through their acquisition of Lijit they have pivoted to make programmatic buying of ads a central part of their offering.

Although companies like FM are still doing high touch, premium campaigns, they are only doing them with very large companies that have very large budgets. For these campaigns the economics of a costly sales cycle still make sense. The question is, for how long?

It is tempting to view this as yet another example of the innovator’s dilemma, but I don’t think all is lost for premium, bespoke campaigns. Yes, the pendulum will overshoot in the direction of programmatic buying, and the days of easy CPMs are over, but there will be a new normal where both high volume programmatic buying and high margin premium campaigns co-exist.

Already there are innovators moving to fill the gap left by larger networks that now only focus on campaigns with 6 figure budgets., founded by ex-FM veterans, is one example of a company looking to define new kinds of carefully thought out, high value campaigns for advertisers that want more precision and more care taken with their brand. And opportunities to advertise in more creative ways than an IAB banner or the latest frightening ad unit that takes over your browser window. Companies like look more like boutique agencies than ad networks. After all, to make the economics of the sales cycle work at a smaller scale they need to be adding a lot of creative value 2.

The key question: How will boutique networks and the niche publishers they partner with quantify the value that premium campaigns can offer to a brand? Intuitively, we all know it is there — we despair of the dumb banner advertising we see all over the web and laud campaigns that are creative, targeted and intelligently tied into social media — but how will we measure awareness and brand uplift that doesn’t translate into immediate business? “We know it when we see it” still isn’t an effective way to present ROI.

  1. “Native” is the buzzword that has emerged to describe campaigns that are more, well, native to the places they appear. The canonical example is the ads alongside search results in Google or Bing. But for publishers it is about getting advertisers closer to the editorial content. If the potential conflict of interest challenges are negotiated properly, this can be very powerful. Advertisers can associate themselves with specific content and specific publishers that reflect their brand values. And their message appears in a context that makes sense. However, I hesitate to use the term “native” to describe what a company like is doing because the now ubiquitous “You might also like” or “Also on the web” sponsored links under articles — mostly automated, sometimes annoying and often spammy — are also referred to as native advertising.

Privacy on iOS 6 is a Wake Up Call for 3rd Party Web Analytics Tools

There is much discussion of the way that the goals of advertising and privacy are at odds. But advertising is not the only casualty when browser makers clamp down their privacy settings. Another big category that this will affect is 3rd party analytics tools. Google Analytics (GA), for example, is an extremely powerful (and free) way for website owners to capture aggregate information about visitors in order to figure out what is working and what isn’t.

This is what I was doing this morning — browsing through our sources of traffic in GA — when something jumped out at me. The direct visitors to our site had started climbing rapidly in mid September, faster than any other category. This is strange, because you would expect that “direct” visitors, who have typed your address into the browser address bar, to grow more slowly than “referral” visitors, who came by clicking a link to your site, or than “search” visitors, who found your site using a search engine.

Digging a little deeper, I found that the increase was strictly among new visitors. This might be explained by press coverage in a printed publication or a TV mention, situations where someone heard about us but didn’t have a link to click. But I didn’t know about any recent coverage, and this sort of influx of visitors is normally a spike, not sustained growth.

What had changed after September 18 to catalyse this growth? Mid September… My conscious brain was working through any changes we had made on the site during the last month when my subconscious brain tapped it on the shoulder and said: “iPhone 5?”.

Of course. The first people who ordered the new model started receiving them on September 20. And I could quickly confirm that all the additional direct visits from new users were people using the version of Safari in iOS 6.

Still, what exactly was going on? I know there are some privacy settings changes in iOS 6, so my first thought was that Google Analytics was no longer able to distinguish between new and returning visitors. Indeed, the pageviews per user did seem to be dropping.

This didn’t account for the difference though. First, the decline in pages per visit seemed to start earlier than September 20. And second, returning visits from iOS 6 did not go to zero, so clearly GA was able to detect return visitors in some cases. It was something else.

Now that I knew what I was looking for — something in iOS that was impacting analytics — it was only a short search before I found the answer in an article by Danny Sullivan (Search Engine Land). According to Sullivan, it is indeed privacy related, but the issue is that iOS is not allowing Google Analytics to see the referrer (the bit of info that tells it where you came from) when you search from the search box in the Safari browser. (For the nerds, Safari is sending the searches through Google SSL search.)

The analytics for iOS in particular supports their theory. Right around September 20, iOS traffic sources start to shift from Google to “direct”. The rate is probably reflecting the penetration of iOS 6.

My take-away from all this is that the move towards privacy might have a big impact on 3rd party analytics services. Not only GA, but any other service that is running on someone else’s server and gathering info about your visitors via a tracking pixel or JavaScript that you put in your pages. In other words, all of them.

Now might be the time to invest in self-hosted solutions. Several years ago it was all about self hosted analytics, then Google bought Urchin and the move towards 3rd party hosted analytics began. Privacy could put a stop to all that and take us back to the days of running something like Urchin on our own servers. Although losing the power of GA would be sad, a resurgence of self hosted solutions would be awesome. If you have ever tried to find a good and affordable analytics tool to run on your server, you’ll know what I mean.

There Will Be No Accounting for (Other People’s) Taste in Tablets

For more than a decade, legions of Apple fans have been wondering how it is possible that people will choose a Windows PC over a Mac. After all, user experience is critical, right? And there is no question, at least in the mind of a Mac user like me, that Apple’s neatly integrated hardware+software experience is vastly superior to the franken-machines that Microsoft and its OEMs conspire to put in Best Buy.

Yet people bought, and are still buying, PCs in far greater numbers than Macs.

It must be the file formats, we say. Microsoft Office dominates word processing and spreadsheets and people need to share those documents. Yes, there is a pretty good version of Office for the Mac, but it is never quite as good as the Windows version.

It must be the Windows hold of the enterprise market. It is much more convenient to have a computer at home that is similar to the one you have at the office.

It must be the ubiquity of Windows. The fact that so many people have Windows PCs makes it much easier to find help and advice.

These are all true, but there is another factor that is much harder for Apple fans to understand: Not everyone sees the beauty. We’re wine connoisseurs watching, with condescending bewilderment, as the rest of the party happily drinks $8 box wine.

Sure, the simplicity of Apple design makes a huge difference for users in the early stages of a new concept’s introduction. What the Mac was to the GUI, the iPod was to MP3 players, the iPhone was to smartphones, and the iPad is to tablets. Apple does a fantastic job of introducing the world to these complex new product categories. But as we as a society begin to get comfortable with the concepts, we can tolerate more complexity. Simplicity is always important, but it becomes relatively less important over time. Other things — like “is it the cheapest one”, or “does it come in neon pink”, or “does it have a USB port” — become more important.

We thought people were choosing the Apple product because of its elegant simplicity, but most of them were really just choosing the simplicity. And to someone who doesn’t resonate with Apple’s aesthetic sensibilities, the elegance becomes a constraint when Apple prioritizes it over things like extra ports, customizability and expansion options.

So it will be with tablets. And while it is too early to tell whether Surface the product will be successful, Surface the concept of a credible, non-Apple iPad alternative is inevitable. And who better to deliver it than Microsoft, a brand on more of the world’s computers than any other.

Ha ha, we scoff. Who wants to do Microsoft Office on a tablet? Office is boring. And tablets have a completely different use case to laptops. Who would want one to run full Windows?

Answer: Lots of people. People with different priorities, working different jobs, living in different countries. People we don’t quite understand.

Lessons from the Failure of Readability’s Author Payment Plan

Yesterday Readability announced that they were shutting down their experiment to take payments on behalf of authors and then distribute them (minus a 30% cut) to those authors. The announcement talks about learning from the experiment, the main learning being that while readers seemed to like the model, it was impossible to get enough publishers on board.

For one thing, most authors in the long tail of content that flows through Readability have no idea that Readability even exists. But another big factor was that many authors who did know about Readability reacted angrily to the model. It seemed presumptuous of Readability to collect payments on the behalf of authors, without said authors ever agreeing to the model. Some prominent authors expressed their anger. Daring Fireball’s John Gruber called Readability ‘scumbags’.

Fanning the flames of the subsequent disagreement on Twitter and in the blogosphere were two Internet heavyweights and Readability advisors, Anil Dash and Jeffrey Zeldman. Both are held in very high regard because they have a record of actively pushing the web forward. Dash focusing on its potential for positive cultural impact, and Zeldman on advancing web standards. In this context it is easy to see why they would be attracted to Readability. The service is tackling the hard problem of finding new business models for authors, and it is making the web a better experience for readers.

The problem: These noble goals have blinded Readability and their advisors to the extent that their sense of entitlement regarding other people’s content angers the people that created that content. So I suggest that there are two important things to learn from the kerfuffle surrounding the Readability payments system.

The Sting of Entitlement Matters More than Money

Dash has repeatedly and correctly pointed out that Readability is not the only company to derive revenue from a read-it-later service. Competitors Instapaper and Pocket do much the same thing for their users. On the face of it Readbility, which tried to compensate publishers, should be more popular with authors than competitors who make no attempt to do so. Right?

Wrong. It turns out that many authors care less about the money, and more about the fact that Readability is representing them without their consent. As a publisher, I understand this reaction and it isn’t entirely rational. Even if the financial outcomes were equivalent for Readability and Instapaper, and even if the user experiences they offered were identical, I would find it easier to accept Instapaper deriving direct benefit from someone buying their app than to accept Readability deriving benefit from collecting revenue on my behalf.

The Readability model just felt dishonest and entitled. Although I thought that John Gruber’s ‘scumbags’ comment was over the top, it perfectly captures the way I feel about Readability in the pit of my knotted stomach whenever I think about their payment plan.

Product Developers Will be Held Accountable

With hindsight it is clear that Readability’s model was impractical because there is such a long tail of content and authors. So now there is nearly $150,000 in a bank account, most of which won’t make it to authors.

Readability made a promise to their customers — that they would compensate authors — and now they can’t deliver on that promise.

Much of Readability’s defense of this failure, and particularly Zeldman’s defense of Readability’s actions, centers around the fact that they are trying to do something good for readers and for authors. But even if the Readability team was entirely motivated by the altruistic goals of improving reading on the web and finding a better business model for authors, it wouldn’t matter. It isn’t Readability’s goals that anger authors. It is their actions.

The lesson is that trying something novel, even with good intentions, does not give you a free pass. As a product developer you will be held accountable for the outcomes related to your product. If you fail, you will have people to answer to and “but, I was trying something good and cool” will probably not be sufficient to get you out of hot water.

Lessons: Unlearned

Readability’s approach to shutting down the payment plan illustrates that their take on the lessons from the failure of the payment model is very different to mine.

First there is the unilateral decision regarding how to deal with the unclaimed payments. Depending on how you look at it, this could be the authors’ money or the readers’ money. Deciding what to do with it without consulting authors or readers shows the same sense of entitlement as collecting the money on the authors’ behalf in the first place.

The decision to explain things in terms of the failure of an experiment, instead of apologizing for getting things wrong, shows a belief that startups can take money from customers without being accountable for delivering what they promise in return.

Then there is their decision to give the money to “non-profit organizations that speak to the spirit of supporting reading and writing”. The implication is (1) that the positives of giving to a non-profit outweighs the negatives of their entitlement in making that decision unilaterally, and (2) that this is some measure of compensation for the failure of the model.

But the end doesn’t justify the means. Giving to charity may well be the only practical way forward, but that doesn’t mean everything’s ok. Without a clear and explicit acknowledgement that this was a flawed idea that failed, and a sincere attempt to involve authors and readers in their decisions around what to do with the money, the flames will continue to burn.

Bridges: Burning

Reading the comments on the announcement post, it is clear that there are both trolls and reasonable people calling foul. Unfortunately, Zeldman’s passionate, angry defense of Readability seems to be directed at both and is rapidly diminishing their chances of making things good.

I can get over being called a troll by my erstwhile heros of the Internet. But I do wish that folks at Readability were internalizing more of the real lessons here. This episode shows all the hallmarks of a classic PR stumble — a sense of denial, underestimation of the concerns, angry surrogates digging the hole deeper — and they will be the biggest losers when it all dies down.

Consumer Startups without Revenue are Products, not Businesses

Nick Bilton recently wrote about stratospheric valuations for companies that don’t generate any revenue and there has been much debate on whether or not this represents a bubble. The more interesting thing from my point of view is how the incentives of venture capital encourage consumer-oriented startups to adopt an ad-based1 business model and simultaneously set them up to fail at executing one profitably.

In the Beginning

Early in the life of a social product like Twitter or Tumblr, there are no ads. The product is conceived without advertising in mind, but rather with a very well intentioned focus on user experience.

The reasons for this focus are easy to explain. First, the people building the product have a vision that is all about end user value. They might be solving a problem for themselves, or addressing a pain point they have identified. They are not driven by the passion to create a vehicle for ads.

Second, for a small user base in the early stages of growth, it hardly makes sense to waste time thinking about selling or serving ads. And until your online business represents a lot of eyeballs, no advertiser or ad network of note is going to be interested in working with you. From my experience, and my chats with people who have much more experience than I do, you need millions of ad impressions to be interesting to a network, and millions of unique users to be interesting to the ad agencies directly.

Third, for any product idea there are soon many competitors, and it is hard to be the one investing some of your focus on ads, while your competitors focus purely on user experience.

We Can Introduce Ads Later. Right?

Maybe not. In the delicate cocktail that makes a great user experience, advertising is a powerful, overbearing ingredient. Adding it to the mix after the fact is almost impossible to do without changing the flavor to something completely different. And probably not very palatable.

Even if the new ad supported product isn’t awful, it is not the same product that users grew to love. To them the addition of ads feels exactly like the bait and switch that it actually is. At best, the ads are an annoyance that devalues the product. At worst, they are a sign of that the makers of their beloved product are “selling out” and have lost their way.

This seems obvious, so why are good product people, who are in other respects thoughtful about the minutia of user experience, so naive about this aspect of their product’s design? The answer is that they never considered it. And this wasn’t an oversight. It was intentional. Advisors and investors encouraged company leaders to delay thinking about monetization. The high order bit, in fact, the only bit that matters, is to grow the number of users and their level of engagement as quickly as possible2.

The result if the company is successful? A product that wasn’t designed to generate advertising revenue, and a company that doesn’t have any advertising DNA.

The Ads are a Lie

The startup got to this point because the focus was never on building a real business with profitable products. After all, the product that matters to early stage investors is not the one the startup is selling, it is the startup itself. History tells them that with enough traction from users, one can sell a business without a single cent of revenue for $1 billion dollars.

When the investors say “the business model can come later”, they are adding “after the exit” in mental parentheses. These days that exit most often comes in the form of an acquisition. The acquirer might want the technology, the users or just to eliminate an upstart competitor. Post acquisition, they will set about integrating the new products into their own business model, whatever that may be. All is well.

In a way, the notion of advertising as the business model has served its purpose as a way to be in denial about the absence of one.

And this is truly why the ad-based model is so popular. It is a model where entrepreneurs and investors can convince themselves and others that a bait and switch is doable. It would be much harder to accept, for example, that an online service was going to launch for free and then start charging later. The suspension of disbelief is based on the lie that ads don’t cost the user anything, when of course, they do.

The race is to get acquired before the lie is revealed. In later stage funding rounds it might be necessary to start dabbling in monetization in order to maintain the lie, but woe is the startup that runs out of growth hype, funding runway and acquisition prospects. It has to stand on its own two feet.

The Incumbents Win

For a company to be truly successful with an ad-based business model, it should start with one. After all, if the product isn’t sufficiently compelling with the revenue generating components in place, then it probably isn’t the basis of a viable business.

Sure, there is the eyeball scaling issue, but on the face of it this challenge — getting over the scaling hump into profitable territory — is exactly the sort of problem that VC money should solve. The wrinkle is that VCs focused on the exit won’t insist that their money is used to build a profitable business. They would prefer a buzz fueled phenomenon with a stratospheric valuation that preferably has no anchor in the boring reality of profit and loss.

These whacky incentives create an environment where many startups really aren’t working towards profitability, but rather towards an acquisition exit. They are building experiences, not businesses. And since they don’t have the shackles of P&L reality, they can do things that make it hard for businesses with real goals of profitability to compete.

The result is a reduced likelihood that large incumbents will be ousted by new upstarts in the consumer market. The incumbent just makes the upstart an offer. And since the upstart has no real revenue prospects on the horizon, its an offer they can’t refuse. And that’s a real pity, because large companies are where successful startups go to die.

In fact, large companies are also where many unsuccessful consumer startups go to die, via the acqui-hire. Typically these are startups that built innovative products, got some impressive user traction, but didn’t get to the nice fat acquisition before they peaked and started the decline. In this case, the acquirer gets a waning user base and some proven product talent at a bargain basement price. The founders save face. The investors get something rather than nothing. Everybody makes out ok, except… the customers.

The Customers Lose

Wherever the outcome is on the continuum from big acquisition success story, to a face-saving acqui-hire, to messy death, the customers are pretty much in the same boat. The product as they knew it, and in which they invested precious time and energy, will go away. If they are lucky, they might have the opportunity to extract their data in some way. And there might even be an alternative service eagerly welcoming the refugees. Even in the best case, however, it is not the outcome that customers hoped for.

The more this movie is repeated, the more consumers are going to take an interest in the business aspects of the services they use. The “too good to be true” spidey sense that we apply so effectively in the real world will start to get applied to online services. We will start to care that the companies we use to store our memories and connect with others are viable businesses. We will want to either pay them money, or see evidence of an advertising model that is sustainable.

That there are incentives to create companies without business models is a systemic problem. This isn’t to say that monetization-free startups aren’t innovating (they are), or that there aren’t exceptional cases where a startup finds its business model later (Google!), but if you would prefer to see a vibrant market of consumer Internet services that is less dominated by giant corporations who gobble tiny ones, then the incentives will need to change.

Fortunately, as the long term results of the current system become clear and as consumers become more savvy, the invisible hand of the market will make it so.

  1. When I use the terms “advertising” and “ad-based” I mean a broad range of business models that involve getting a third party to pay you in return for some sort of promotion to, or information about, the people who use your product.
  2. Revenue is sometimes even considered less than a “nice to have”. Without it, the sky’s the limit on the valuation of the company. With it, the valuation is brought sharply back down to earth.

Thanks to @toddwseattle and @sp990 for reading and commenting on earlier versions of this post.

China Smartphone Share: Demand Side Fundamentals will Beat Supply Side Tactical Advantages

Idiocy in the tech press is particularly acute recently as writers pounce on the news that Samsung now has triple Apple’s share of the smartphone market in China.

Let’s break this down. In one corner, we have Samsung, which has been manufacturing and distributing phones for decades. It has relationships with all the mobile operators in China.

In the other corner we have Apple, who has been breaking into this market for just over two years. In a notoriously tough carrier ecosystem, Apple originally partnered with China Unicom, and more recently with the smaller operator China Telecom. But not with the nation’s biggest operator, China Mobile. Despite this handicap, and thanks to demand for the iPhone, they have one third the market share of entrenched Samsung. Demand is so high, in fact, that 15 million people on China Mobile are using jailbroken iPhones at slow 2G Internet speeds just to have one despite the lack of official support. And Apple had to stop selling the iPhone in its retail stores because demand for them caused “crowd disturbances”.

Yes, definitely Apple that should be worried.

Sarcasm aside, the mistake being made here is one of predicting the future of smartphone market share in China based on a tactical supply side advantage, rather than on the demand side fundamentals. Of course Samsung is in a better position with respect to the carriers. And of course this is something that Apple should address as a high priority. And they are. In 2006, Apple had relationships with zero carriers and Samsung had relationships with all of them. But Samsung has been losing their distribution lead ever since the iPhone launched and Apple started collecting mobile operators.

Not China Mobile though. Maybe integrating support for China Mobile’s unique brand of 3G didn’t fit into Apple’s schedule, or maybe they just haven’t agreed on terms. Based on the couple of times that I was in meetings with China Mobile executives, I wouldn’t be surprised if it’s the latter. They own the mobile market in China. Whatever the details, Apple has made some tough trade off decisions and decided not to support China Mobile’s network. For now.

And while the supply side factor of distribution reach is important, it is not nearly as fundamental as the demand side factor of people wanting your product.

At this point some people will say that Mac was always nicer than Windows and that didn’t seem to help Apple much in the PC market. But Windows was dominant because of the demand side economy of scale (or network effect) associated with the app platform.

Others will say that since Samsung is Android based, it will create a Windows-like, app platform network effect, but they are confusing OS footprint with app platform footprint.

If the logic isn’t convincing, maybe a case study will be. Consider Japan. iPhone launched there in July 2008 on the number 3 mobile operator, Softbank. By September that year it appeared to be hitting a bump in the road. Sales were slowing and according to Yukari Iwatani Kane of the WSJ this was because of the high price point and a lack of features that the mobile-sophisticated Japanese market expected.

Last week we learned that Apple became mobile phone market share leader in Japan. Not share of smartphones, but share of all phones. This despite the fact that NTT DoCoMo, Japan’s largest mobile operator by far, still doesn’t carry the iPhone.

Samsung is definitely winning some battles in China, but unless it produces a product that can cause a riot, it will lose the war.

Technology First Movers do not Market Leaders Make

It feels a bit unfair to pick on one author for mistakes that so many technology writers are making all the time, but this particular article in Forbes is such a great example of the problem that it is impossible for me to resist.

The gist of the article is that Apple hardware is falling behind. The company is “under pressure” and will soon be “feeling the heat” from new technologies like Samsung’s foldable display. This argument is flawed on several levels.

1. The Many Faces of Samsung

There is not one “Samsung”. The Samsung division that makes display technologies is separate from and largely independent of the division that makes phones. The former will happily sell whatever they can to Apple, including fancy new flexible display technology.

In fact, you can bet that if their flexible displays are close to production-ready, then Samsung’s display technology team is actively trying to sell Apple on incorporating the technology into future iOS products. After all, there may be gazillions of Android phones being activated every day, but there is no single purchaser of specific component SKUs larger than Apple.

2. First Movers Take One for the Industry

History teaches us that being the first to use a new technology is hardly a guarantee of success. In fact, in the case of a technology that makes fundamental changes to user experience it is much more likely to end badly for the first mover. Apple was nowhere near the first company to make a phone with a touch display. It was the first to do it right.

Understanding this point is one of Apple’s strengths, having learned hard lessons from products like the Newton. So there is zero anxiety in Cupertino about Samsung’s phone team beating them to the punch with a foldable display. More likely, having hosted the Samsung foldable display sales team in Cupertino, Apple’s team has concluded that the technology is interesting and something to watch, but still way too risky to be incorporated into the iPhone or iPad. Besides, they can wait for other products come to market with this novelty and observe how consumers react. They can wait for the hardware technology to reach maturity, because you can bet that the first few generations will be far from consumer-ready.

3. Remember the Software

One common misconception about a technology like a foldable display is that the hard part is done once you have figured out how to make a display that is flexible. But that is only the beginning. A flexible display is a fundamental change. Until now, ALL computer and device screens have been hard, flat surfaces. All user interface design assumes that they are hard, flat surfaces. When this changes, the user experience implications are probably as extreme as they were for moving from punch cards to keyboards, or from keyboard only to keyboard plus mouse, or from physical keys to touch screens.

What often happens with new, game changer hardware technologies is that the first-to-market products combine the new technology with the old software user interface. So if we do see Samsung phones with foldable screens, the software UI will probably look almost identical to phones that have normal screens. It is only later when the user experience design is rethought from the ground up for a flexible display that the potential will truly be realized. And when this happens, it is very likely that a vertically integrated player, with their end-to-end control of hardware and software, will do it.

Samsung and software? Samsung and User Experience? Not so much.

Flexible displays will probably be big. First movers will initially generate some buzz, and later be ridiculed as the first products fail to meet expectations. At some point, perhaps a lot later when flexible displays are a forgotten technology, other companies will bring products to market that set reasonable expectations and meet them with delightful user experiences. These companies will get history’s credit for being first to market.

Weathering the Vertical Storm

The early mainstream successes of personal computing were products like the Commodore 64 and the Apple II. They ran apps from third parties, and supported third party peripherals, but the devices themselves were designed, engineered, manufactured and distributed by a single company. In other words, the suppliers of these computers were vertically integrated.

When the IBM PC was born, the relationship between Microsoft (operating system software vendor) and IBM (hardware vendor) ushered in a new, horizontal model. With hindsight, IBM made a blunder when they allowed their contractor, Microsoft, to retain ownership of the operating system (OS), but this was only one part of the equation. Crucially, a layer of software decoupled the OS software from the electronics hardware, and the PC’s brain was a micro-processor that was available from Intel, a 3rd party silicon vendor. These factors and others enabled the industry of PC clones that followed when Phoenix cloned the key interface layer between hardware and software (the Basic Input / Output System, or BIOS).

Lawsuits followed, but when the dust settled the industry was horizontal. And Microsoft owned the most important piece of the value chain — the OS. The OS was the key because it defined the app platform, and apps, in turn, are what created the network effect that exploded into 95% market share. Hardware, on the other hand, was “commoditized” because virtually anyone could make a PC from components. In fact, college students like Michael Dell started doing just that from their dorm rooms and grew companies that later eclipsed IBM in the PC market.

So in the PC value chain, there was intense competition in the hardware part, and little to no competition in the OS part. The rule is that value flows to the part of the value chain that is most “concentrated” (fewest competitors) because that is where companies aren’t pressured to lower their prices, and nowhere is this more apparent than PCs. Microsoft with the biggest profit margins we’ll ever see, and the PC manufacturers struggling to make a few points of margin.

Mobile Computing and the First Signs of a Cycle

In the late ’90s and early ’00s three precursors to mobile computing devices started to go mainstream, the phone, the personal data assistant (PDA) and the portable media player. Initially these were all vertical plays, but Microsoft saw the potential of mobile computing and got in early with a compact version of its Windows operating system, Windows CE.

The problem was that the market for mobile computing was nascent. Only the earliest of early adopters were prepared to do things like edit documents on a device, or send email from one, or download music to one. At a time where the thing people needed most of all to convince them to buy one of these devices was simplicity, Microsoft’s horizontal approach led to confusing products. Not only did they have too many features, they were confusing in that multiple companies were collaborating to bring them to you. It is a Compaq piece of hardware with Microsoft software and… what does it do again? Who do I call when I’m confused?

The devices that did take off were different. Simpler. Palm Pilots acted as an extension of your PC address book and calendar. Blackberrys delivered only your email. iPods were walkmans that could store all your music. The successful products were focused on one single value proposition.

This might have represented the beginnings of a pattern that we will see repeated whenever there is a fundamental shift in user experience, or a new computing modality. In the early stages, simple, vertically integrated products that minimize complexity are successful. I don’t only mean user interface complexity, I mean everything about the user experience. “What does it do?”, “Who makes it?”, “Where can I buy it?” and “Who will help me use it?” must all have simple answers.

How simple? Much, much simpler than you think. Very few people in technology understand the level of simplicity required to create a new category, so the odds are that you aren’t one of them.

Steve Jobs was one of these people and he hired a lot of the others. They made the iPhone, which catalyzed the mainstream smartphone market. Even Jobs’ description of the iPhone during his launch keynote reflects the need for simplicity. He teases the crowd by describing three separate dedicated devices, a new iPod, a phone and an Internet communications device.

“An iPod, a phone and an Internet communicator. An iPod, a phone… Are you getting it? These are not three separate devices. This is one device. And we are calling it, iPhone. Today Apple is going to reinvent the phone.”

Beautiful simplicity that launches Smartphones across the chasm.

Steve Jobs and the Horizontal Storm

With the iPhone, Apple is in its vertical element. The capability to design and execute on all aspects of the product is their home turf. This gives them the edge in the new mobile computing cycle. It represents their emergence from a long period where their model was not suited to the dominant horizontal model in the computing industry.

But how did they weather the horizontal storm that Jobs inherited when he returned to Apple in 1996?

  1. They hit bottom. When Steve returned, he could cancel products left, right and center without sacrificing much profitability. There was little to lose by trying a new direction.
  2. They focused their vertical chops on a profitable niche. A beautiful, differentiated iMac that had high appeal with a smallish subset of the PC market created a profitable base to start the return.
  3. They found and dominated an adjacent vertical opportunity. Who knows whether Apple or Jobs knew the potential of the iPod at first, but when it became evident they executed flawlessly doing what they do best: an end to end offering that took a technology based value proposition mainstream.

By 2006 they had the brand, the capability, and the resources to introduce a product that led the world into the era of mobile computing. And also, back to an industry where the dominant player is vertically integrated. I don’t think they sat around the boardroom plotting the overthrow of the horizontal model. They were probably just playing to their strengths and passionate in a belief that “if you love software you have to build your own hardware”. Their specific motivations are irrelevant though. The outcome is that their own product changed the industry rules and brought the game back to their home field.

Once Horizontal, Always Horizontal

Apple didn’t make a big wrenching change when Steve Jobs returned in 1996. It stopped trying to change and it returned to its roots — making profitable end-to-end products. Prior to that the company had bowed to the pressure of a conventional wisdom that Microsoft’s horizontal platform strategy was better, but their attempts to make a business out of licensing their platform wasn’t going well at all. There were any number of reasons, but most of them derive from the simple truth that making platforms wasn’t Apple’s bag. They hadn’t hired the kind of people that were good at it. Passionate Apple employees didn’t have their heart in it. It wasn’t in their DNA.

And here is a fundamental truth that so few people acknowledge. Building a business on a horizontal platform and building a business on a vertical product are completely different things. So although Apple and Microsoft have many superficial similarities — they both build operating systems, write application software and design user experiences — their differences run far more deep.

The horizontal platform is all about strategic collaboration with partners. This is a technical challenge and a business challenge. Technical, because you need to support all of your hardware partners all the time. Not only their current products, but their previous generation hardware too. And a business challenge because your economics are completely dependent on extracting profits from the value chain at your partners’ expense, while keeping them happy enough not to flee your platform.

Try to imagine the mind-bending complexity of building something as sophisticated as Windows for the thousands of distinct products that use it as their OS. Then imagine the constant negotiations with your big partners, and the massive programs you need to set up to manage the thousands of small hardware partners that you can’t deal with directly.

Apple deals with orders of magnitude less complexity when it comes to hardware supported and partnerships with other companies. Windows engineers can only test a new OS on a tiny fraction of the actual PCs that will run it. Apple can test a new OS on every machine that will run it. Apple has few partnerships, and when it does partner it does so when it has a lot of power in the relationship. It doesn’t have partners that are also customers. It doesn’t have any confusion about who’s more important. It’s you, the consumer. Microsoft, on the other hand, tries very, very hard to do what you want, but it also has to answer to Michael Dell.

What Apple does have is the ability to execute almost all aspects of the end-to-end product. And they can afford to have an undiluted, single minded focus on the people that purchase their product. Unlike Microsoft, it has no confusion about who the customer is. Engineers, designers and managers spend all of their time think about products, the problems they solve, and the people that use them.

All of this is to say that when people talk of Microsoft becoming like more Apple and adopting a vertical strategy — or back in the day, Apple becoming more like Microsoft and licensing their OS — they completely miss the massive, wrenching change that this would represent. Microsoft will have product successes, like Xbox, but it is, and probably always will be, a platforms company.

The Platforms Guy

If Apple is made in Steve Jobs’ Image, then Steve Ballmer is made in Microsoft’s image. The large, blustering, bumbling exterior belies a very smart, capable core that is struggling to weather two storms: (1) the shift in power from from client to cloud and (2) the return to dominance of a vertical industry structure in the fastest growth area of computing: mobile.

Steve Ballmer can’t take many lessons from the way Steve Jobs weathered the horizontal storm. Microsoft’s remarkable profitability means that it will not hit bottom for a long, long time. It can’t easily do the iMac trick and retreat into a niche, because horizontal platforms need market share and scale in order to succeed. And it can’t do the iPod trick of finding a nascent category to dominate, because horizontal platforms don’t do well in the early stages of a market when consumers need clarity and simplicity more than anything else about the value proposition.

Microsoft has to do something different. It has to find the next important horizontal platform, accelerate us towards it, and then dominate it completely. Is it in the living room? Maybe, so let’s build the leading entertainment platform that connects to the TV. Is it the preeminent starting point on the web? Maybe, so lets make MSN the most important portal. Is it the communications network? Maybe, so let’s buy Hotmail (then) and Skype (now). Is it the index of the web? Maybe, so we better not cede that to Google. Can we make it the mobile OS? Maybe, so let’s get in early, build one and try to replicate the PC licensing model. Is it the platform for web applications? Maybe, so we’ll create a platform for the cloud called Azure.

Is this hard? Yes. It might be impossible. Is it crazy? No. If you are a platform company and you plan to stay true to your nature, then this is a completely rational strategy. It does, however, take its toll.

First, Microsoft can’t go anywhere quietly. As soon as the giant appears at the door of any category, every journalist, pundit and competitor knows exactly what Microsoft has in mind. There is very little scope for subtlety. Gulliver can only make a frontal assault, because the front moves to Gulliver.

Second, anything but domination will be considered a failure. Both internally at Microsoft and by the press.

And third, there will be many failures along the way. Some of them successes by any other standard, and others hopeless failures by any standard. Either way, they will represent opportunity after opportunity for gloating pundits to ridicule Microsoft. And to ridicule the personification of Microsoft, Steve Ballmer.

Opportunity in Obscurity

Ironically, all the ridicule does have a silver lining. Although Microsoft is still making money hand over fist on Windows and Office, it has all but left the industry consciousness. Mary Meeker’s “Internet Trends” presentation at the Web 2.0 summit in 2010 mentioned Microsoft precisely once, more than 30 slides in, as one of three companies introducing innovative gaming input methods. Notably, it isn’t listed in the Global Public Internet Companies about 5 slides earlier. The surprise isn’t so much that Meeker reinforced perception rather than reality, but rather that almost no-one pointed it out. In an even more striking contradiction, Meeker’s 2011 presentation omits Microsoft in her list of “mega-leaders”, but acknowledges that they are behind only Google in global monthly unique visitors in the very next slide.

More recently, Tod Hixon’s Forbes article, “The Post-PC Era Starts to Make Sense”, doesn’t mention Microsoft at all. Even if you don’t believe that Windows Phone 7 and XBox warrant inclusion, surely Microsoft at least deserves a mention for representing the PC-oriented incumbent. Hell, even IBM makes it into Hixon’s piece.

The silver lining is that competitors will underestimate Microsoft’s ability to return to relevance. It is looking less and less like Gulliver. At the same time, it is continuing to adapt to the new realities of cloud and mobile computing.

In the cloud, Microsoft has its Azure platform. Here are two interesting things about Azure. First, it is apparently one of two 3rd party cloud platforms providing infrastructure for Apple’s iCloud (the other is Amazon’s AWS). The cloud is the area where Apple’s DNA is weak, and that is one thing to draw from this. It also brings into focus the extent to which Google has replaced Microsoft as Apple’s nemesis.

The second interesting tidbit about Azure is that it has recently released support for Node.js, a bleeding edge open sourced technology for building code that runs on servers. This is a different Microsoft.

In mobile, Microsoft has Windows Phone 7, widely acknowledged to be an excellent user experience, but struggling with market share percentage in the low single digits. Microsoft has partnered with Nokia, another powerful company that pundits have already counted out. The result of their collaboration, the Lumia 700 and Lumia 800 phones, are receiving extremely positive reviews from everyone who sees one, and initial market response in Europe is good. Lumia 900 is eagerly anticipated in the US, and tech bloggers are raving about it. Having a good product is definitely the first step, and Microsoft has achieved that.

In a Mirror, Darkly

With competitors looking the other way, Microsoft’s biggest obstacle is itself. The cash cow businesses will not cede the throne at Microsoft easily. My knowledge of internal Microsoft politics is more than two years stale, but I would guess that even today, Windows, not Windows Phone or Windows Azure, is the more powerful internal group.

Perhaps an even more difficult challenge for Microsoft is switching to the mindset of the underdog. Microsoft was so dominant for so long, that hubris is part of its DNA. In the quest to find a market for Windows Phone that is “Microsoft large”, executives are likely to favor a massively broad approach that maximizes addressable market and distribution reach. In the process they will smash themselves on the battlements of iOS and Android.

If, on the other hand, they have the humility to focus on a few key market segments that they can dominate completely, then they might gain a solid foothold that positions them for a comeback. Microsoft is more focused on user experience than Google, and it is more experienced in the cloud than Apple. That powerful combination may end up weathering the vertical storm.

Disclosure: I was a Microsoft employee and I still own Microsoft stock.

The Growing Business of Monetizing Other People’s Content

For the last couple of years I have been an active user of Instapaper, Marco Arment’s excellent service for saving online articles for later reading. According to his press kit the service has more than 1 million signed up users and is profitable.

Instapaper is successful for good reason — it offers a lot of value to users. If you have the iPhone or iPad app and you have opened one of these apps while connected to the Internet, then the text of the articles you have saved will be available from that device even if it is offline (e.g. in the subway). The Instapaper UI allows you to read the articles without the distractions that the original publisher of the content saw fit to place near the article. Ads, for example, do not follow articles into Instapaper. The saved articles also serve as a useful archive of the things you found interesting enough to save for later.

So it was from the perspective of an appreciative end user that I recently started questioning the Instapaper model. After all, Angie and I make our living as a small web publisher. There are probably YLF readers using Instapaper to save Angie’s articles for later. When this occurred to me I immediately experienced a rush of involuntary emotions. The knee-jerk reaction of someone with a small online business where advertising revenue puts food on the table, who has invested time in designing and building the site, and who watches his wife spend many hours researching and writing most of the site’s content.

I did the healthy thing. I tweeted.

I’m torn about @instapaper & @readability. As a user I like them. As a publisher I feel that they monetize our content without permission.

Marco responded:

Instapaper’s mechanics respect publishers and their pageviews more than any similar service.

I agree with this. On Instapaper’s information for publishers page you can read about preparing your webpage so that Instapaper will parse it correctly (as a tech-savvy user you can even customize the way Instapaper represents ANY site when you save it for reading later). And if you object to Instapaper using your content, you as a publisher can opt out.

And as Marco pointed out to me in email, Instapaper has engagement benefits for a publisher:

The most common use-case by far is this: someone opens up a page on the publisher’s site, looks around for a bit and reads the first few paragraphs, clicks the Read Later bookmarklet, then reads the rest in Instapaper. This gives most publishers as much value as any other viewing method, and since it helps people read longer articles attentively instead of just skimming them briefly and clicking away, most publishers have noted increased engagement and return rates from Instapaper users. (In fact, almost every staff member at publishers I’ve spoken with has used Instapaper themselves.) Publishers almost universally agree that this use of Instapaper is beneficial to everyone.

But here’s the thing: However much Instapaper has our benefit in mind, it is monetizing our content in a way that we did not explicitly approve. And while there might be an engagement benefit for us as publishers, as a user of Instapaper I know that I spend much less time in the publisher’s world and more time in Instapaper than I used to. After reading a Daring Fireball article on John Gruber’s site, I’m very likely to read another DF article. After reading a DF article on Instapaper, I’ll probably go from that to something completely different. Bottom line: The benefits for publishers argument isn’t a slam dunk either way.

And it isn’t like RSS. Although many people consume our content in RSS readers, it was our choice to provide a full RSS feed in the first place, and if we choose to do so we can put ads in the feed.

Drawing the Line

In the market of read-it-later apps, Marco is the good guy. While speaking to him it becomes clear that he has been thoughtful about the impact on publishers when designing Instapaper. His competitors, on the other hand, are going much further in repurposing other people’s content without much consideration for the original publisher. Primarily this comes in the form of being able to click links and save them for reading later without ever having gone to the publisher’s site. When this happens, the publisher never sees any advertising revenue and doesn’t get the opportunity to present their content in the context of their own design.

Then there is the feature of rendering multi-page articles in a single page. We don’t do multi-page articles on YLF, and I am sometimes annoyed by reading them on other sites, but I do respect the publisher’s right to offer their content in this way. We as readers can then make the trade off and decide whether the value of an article overrides the annoyance of clicking multiple pages to read it. Despite a lot of user pressure, Marco does not offer the feature of piecing multiple page articles into a single page (his competitors use this as a key differentiator).

Update on 3/4/2012: A few days ago Marco updated the Instapaper bookmarklet to support the saving of multiple page articles. It seems the competitive pressure finally outweighed the negative impact he knew this would have on publishers. This is completely understandable, but underscores the point that read-it-later services will struggle to prioritize the interests of the publisher.

So as a publisher the concern isn’t as much the impact of Instapaper as it is the impact of a wide range of parasitic apps that have repurposing your content as their business model. If it is ok to ingest a publisher’s content without their consent and then monetize that content, where does one draw the line? If the argument is that the app offers a better experience than the original publisher, then is it ok for me to ingest the content of the and present it in a nicer, more convenient interface alongside my own ads?

A Middle Ground

A couple of years ago one of our readers brought it to our attention that a small company was copying Angie’s blog posts (almost verbatim), compiling them into a PDF document, and sending this document to their members as part of a paid subscription. I don’t think anyone would condone this. It is a blatant example of someone stealing and monetizing our content. One clear difference between this and a read-it-later app is that these people weren’t even crediting the original source of the material. But let’s say they had clearly noted that all the content was taken from YLF and was written by Angie, would it have been ok to package it and sell it to their members in a format that their particular readers found more convenient?

I don’t think so. I think there must be a model where publishers have more say in how their content is monetized AND and where users get the benefits of read-it-later features and apps. Napster showed us a better way to distribute digital music, and then iTunes showed us a way to do it while respecting the creators of the music. Will there be an iTunes of online publishing?

One Instapaper alternative, Readability, has already tried a middle ground that includes publishers in their monetization model. In Readability’s original model readers could optionally become subscribers, and then 70% of their subscription fee was paid to the original publishers of the articles that were saved by the user. I don’t think this is the solution. For one thing, publishers had to know about Readability and sign up to receive their payments. And while it felt like nice gesture on their part, we as a publisher did not agree to participate in Readability on these terms. We have not had any involvement in the pricing of the subscription and we did not negotiate our cut. In a way, Readability was on more shaky ground than Instapaper because it is making a tacit acknowledgement that writers and publishers should participate in the economics of a save it for later reading service.

Also, the Readability subscription is now free and I can’t see any mention of the original model on their website, so that experiment doesn’t seem to have been an unqualified success. I talked to a Readability representative and they told me that although it is still available “We don’t really mention the paid option on the site, since paying is entirely opt-in.

Update on 2/24/2012: Max Fenton of Readability has subsequently clarified that once a reader is logged in, the opt in subscription option is mentioned on every page of the site except articles themselves. Max tells me that they learned through feedback that explaining the payments model before a new reader had signed up was complicated. For me the salient point is that, whatever the reason, the publisher support aspect of Readability’s model is far less prominent today than it was when I first learned about them. Also, I originally stated that the writer share of opt-in subscriptions was 30%, but it is 70%.

The App Response

Read-it-later apps can function because web pages are designed to be machine readable. Ultimately human readable of course, but before you see a page, your browser needs to interpret its native language of HTML, CSS and Javascript. This means that apps like Instapaper “scrape” and ingest the content. Few publishers created their web pages with this in mind. It is just a side effect of the openness of the web.

Traditional publishing business models were in trouble long before read-it-later apps came along, but new publishers with lower overheads are doing quite well on an ad-supported basis. That might change as more and more of a publisher’s content is consumed somewhere other than the place they intended. These new publishers might find themselves looking at some of the same options that traditional publishers have been wrestling with.

The first of these options is the Internet paywall, which major publishers like the Wall Street Journal and the New York Times have embraced in different ways. A more recent approach rides the resurgence of native apps that was catalyzed by Apple’s iOS App Store 1.

Publishing via an app allows the publisher to tightly control their user experience and business model. And it allows them to side-step the ugliness of any public conflict with apps that are scraping their content.

On the other hand, by going to an app the publisher gives up a medium that was designed to deliver hyperlinked content to the masses. And they give up the ease of web based sharing and all the viral advantages associated with that. There’s a lot of baby going out with the bath water.

Trade offs notwithstanding, this shift is happening and in the process some content is migrating from the web to native apps. Of course, the content will still be flowing over the Internet, probably even using the web’s HTTP protocol, but the point is that it won’t be readable in a general purpose web browser, but rather in the custom, publisher-specific web browser that is the publisher’s app.

The Web as Solution

I like apps for many things, but I also want to live in a world where the open web is alive and well. When people talk about the web versus apps, it is surprising how often they discount the single, incredible innovation that defines the web and is not available to apps: the hyperlink. Apps create silos. Delightful silos, but silos nonetheless. They are connected to the web, but they are not in the web. You can link from an app to a web page, but you can’t link from the web to an “app page”. You can’t find an “app page” using web search unless the app developer creates a page that mirrors that content.

I think that the massive advantages of being in the web as opposed to being attached to it will keep many publishers there. And if the hyperlink-fueled  growth of the web itself is anything to go by, they will thrive, even as apps and services scrape and monetize their content. Their user experiences will improve, and they will figure out how to make the in-situ consumption of their content so compelling that readers will be loathe to read it elsewhere.

And for inspiration they need look no further than the class-leading read-it-later apps like Instapaper and Readability. These apps succeed because they solve a problem for people. There is no doubt that much of their value proposition exists precisely because publishers have such broken user experiences today. Publishers should ask themselves why people need a third party service to get a “reading friendly” experience. And they should think about how they might partner with offline consumption services like Instapaper in a model where everybody, especially the reader, wins.

  1. I don’t view opting out as viable option because (1) it probably isn’t practical to for publishers to play whac-a-mole with emerging read-it-later apps and (2) opting out paints the publisher in a negative light when the annoying message appears telling users that this publisher doesn’t allow their content to be read later.