GP vision
Recommended by: Zhu Xiaohu
Invisible intimates
Eugene Wei
From: Blog Remains of the Day
Jinshajiang Venture Capital (welcome to click on the link at the end of the article to read the original article)
Recommended language:
Eugene Wei is the first analyst in Amazon’s strategy group. He believes that the most important part of strategic planning is to find and destroy the so-called invisible ceiling in advance.
This article analyzes the ceiling and solutions of Amazon, Facebook, Snapchat, Instagram, Twitter and other giants. It’s thorough and thought-provoking.
What was once the core advantage turned out to be the biggest obstacle at a certain point. Some are limited by the product itself, some are lost to The Times. Destroy a ceiling to unlock a new round of growth. There is always a next ceiling, but the strong never stop…
Sun Tzu’s art of war goes: Know yourself and know your enemy, and you will win a hundred battles with no danger of defeat. Knowing each other and knowing each other, one wins and one loses; Without knowing your enemy, without knowing your friend, every battle is perilous.
I was amazon’s first strategic planning analyst, which was my first job at Amazon. Strategy is contrasted with accounting, which records the past, and accounting, which focuses on the future. We do forward-looking analysis over a range of time horizons, from regular monthly and annual forecasts to five-year and 10-year forecasts for financing and corporate strategy.
One of the hardest numbers to predict is product penetration. Amazon has gone public, but that doesn’t mean strategic planning can be slackened. Jeff Bezos has publicly said, “In the short run, the stock market is a voting machine, and companies’ stock prices fluctuate, but in the long run, it’s a weighing machine that ultimately reflects the true value of companies.” Your job is to learn as much as you can about what will happen in the future. Every chief financial officer of a public company will tell you how much they value forward guidance at work. Because of the information asymmetry, the institutional analysts responsible for your firm adjust their investment forecasts and layouts for clients largely based on quarterly earnings calls. It’s not just that wrong forecasts affect share prices, but that if they’re wrong, it means you have no idea where the business is going, which is definitely more devastating in the long run.
It didn’t take long for me to discover that our forecasts for months, quarters, and even a year were accurate and accurate. And it’s the longer term predictions that we’re most confused about. Every successful business model goes through the famous S-curve, and most companies and their investors focus on finding that tipping point. But equally important, and less appreciated and studied, is precisely the less pleasant inflection point at the end of the S-curve, when growth begins to slow.
One of Amazon’s great strengths is that we can easily predict the size of our market, as a rule of thumb we can do if we capture 10% of the global book market. You can be even more optimistic, you can forecast bigger, but you still tend to be conservative financially.
As soon as I joined Amazon, I was working with a bunch of Mbas on various business plans, including music, video, software, magazines, and international business. I’m starting to think that the total scale we can do in the long run is the sum of the segments we can cut off from the retail track.
However, product penetration remains a mystery. In our analytical model, an important prerequisite for product adoption is exposure. Because if someone has never heard of Amazon, they can’t become a customer. The channels of exposure or reach are well established: from other affiliate sites (we call them partners), from portals (AOL, Excite, Yahoo, etc.), and word of mouth. (Note that this was after email and before social networking, so word of mouth wasn’t nearly as effective as it is today.) Of course, awareness can be measured through sophisticated market research techniques.
So the question remains: how do you predict whether someone who already knows about Amazon will make a purchase? Why do some of them successfully convert to buying users while others don’t?
For many startups, and even larger tech companies, it’s a mystery when they reach the “S-curve” of boom-bust growth. I think it’s entirely possible that something could have been detected earlier. The advantage of this is that early detection and early resolution. We have paid too much attention to product-market mismatch, but as the company grows to a certain point, we should spend more time focusing on and addressing product-market mismatch.
In my view, the most important part of strategic planning is to find and remove what I call the invisible ceiling. The so-called invisible ceiling is the growth bottleneck that we are bound to hit on the current path of the company. This concept is very important to many people in the company: whether it’s the CEO, whether it’s the product department, whether it’s the accounting department that I mentioned earlier.
The Amazon’s invisible ceiling
Fortunately for Amazon, we found one of the most important ceilings early on. Amazon continued to grow for years because we figured this out and addressed it early enough. If we hadn’t adjusted our strategy then, our growth would have stalled, and much of that would have depended on a single factor.
We try to find and destroy this enemy in two ways. For people who have already made a purchase, we used to pop up at the end of the shopping cart process after they placed an order and ask, why don’t you buy things from Amazon more often? And for people who haven’t bought anything on our platform, we ask third-party companies to do market research and ask them, why don’t they buy something from Amazon?
The two paths agree surprisingly well on the same point. I don’t even need to recall the conclusion, because I think this factor is still very important in e-commerce and related businesses today.
It’s freight.
People hate paying freight. This doesn’t sound new, even self-evident. But I think really understanding that was the key to amazon’s continued growth for many years after that.
People don’t simply hate paying freight anymore, they actually hate it to an irrational degree. We know this because our first solution to this problem was to tell users at the shopping cart and checkout that it was still cheaper than buying books at a local brick-and-mortar bookstore, even with shipping. At the time, most Amazon customers didn’t pay sales tax. And we’re not even talking about the cost of a trip to the bookstore, the depreciation cost of a car, their time, etc.
People do not buy this rational analysis. In general, people are terrible at evaluating the value of their time. Perhaps because for most people, spending time and getting money are separate. The vast majority of our time is spent without immediate feedback.
Rich people have faster and more direct feedback on their time returns, so they tend to be better able to value their time. It also explains why the first thing most super-rich people I know do when they get rich is hire a chauffeur or start flying private jets. For most ordinary people, the opportunity cost of their time cannot be measured in real time.
You can’t imagine how relieved it is as a product person to find that single hurdle to overcome. This applies to anyone who wants to solve a problem. Just as certain diets promise to lose weight if you cut out sugar or carbs, it often feels good, not because you’re actually losing weight. More than half the pleasure comes from feeling like you can solve a complex problem with a simple solution.
Amazon has been trying for years to solve the problem of people hating shipping. We offered a shipping savings plan: Free shipping for eligible purchases over $25, which includes most amazon products.
The problem with this scheme is obvious: users will reduce the frequency of their amazon purchases and save up enough for free shipping to make the next purchase. In certain cases, the strategy of getting users to consume less of your product or service can be beneficial in the long run, but Amazon is clearly not one of them.
Then we came up with Amazon Prime, which sells a membership for free shipping. To state the obvious, there must be a cost to shipping physical goods. So how Prime affects sales and average unit prices makes a huge difference to the financial model used to judge Prime.
Thankfully, Jeff decided to skip the test and go straight to execution. It is not uncommon for tech companies to put growth first and then monetise their businesses over the long term. But this model is much more acceptable in social networks than in retail businesses, where the unit economy model is clearer. The more you sell, the more you lose. This business model is definitely not sustainable. We’re frustrated that people confuse this with Amazon’s business model, and the skepticism continues to this day.
It turns out that you can get people to prepay shipping through membership schemes like Prime, and they’re more than happy to make subsequent deals. Yes, on some orders, on some users, we didn’t make money, but across the network, the huge increase in the demand curve was amazing and changed the rules of the game.
And, as Jeff always says, in the long run, you can always make a profit by tweaking some of the details. For example, if something is really big and heavy, you can ask for a small surcharge or remove it from Prime free shipping. Amazon now has some items marked as “add-on items,” which users need to buy along with enough other items to get free shipping, so they can be shipped together rather than individually.
Jeff used the same “fix it later” strategy in our early days, before we had a good return tracking system in place. In the early days of Amazon, there was a time window when if you returned a box of books to us, we could not easily tell whether you bought the box of books from Amazon or not. So we choose to trust you, unconditionally. Then a woman took advantage of our loophole and returned box after box of books. Given our relatively limited software resources at the time, Jeff said we would just skip it and fix the bug later. It was such a painful time that our customer service staff spontaneously shared the woman’s name internally so they could notice the woman’s return request before our software and return mechanism improved. It’s like a “Beware this customer” note on every monitor. Ma ‘am, wherever you are, you deserve our exclusive treatment, because you have shown the ultimate entrepreneurship in taking advantage of our loopholes!
The Prime system has high barriers that other retailers will struggle to emulate, both economically and logistically. As mentioned earlier, to deliver physical goods, there are transportation costs. So delivery startups like Postmates, with their challenging unit economy model and people’s aversion to shipping, are walking on thin ice. It may not take off until self-driving cars or drones make it much cheaper to deliver physical goods.
Also, few customers can buy enough from retailers other than Amazon to make prepaid shipping like Prime worth the price of the ticket. Even if they did, our economic model for delivery would be optimal, based on Amazon’s economies of scale and deep understanding of inventory allocation.
In essence, we did one thing right: we identified and solved the core problem of the e-commerce ceiling long before it was even close to hitting the ceiling.
The invisible ceiling is invisible
The obvious problem for many companies, however, is that they have a hard time finding their ceilings when they are trying to create entirely new businesses and services. That’s because their business doesn’t have clear, trackable metrics like Amazon’s, such as its reach based on the size of the global book market.
If you look at social networks, Facebook, Twitter, Instagram, Snapchat, where would their ceiling be?
Some of these constraints on growth are easy to spot. Instant messaging apps and similar tool-centric social networks are, in many cases, geographically constrained. These apps are based on real-world social graphs and are mostly geographically aggregated, so there are winner-takes-all situations in many countries, such as Kakao in South Korea or Line in Taiwan. There will also be a combination of geography and policy factors, such as wechat’s dominance over all its competitors in China.
Other constraints are much more subtle. It takes more product insight, or even what some call intuition, to feel the core constraints before they hit the ceiling. The real challenge for employees and investors is to find their OWN PMF, which is a product-market mismatch followed by a deep understanding of the product-market mismatch.
Without internal data and research, it was difficult to analyze a company’s potential ceiling using only public information and my own product intuition. But we can take a quick look at some of the industry’s biggest players and see what’s at the core of their ability to reach the ceiling. Of course, these companies are already big, and they may have many ceilings, but I’m going to focus on just one of the most critical macroscopically. You can use this approach to analyze startups as well, but it’s important to note that there may be a difference between the initial strategy for product-market matching and the strategy for finding the ceiling and trying to break through it.
Let’s start with Twitter. Its status is far from ideal, and in that sense it may be the most depressing of all tech companies. Twitter’s user growth has been stagnant for a long time, so it’s safe to say it’s hitting the ceiling with all its might. On quarterly earnings calls, management simply can’t say if, when or how this will change, and they collectively have no idea what’s going to happen and nothing they can do about it.
A popular view of Twitter is that giving new users the experience of power users or existing users quickly helps convert them into active users, a strategy that is thought to work for most social networks. Most of the social networks that are growing strongly have achieved a key metric that unlocked their PMF — product-market matching. For example, when you have 30 friends on Facebook, you are locked into Facebook. For Twitter, the key metric is the number of accounts you follow, and when you follow enough accounts, you can generate a stream that suits your tastes.
That meet the conditions for productive pattern matching theory in silicon valley is all the more markets than in any other place, such a story between all employees, the board of directors, and other rich scientific and technological elite telegenic cluster, so it’s not surprising, when enterprise touch up to increase the ceiling, someone mentioned that this kind of solution.
The idea is a bit like Geoffrey Moore’s “crossing the chasm”, that is, going from seed to mainstream is to convince more people to use the same product or service as the seed.
But in the case of Twitter, I don’t think this theory applies. Based on the current form of the product, I don’t think there will be any meaningful user growth, and product tweaks on top of that won’t lead to new growth. The longer they wait to realize this, the longer they are stuck in a vicious cycle.
Sometimes, PMF, or product-market matching, that works for early adopters is only for early adopters and can’t be pushed to the wider mainstream because other people don’t want or need the product. In this case, the key strategy to unlock growth should be to segment users and create different products for different users.
Mistaking one business for another can be deadly, because the strategy to reach the market is very different. A common symptom of this mistake is that you simply cannot see the ceiling of your business because you do not have a proper and deep understanding of the nature of your business’s product-market mismatch.
I believe Twitter’s core experience has reached the majority of people around the world who like it. Let’s take a look at the core attributes of the Twitter product. (I separate the concept of A Twitter product from a Twitter service, which refers to the utility properties that deliver public messages.)
Twitter is heavily text-based, 140 characters or less, and now 280 characters or less, displaying tweets from the accounts you follow in an algorithmic order in a vertical scrolling stream. In this study, we default the algorithm order to time.
For Twitter’s followers, many of whom are heavily dependent on information, the core of product-market matching, like any other technology product these days, is addiction. Because the text is short, if the user finds it boring, he can scan it and easily scroll to the next one. Finding tweets of interest in an almost random order is a proven dopamine-inducing experience. Rats get their food by hitting a joystick, and powerful Twitter users get their tasty text food by pulling up and down the scroll bars on their phone screens.
For people who are heavily dependent on information, text has an incomparable speed advantage over photos, videos or music. Scanning text quickly to get information can bring them an excellent user experience with controllable speed, which is different from other media with fixed speed (especially video media, for people who are heavily dependent on information, they hate video because they cannot quickly scan to get information).
Over time, the cycle tightens and accelerates through the interactions of all users on Twitter. Fan likes, retweets, and other forms of feedback motivate users to write more tweets that generate patterns of positive feedback. The ideal tweet, one that gets a lot of positive feedback, generally has the following properties:
-
Short and concise. Character limits, like fortune cookie quotes, encourage this style of text;
-
Amazing words. It can be a different idea, or it can be something new in a new bottle.
-
Give their loyal fans with high recognition, directly to the readers beat chicken blood;
-
Take a swipe at someone your fans don’t like.
-
Tweets by celebrities, especially those with large followings who don’t make a living on Twitter, like Chrissy Teigen or Kanye West;
-
The topic should be one that most people think they know or have something to say.
Of course, the attributes of the ideal tweet also vary from group to group on Twitter. Unlike intellectual tweets, which have a different style from NBA groups, the point is that the cycle of inertia within each group also gets stronger.
The problem is, for those of you who are not Twitter users, all of the above desirable attributes for early adopters are unreadable and uninteresting. A lot of people don’t like the text-heavy nature of Twitter. In fact, most people do.
One of Twitter’s recommendation algorithms is to post content that existing heavy users find interesting in the trending posts, often leaving new users confused: Why chase it? Why are these conversations so hard to understand? (In fact, even for Twitter aficionados, reading every buzz is a challenge.) Why do people make such an effort to analyze the context of tweets?
It’s very harmful to assume that the user feels what you feel all the time. Because the person making the product is likely to be a member of the target audience, the initial inspiration for building the product comes from intuition. However, if you’re building a product that needs a niche audience, continuing to adopt the traits that early adopters prefer can lead you in exactly the wrong direction, and the company can quickly hit a ceiling without realizing it.
In fact, the concept of “crossing the chasm” is conditional. If the gap is as big as a chasm, the same product cannot bridge it. On the other side of the divide, on the other side, is a whole different world, with a whole different set of people and needs.
I’m a heavy Twitter user, and I recently received a tweet congratulating me on my 11th anniversary on Twitter. But the rest of my family, from my parents to my siblings to my girlfriend, including my nieces and nephews, have all tried and abandoned Twitter. Because Twitter doesn’t satisfy any of their deep needs.
I’m not surprised. It seems to me that the die-hard users of Twitter are journalists, techies and intellectuals who are already heavily dependent on information. For them, opening Twitter is like wearing Cerebro, connected to thousands of brains around the world, as if the cerebral cortex were spreading out across the world and sitting on top of millions of brains around the world.
SHH, be quiet. I’m reading Twitter.
For this group of people, Twitter has become an integral part of their lives and work, and it speeds them up as if they were riding a bicycle for information and intellectual exchange. One of the things Twitter does best is provide a sense of knowledge sharing in near real time, satisfying some needs like SoulCycle or Peloton. For people, a sense of communication can also lead to a sense of gain.
If my hunch is right, then all the iterations around Twitter’s minutiae won’t do anything to boost user growth. This iteration of the product does improve the experience of current users and increase activity, but does not address the core attributes of the product. So, no matter how iterative, it’s still not going to work for people who have abandoned Twitter. They tried to enter the world of Twitter and immediately found it to be a community of tech geeks, La Croix drinkers talking about bitcoin and meditation.
The good news is that the Twitter service, a public messaging protocol with a one-way tracking model, could be the basis for many other potentially attractive products. Finding the right track prevents you from wasting time on the wrong strategy.
Unfortunately, one of the main ways to build new products on top of agreements is through third-party developer programs, and third-party developers have long been treated like stepmother’s children on Twitter. And somehow, Twitter’s own internal product development has been incredibly slow. With Twitter’s well-established public messaging protocols and data sets, a vibrant third-party developer program could have made development much more efficient.
[but also please note, I quite agree with some technology companies limit others call API interface to make a clone products. No company shall have the duty to open interface, convenient others on the basis of a competing goods. Most people may not remember, amazon’s first service is assigned to the partner, let they sell goods to site. Someone used that mechanism to make a bunch of clones of Amazon’s sites, and Amazon Web Services had to evolve into something else.]
I also wonder if Twitter’s decision to shut down third-party developer access is meant to pool and enjoy all the advertising resources and value. All of these problems can be solved by tweaking third-party development plans. Third-party developers have two options:
One option is to insert an official Twitter AD for every number of tweets. That leaves the possibility of Twitter supporting third-party Twitter clients like Tweetbot, whose advertisers may compete with Twitter’s own advertisers. Some third parties may come up with a better experience than Twitter itself, but that will be good for Twitter’s total AD revenue.
The second option is to pay a flat fee for each number of tweets. This will force developers to come up with their own ways to monetize the cost. But at least that option should exist. No doubt there are enterprising developers out there who can think of good monetization scenarios, such as business research.
In short, Twitter products have hit the ceiling, but protocol-based development platforms still have potential.
Snapchat
Snapchat is another example of approaching the growth ceiling. But unlike Twitter, I don’t think the ceiling lies in the core attributes of the product, but in the generational divide of users.
Making the interface simpler and friendlier might help a little, but not much. In fact, it was the complexity and opacity of Snapchat’s interface that made it popular, intentionally or not, when it first launched. Snapchat came along at a time when parents were flocking to Facebook in droves, and Facebook had been around for a long time, with its early young users accumulating traces of social content, some of which needed to be erased completely from their parents’ eyes.
Snapchat was born, a perfect match for a service that not only deleted content by default for short periods of time, but also had an interface that was confusing and confusing to intrusive parents. In fact, I think for many products, a cluttered Easter egg looks better than a clean, elegant, easy-to-use interface, and in the long run, this is the pinnacle of UI design. (We’ll talk more about that another day.)
I previously wrote about selfies as a second language. The root of this phenomenon is that for a generation of young people who have grown up with smartphones with front-facing and rear-facing cameras, the most efficient means of communication is no longer a keyboard, but a camera. But that doesn’t apply at all to older users, whose first choice isn’t a selfie. So Snapchat’s camera as the default interface is a bold choice, and because of this design, older people will never choose Snapchat as their first chat tool, no matter how Snapchat moves and aligns the other button panes.
More importantly, I believe that every generation needs its own territory, which can be acquired cheaply and left behind in a short period of time. This applies in both the virtual and physical worlds.
If you look at the habits of older Snapchat users, you’ll see a preference for Stories. Younger users are more likely to use peer-to-peer photo sharing. (I knew, of course, that the peer-to-peer messages I received might have been sent in groups, but the product design concealed that group activity and still made me feel one-on-one.) From the example of Snapchat, it is not difficult to find that the same product can satisfy different user groups with different functions. Of course, some user experience will be sacrificed in this way.
On a deeper level, I think the need for instant information over a person’s lifetime changes with age. Young years, especially adolescence, are when the value of instant information is at its highest. As we get older and time slows down and nostalgia begins to creep in, the value of permanent content, especially those that record the good old days, slowly rises. And crucially, as people get older, they become more accustomed to managing their public image, reducing the need for immediacy.
In short, my point is that making the interface simple and user-friendly is not going to help break Snapchat’s growth ceiling. There is a difference between the underlying reason for the interface and the core factor behind the growth ceiling.
The good news for Snapchat is that I don’t think Facebook will appeal to young people either. It wouldn’t have helped if Facebook had copied every feature of Snapchat. The bad news, of course, is that Snapchat is also failing to attract older users. Which raises an interesting question: Will Snapchat’s users stick with it as they get older? And, for the next generation of young people, when they have their own smartphones, which social network will be their favorite? Will they, like generations before them, need a new social network for their generation? Sometimes I think they just need a name that hasn’t been taken by anyone before. Who wants to be Joesmith 43213 when you can be the first Joesmith in a new social network?
For Snapchat, however, the competitor that should worry more than Facebook is Instagram. It emerged, like Snapchat, after Facebook, giving it the opportunity to become a social network for a new generation of young people to tap into and tap into untapped social capital. And Instagram is a definite, more image-oriented social platform, which fits nicely with the new generation’s preference for visual information.
When Messenger copied the Stories feature, it felt like a middle-aged couple trying to dress up as cowboys for Coachella. And when Instagram added Stories, it was a perfect fit, addressing the lack of content creation for young users, a demographic that overlaps so closely with Snapchat. The problem for Instagram at the time was that users were so focused on the quality of their posts that they were Posting less frequently. With the addition of Stories, a new mechanism is created where content published in the Stories module is not forced into the news stream, and its short timeliness allows users to create and share content more freely and without pressure.
In order to attract more users, more usage scenarios need to be expanded. This is a common path for social networks and for any product in general. It is almost impossible for a product to stick to its original product form without ever hitting the growth ceiling. Compare Today’s Facebook to the original Facebook; Or what Amazon’s selection looks like now versus when it first launched.
Product teams need to be mentally strong to make such changes. The better your initial PMF is product-market fit, the more your early adopters will react to product changes. As products and services reach more users and adapt to more scenarios, it is inevitable that those distinctive features and preferences will be diluted.
Remember when Twitter decided to expand the character limit from 140 to 280, there was a huge backlash from leading users, who threatened to abandon the platform. Ironically, the fact that a larger word count actually improves the product experience for existing users doesn’t make any difference to those who don’t use Twitter at all.
Back to Snapchat. I wrote a long time ago that the power of social networks lies in social relationships. That means a lot, and in Snapchat’s case, it’s a double-edged sword. Snapchat is a social network embraced by young people. If there were a sudden influx of older people, the attraction would be lost, as Groucho Marx put it: “I wouldn’t want to join a club if it were all me.”
From an instrumental dimension, Facebook’s network effects remain pure and infinite. The more people on Facebook, the more instrumental it becomes, similar to the value of a global index. Some people don’t use Facebook very often, but if I don’t have their email or phone number, I can still find them on Messenger. Analyzing Facebook is complicated, of course, because it serves so many different needs in so many different geographies and markets, and social networks tend to be highly dependent on usage scenarios. In some places, for example, Facebook is synonymous with the Internet; In some countries, some companies only have a Facebook page, which is hard for Americans to understand. I, for one, have a habit of doing Yelp searches first.
On the social dimension, the picture is less clear. Here, I will focus on the US market for analysis, because this is the market I am most familiar with. Facebook is the largest social network in history, and it faces challenges to scale that no other social network has ever seen.
The power of social networks comes from social relationships, but it also creates problems. One is that there is a point at which having lots of social connections, which is undoubtedly a huge advantage, can become an impediment to continued growth. For humans, who have long been more used to living in less large groups, suddenly having to post to everyone we know on a social platform can be daunting. Except, of course, celebrities, marketers, and people who share all the time, regardless of the audience, and are immersed in their characters all the time. Of course, you know where you belong.
This is the first time in the history of social networks that we have experienced diseconomies of scale, and it happened to Facebook because it was the first to achieve this unprecedented scale. Imagine being in the same room with all your family, friends, colleagues, acquaintances, even people you’ve only met once, because you can’t turn down a friend request from anyone. That’s hundreds, maybe thousands of people. What can you say to them in this situation? We know that people have different identities with different people in life, but we rarely need to forge a unified identity with all the people we know. In the real world, there are probably only a few times in life when you need to do this, and one of those times is at your own wedding, and then many years later, at a funeral. And in the online world? This simultaneous display of identity for everyone is the default mode in Facebook’s news stream.
It is not surprising, then, that public figures who have to deal with this issue on a daily basis are often particularly defensive. The more people you meet, the more likely you are to accidentally offend someone, it’s 100%.
When I check my Facebook feed, I notice that fewer and fewer of my friends are still sharing anything about themselves on Facebook. My guess is that if you plotted the number of friends on Facebook and how often they shared it, it would look something like this:
Of course, this isn’t true of everyone. There are always some sharing fanatics who will share their opinions freely, regardless of venue or audience size, but these people can be downright annoying. Imagine being at Thanksgiving dinner when you’re about to carve the Turkey and someone suddenly starts talking politics… This leads to a vicious cycle — regular users share less and less, and annoying people flood the screen…
Under the algorithm, not every piece of your content will be sent to everyone, but you also don’t know who will see it, so for you, it’s still like everyone will see your content. It’s the same with the chilling effect.
Another diseconomy of scale occurred when parents flocked to Facebook, forcing young people to start moving to Snapchat. The quickest way to clear a club or dance floor is to have their mom and dad pop up. When your parents show up among your Facebook friends, you are constantly aware that they are watching you. It’s like trying to tell a secret in a room that might be bugged, but you can’t turn up the radio or turn on the faucet like in a spy movie. All you can do is communicate in some code that mom and dad don’t understand, or escape to Snapchat…
What I’m trying to say is that I think the most likely breakthrough for Facebook to regain value in social features is to break the existing interface. The current ad-friendly interface for news flow is at the heart of the diseconomies of scale.
One obvious solution is group mode. It can subdivide large social relationships into groups that are more unified in purpose or ideology. Google+ Circles is an attempt to explore this, but since they haven’t reached that scale yet, the problem they’re trying to solve hasn’t really happened yet.
Where is the Ceiling on Instagram? The problem is tricky because until now they have been growing at a high rate and there is no sign of a slowdown.
One of the advantages of Instagram is that it happened to be in the transition period when Facebook was expanding from text to photos and videos. Instagram was an image platform from the beginning. The initial format of Instagram was square pictures with simple titles, which could not be linked and did not support forwarding.
There are many advantages to this form. One is that it’s harder to be provocative or intolerant in the form of pictures than it is with words, because pictures naturally soften both bragging and provocation. People tend to be better at using words to do aggressive and hurtful things. And Instagram has gone further than other platforms in policing emotional behavior. Especially when compared to Twitter, which has only recently begun to regulate offensive comments.
Of course, photos are not completely immune from attack. Instagram is full of “Look how perfect my life is” photos, and this is a major complaint. It’s also true that some people are uncomfortable with the constant presence of such content in their news feed. But given that Instagram isn’t a completely public, one-way platform like Twitter, it’s harder to complain publicly. The only way to do that is with comments, and Instagram is clearly already doing that.
At a time when Facebook and other social networks are rushing in the direction of all-media, Instagram stands out as a pure visual medium. There’s a sizable group of users who think that the best part of Facebook is photos and videos, and that’s a natural segue to Instagram, which is a perfect fit for them. And, as we often say, and also mentioned in the Twitter section, we tend to overestimate the total market size and fit of text fit. If Facebook only showed photos and videos from a week, I think their activity would go up. But now that Facebook has acquired Instagram…
Instagram, like other social networks, is breaking through the format ceiling created by those early format limits: photos are no longer forced to be square, and videos are longer, widening the range of usage and, most importantly, lowering the bar for producing content.
Instagram copied Snapchat’s Stories feature, an important move to break through another giant ceiling. The inherent nature of the Instagram platform for Posting the best and the best content seriously affects how often users post content. Stories addresses the supply-side problem in several ways. First, rather than being pushed directly to the user in the homepage feed, the user has to click on the publisher’s personal page to see it, which puts the choice in the user’s hands and the publisher doesn’t feel pressured to cram too much content into someone else’s feed. The automatic deletion nature of Stories further alleviates publishers’ concerns about too much and too much clutter affecting tonality. Stories freed up content productivity at a stroke, resulting in so much content that I could barely read a fraction of it. Even my friends who aren’t very shareable often share content in Stories rather than in the main feed.
Instead of pandering to the landscape that video viewers are more used to, Stories went portrait full screen, because portrait is creator-friendly and it’s the most natural and convenient direction for us to shoot with our phones. Stories also offers personalized stickers, text tops, face filters and other gadgets that aren’t available on Instagram’s main interface, presumably in an effort to create a differentiated aesthetic in the feed and what’s presented in Stories.
Instagram is also a natural fit for advertising: anything on it is advertising. Some people complain about too many ads on Instagram, but if you check your feed, 100 percent of them can be called ads.
I opened up the news stream and only looked at the first 20 and classified them all as ads: Some how good recommend to eat, some recommended beautiful destinations, some recommended photographer, cameraman virtuosity, some recommended upcoming film actor, and rihanna’s latest underwear series and cosmetics, as well as those for friends cooking delicious food, is touring concert, interesting people, or some sports or teams. Yes, only some of them are official Instagram ads.
I don’t think it’s bad. All social networks have this appeal, but Instagram’s visual nature and tonality make advertising a perfect fit. For example, messaging apps themselves deliver messages rather than ads, which is why Messenger is unlikely to become a lucrative advertising platform like Instagram. On Instagram, if an AD isn’t clearly flagged, or if it’s not from an account I don’t follow, I can’t tell which ads are, because they’re no different than anything else in the feed.
Ironically, as Facebook continues to broaden its usage landscape and move into all-media, Instagram has more room to grow because of its simplicity and purity.
Of course, every product and service has its own natural limitations. Talking to people on Instagram, for example, still feels awkward and redundant. But consider how many young users use Snapchat as their primary chat tool, and in that sense, Instagram has plenty of room to imagine.
Instagram is rumored to be considering a separate messaging App, which I think makes sense. And it would be ironic for Snapchat if Instagram were able to separate its core messaging and broadcasting functions before Snapchat did. Snapchat has done a lot to balance the instant-messaging nature of the App with radio elements like Discover and Stories. A more efficient solution would be to split the two and have separate apps host different functions.
All social networks exist on mobile phones, so nothing you can design can jump out of that screen. So, there’s bound to be a limit to how much optimization you can do in a single App. The evolution of the phone itself has made possible new designs and billions of dollars of business, but Instagram, like other apps, faces a ceiling on how much you can cram into a single App.
Instagram has already moved in this direction, creating separate apps like Boomerang or Hyperlapse that focus on different scenes and needs. These separate apps can help you take advantage of the latest technology on your phone while keeping Instagram’s main App pure. I often hear people argue against developing new apps because it’s a hassle for users to get used to a new App. But in a sense, splitting different functions into different apps is the most elegant way to break through the limitations of mobile devices.
At the moment, Instagram is still dominated by short content, and long content is less numerous and not better supported. Instead of trying to optimize it into something like a TV show or live broadcast, the current model of news flow is for algorithms to show you a huge amount of content based on the accounts you follow. The closest thing to long content on Instagram is Live. But for now, most of them seem to be just longer selfies, loosely formatted and uncurated. In fact, compared with the pursuit of long narrative, a better way to aggregate existing content and prolong users’ use time is to evolve towards the direction of “sense of presence”. Similar to what Snapchat does with location-based distribution, aggregating content from a single location into a news stream. This is valuable for concerts, sporting events, or emergencies such as natural disasters or protest marches.
There is also a more general ceiling on how long we can become addicted to algorithmic recommendation based information flow patterns. Looking at everyone’s carefully planned snapshots day after day, we may one day reach a tipping point of jealousy and fear of missing out. When enough users reach that threshold, the ceiling will be reached.
However, in this age, we have reached an unprecedented level of tolerance for vanity, which previous generations would have considered disgraceful. Just as there is a glut of information, there is a glut of identity. In this world, all you have to do is leverage the power of social networks to maximize your individual impact. This drives the underlying structure of social networks to change from centralized to decentralized.
The final possible ceiling comes from a basic judgment I have made about huge social networks like Facebook and Instagram. I think for them to continue to grow, they need more structured interactions and unit-like content, for example, lists are a kind of entry point to a structured unit. This certainly raises the bar a bit for content producers, but the benefit is clear: it breaks up the overall flow of information into different units, giving users a sense of what to expect when viewing different categories of content, similar to the way magazines are arranged.
The information flow model has become perfect at filling people’s free time. A huge amount of information can be inserted into every gap in people’s schedule and attention. On the other hand, locking time is a higher order task than filling gaps, and structured content has an obvious advantage over information flow. People set aside a whole block of time to play Fortnite or watch Netflix, rather than a dedicated block of time to surf the news stream, unless they’re on a boring flight. Silicon Valley’s ambitions don’t stop at padding, and every tech company wants to be in control to some extent, either by itself or through acquisitions, like collecting all the Infinity Stones in the Marvel universe. Facebook’s acquisitions of Instagram and Whatsapp are the best examples.
Where is the amazon’s next ceiling?
We started our conversation about ceilings with Amazon, a case worth revisiting because amazon’s ambitions are far from over. After freight, where is the next possible ceiling?
Also in the case of delivery, the other natural barrier after freight is speed. Yes, free shipping is great, but over time, expectations will rise. As Jeff put it in his latest annual letter to shareholders:
One of my favorite things about users is that they can’t get enough. Users’ expectations are never set in stone, they are constantly rising. It’s human nature. Human beings have never been satisfied with farming and hunting. They are always striving for “better”. “Wow” can quickly turn into an ordinary “um,” and today the escalating cycle is faster than ever. That’s probably because today’s users have so much information at their fingertips: scrolling through reviews, comparing prices across platforms, checking inventory, measuring shipping speed, and so on. Just take the retail business as an example. Everything we do at Amazon, and every other business, points to the same conclusion — that user initiative is on the rise. Because of this, you can never rest on your past success, because your audience never stops.
Free shipping is great, but why wait two days for delivery? Is it possible to receive it tomorrow? Even today? Even right away?
Amazon has been trying to solve this problem for more than a decade. To this end, we built a higher density of small warehouses, rather than the previous model of fewer and farther large warehouses. When we first mentioned drone delivery on 60 Minutes, everyone thought it was a joke or a gimmick, but it was for the same reason that we had lockers in our local retail store, which was to solve the delivery speed problem.
Another possible ceiling is that What Amazon is good at is meeting customer needs. We can be the best in this dimension and be the first choice for users when they have clear needs, but we are not good at stimulating needs that users themselves do not have. The tech company that does this best is Apple, and if you look at other industries, luxury brands are the best.
At Amazon, the typical shopping pattern is what we call “spear fishing.” Users have specific shopping needs, they go to Amazon, type in what they want to buy in the search box, and select the one they want with one click. On the other hand, if you’ve ever been to a mall with someone who really enjoys shopping, the whole situation is very different. It’s a combination of broad convergence and target – you can’t have one without the other. For example, when buying clothes, they will pick out a few clothes from the display shelf, try them on, look in the mirror, spin around to observe the overall effect of the upper part of the clothes, and hold them in their hands repeatedly to make the final decision. Sometimes they don’t know what they want to buy, but they can’t help running their hands across the rows of clothes.
It’s a state similar to browsing, and Amazon’s interface has some services based on it, but not enough. If you have a general idea of what you want to buy, Amazon, based on algorithmic recommendations, can guide you through the comparison process and provide relevant user reviews for you to reference. But the process is still more purpose-oriented.
Amazon’s offline stores do the same. I’ve been to amazon’s offline bookstore in Seattle College Town. It’s a 3d version of Amazon’s online bookstore, except it doesn’t have as much stock as it does online. Unmanned Amazon Go sounded more fun, and I went with gusto, but it had the same look and feel that its main selling point was self-service shopping to enhance the shopping experience.
Speaking of creative desire, I think of my recent and only trip to Milan. I wandered into a luxury store, where the salesgirl managed to convince me to buy a sports coat I didn’t need. In fact, they had such a low stock that the coat was not on display at all.
The clerk just looked at me, asked me a few questions, then went behind the counter and pulled out a coat. Notice, instead of giving the customer a lot of options, just give them one. She first convinced me to try it on, then complimented me on how good it looked and introduced me to some of its special features. Slowly, SILENTLY, I began to nod my head in my head, and finally I fully agreed that the coat belonged to me, and that I should be wearing it the way I should.
This is not amazon’s fault alone. Most tech companies have been used to machine learning and algorithm-driven ROI for years: getting more data, improving recommendation algorithms, making better human-goods matches, and so on. The beauty of luxury is that they are not data-driven at all, which is to convince you that this is what you want. The same is true of Hollywood, which is admirable. Seduction is a gift, and technology companies tend to overestimate the power of algorithms, assuming that happiness is a problem that can be solved by algorithms, and underestimate the value of seduction.
Netflix offered $1 million to improve its recommendation algorithm, but millions of users scrolled back and forth on the Nextflix home page every day to find what they wanted to watch. This isn’t just Netflix’s problem; it’s the same with any streaming platform. Apple TV, as a video-watching device, is most appreciated by users for its screen saver. It’s like a restaurant that’s not known for the food, but for the typography on the menu. I’m not saying that data drives don’t work, but tech companies need to realize that the returns on those enduring temptations can be better than you think.
But that vulnerability is less fatal for Amazon than for other retailers. The vast majority of goods sold on Amazon are generic, and users may have been groped on other platforms, but they often end up coming back to Amazon to complete the purchase. Amazon’s strength in logistics and customer service is the biggest barrier, because logistics and service are the closest part of retail to payment, which is why Amazon earns nearly all the profits in the retail chain.
As Jeff emphasized from the start in his annual shareholder letter, Amazon’s mission is to be the most customer-centric company in the world. One of the most effective ways to find drivers of sustained growth is to keep up with users’ dissatisfaction, which is almost natural, and in the right circumstances, users are happy to voice their dissatisfaction.
How to find your own invisible ceiling?
One way is to ask your users directly. As I mentioned at the beginning of this article, we found the shipping breakthrough by directly asking users, both those who use our platform and those who don’t.
Mr Ford had an oft-quoted quote: “If you asked people what they wanted, they would say they wanted a faster horse”. That seems to contradict my point.
Indeed, consumers often can’t say what they want. But an oft-overlooked point is that they can at least express what they don’t want. When users say they want a faster horse, what you should hear is not literally, but that they feel that riding a horse is too slow, and clever product people can take that point further — people want a faster way to get around.
Of course, this is not a panacea. Because sometimes people lie and hide things they don’t like, and sometimes they can’t even articulate their dissatisfaction. But if you combine their feedback with good user behavior data analysis, or well-designed user testing, you’re more likely to find the real problem.
A popular view in Silicon Valley is that ToC is more of a challenge on the product side than ToB, because enterprise customers can articulate what they want, while consumer users can’t say what they want, as Mr Ford says. I must reiterate that this statement is only half true. I have consulted with ToC companies and found that they do not put enough effort into researching objections to their decision not to use their products.
We talk a lot about the demand curve in economics, but there’s also a demand curve in product, which charts what users want for your product or service. As Jeff stated in his shareholder letter, the arc of customer demand is long and upward. Every company, especially its product team, needs to keep abreast of this demand arc.
A lot of companies focus on analyzing the user conversion funnel to see who gets converted and how. However, as the company grows, it becomes more important to focus on users who have never entered the funnel, or who leave as soon as they enter. If your product-market fit is different for each of your current and potential customers, it’s a never-ending task for the company to understand why and where the differences are.
When companies test their products in focus groups, they often show me positive feedback. And I’m always more interested in people who give negative feedback. Of course, the discomfort caused by this negative feedback to the product team is self-evident. Maybe some of the negative feedback isn’t useful right now; It’s also possible that your PMF is so strong in another user group that some negative comments have no real value. But I would argue that these negative reviews are valuable in the long run.
All honest negative feedback hints at the ceiling of your product in some demographic, even if that ceiling is far from there. I’ve met a lot of companies and even if they don’t know how to solve the problem, I’m still curious if you can at least articulate what your problem is.
I’m often asked to give feedback on products. It has often been the case that as soon as I raise a concern or question, the product team will jump in and interrupt me with an explanation to prove that my objection was anticipated.
My advice is to listen to the user’s feedback in its entirety, rather than interrupting it. You will never have the opportunity to explain your product or service to every user. The product will do the talking. Most product people blame the user, constantly trying to tell the user what they did wrong, which is not a good habit. They won’t give you any feedback in the future.
Even without external feedback, you can train yourself to find product ceilings. I used to use a method that worked. When companies launch new products and are exploring their PMF, or product-market matching, I ask them the question, of all the people in the world who don’t use your product or service, what are the reasons for each of them? If you ask yourself this question, you will get a lot of very clear answers, and if you continue along this line of questioning, the boundaries of your actual accessible market will become clearer and clearer.
Finally, it is possible to keep a business from going to the ceiling by simply relying on superior product intuition. I don’t really want to share that. When I meet Silicon Valley ceos who idolize Jobs, I object to their product intuition for exactly the same reason they promote it — product genius like Jobs is so rare that it’s as precious as a unicorn. So relying entirely on product genius is much riskier than relying on a complete methodology.
Of course, that’s an argument that product people don’t listen to. The enduring image of Jobs as a technophile who kept coming up with brilliant ideas was heroic and romantic. A look at Jobs’s biography, however, reveals how rare life experiences and choices shaped him. Still, it has spawned a crop of ceos wearing The same Kind of Jobs pullover and equally keen to pick apart design teams. It was as if you could learn the essence of Joe by imitating his appearance. We greatly underestimate the role of Jobs’ unique product genius.
Mr Jobs’s gifts are too rare for Apple itself to replace him. It’s no coincidence that the apple products that have disappointed me most have the word “Pro” in their names. MacBook Pro, Mac Pro, Final Cut Pro, iMac Pro, they’re all full of flaws.
The users of these Pro products are usually the ones with the clearest needs and workflow. Therefore, listening to them is a very reliable way to establish and maintain a product-market match. But Apple doesn’t seem willing to do that, so you can imagine the mistakes they’ve made in those product lines.
I’m glad to see that the next Mac Pro is being widely heard from the media and professionals. It’s disappointing that it took 2019 to do this, but at least Apple is finally getting off its pedestal. That’s one of the best things I’ve ever heard about Apple.
Success is intuitive, failure is intuitive. Snapchat is another company whose existence depends on one person’s product inspiration, and in that light, it’s no surprise that it stumbled with its latest redesign. Companies are always true to the methodology on which their success depends, and as a result, their greatest strength is often also their greatest weakness. Apple and Snapchat both worked wonders with one man’s product inspiration: got us rid of the CD-ROM drive, invented the iPhone, Airpods, camera-focused apps, Stories, and on and on.
Our products and services are facing an increasingly dynamic external world, especially ToC products and services, which can no longer apply the same old truth rules. Most product people’s perception of the outside world often lags behind the pace of change in the world itself, so I recommend incorporating data and user feedback into product intuition. The success of product people often comes from being right about a group of users at a given moment. So called PMFS are post-hoc. Analyzing their success leads us to confuse luck with a truly sustainable product intuition, which, again, is extremely rare.
Just like an investor who can be regarded as an investment genius for a lifetime because of his luck in making a few very successful projects; Or an impressive multiple of returns often creates a lasting brand effect for the investment institution. In fact, a product person’s intuition may have coincided with a moment in history to create a hit product. But one’s perception cannot always keep pace with the rapid changes in the world.
This explains why sustaining competitive advantage is so difficult. In the long run, internal product changes tend to go in a negative direction. But we may be too internal and not pay enough attention to external factors. Ben Thompson has a wonderful analysis of this in Sacred Discontent: Subversive Weapons:
Bezo’s shareholder letter revealed another big advantage of focusing on the consumer: there is no ceiling. When I wrote that article five years ago, I used to think that there was a ceiling to the business opportunities that came from focusing on users, and that you could only get closer and closer, but not fully satisfied, as this chart shows:
But in fact, we’re missing an important point. Is that users’ expectations are not static. As Bezos describes it: sacred discontent. What’s extraordinary today may be extraordinary tomorrow, but it can be a huge opportunity: if your company is focused on creating the best possible experience for its users, that means there’s no ceiling on your growth. As shown below:
In Amazon’s case, their user-centric mission means ambitious and dynamic growth, plus they’re particularly good at constantly incubating new businesses with profits from existing ones. Like a perpetual motion machine! I’m not sure Amazon can beat Apple’s trillion-dollar market cap, but they’re definitely one of the top two.
The so-called pattern inductive approach has become almost standard in Silicon Valley and beyond, but it has an after-the-fact quality to it. We can’t blame most people for sticking to a steppy-as-you-go approach, which is the least risky.
But if the consumer’s is changing, then sooner or later this pattern of inductive path to the ceiling. Why is it that startups tend to create some kind of technological innovation? Because they can’t live by following the market to the average level. Achieving absolute product-market matching is a very real problem for startups, and there is no turning back.
Companies can stay alive by constantly changing product owners, but for product professionals, our self-requirement is to always be aware of the market, even if your own life doesn’t change that much. I’ve found that an effective way to free myself from my ingrained product obsession is to absorb more new things. Also, the older I get, the more COMFORTABLE I am with dancing with my users. The lead dancer may seem more in control, but these days, it’s not the solo era.
One of my favorite posts by Ben Thompson is “What Clayton Christensen Did Wrong”, where he builds on Christensen’s theory of disruptive innovation and suggests that the way to avoid low-end disruption is to improve the user experience. To some extent, the user experience can never be improved enough. When Tesla came along, electric cars had a shorter range than petrol cars, charged more slowly (given the scarcity of public charging points at the time) and were more expensive than petrol cars — hardly a low-end disruption.
But Tesla has clearly found its own special footing. For example, it has a better range than other electric cars and can be charged twice a day for commuting, but more importantly, the Tesla is a symbol of environmental awareness for the wealthy. Of course, Tesla can’t rely on this gimmick for long, because key metrics of user experience change. But the biggest takeaway from this is that the breadth of user experience is broad enough to be possible, and some of it is hard to quantify.
Thompson believes that we can never overserve our users. As a product person, I would say that it’s actually very difficult, if not impossible, to have a very deep understanding of customer needs. Amazon has made it its mission to be the most user-centric company in the world, and that’s a truly long-term strategy because it’s truly endless!
In my personal experience, the most successful people tend to recognize their ceiling before others do. They realized it early and avoided it decisively. Some people decide as early as elementary school that they’re not going to be a world-class pianist or tennis player. Another mentioned realizing in his freshman year that he wasn’t going to be the best at math in his dorm room, let alone in the world. Another guy worked for a year and knew he wasn’t going to be the best programmer in the company, so he switched to management and eventually became CEO.
They recognize their limitations early on, and as a result, they are quicker to figure out where they can grow. The most frustrating thing about product development is that it’s a multi-dimensional complex problem, but the advantage is that it’s not expensive to give up some of these dimensions, so this idea can be applied broadly.
This is not about advocating that everyone be the best at what they do. I’ve made peace with my mediocrity in life: accepting that cooking for myself is average, that I can’t win the Tour de France, that I’m better off in the background than in the front… But when we enter the realm of business, where the winner takes all, you have to do business if you want to survive the increasingly cutthroat competition. Sticking to yourself and doing what you want won’t get you far or better here…
GSR Venture Capital
Focus on Internet early stage investment
This public number content does not reprint without permission, but welcome to forward circle of friends.
