Hacker News new | past | comments | ask | show | jobs | submit login
The Facebook Offering: How It Compares (nytimes.com)
296 points by jashkenas on May 17, 2012 | hide | past | favorite | 83 comments

For the curious, this started out as a D3/SVG chart that we ended up porting over to be 90% canvas, just for performance reasons (especially in Firefox). There's still lots of nice ugly "if (USE_CANVAS) {" stuff in there.

It's excellent work. Good idea and execution - in particular, I like how you shift from an absolute scale to a logarithmic scale. I think the animation should help people who don't normally have an intuition for logarithmic scales.

The most fun I'm having is watching "0" fly off to infinity when you switch from page 1 to page 4 or 5.

I guess negative infinity is pretty close to positive infinity in that toolkit.

Awesome chart. NYTimes always does such a great job with charts and graphical data representation. Many would do well to learn from NYT, instead of chartjunk with clipart and few data points.

I was going to come in here and ask just that question. Now a followup--what was wrong with d3? was it just too slow to render? Or too slow to load?

When the bubbles were done as SVG elements, the animation in Firefox became _painfully_ slow. Felt like just a couple frames per second, maybe. (In contrast, the animation was fine in WebKit-based browsers).

Is the original available? I'd like to look into this. Thanks!

I imagine it's the performance hit of moving around so many DOM elements on the page. We had a similar experience with this visualization that started as D3 http://trends.truliablog.com/vis/greener/

Are those real or nominal dollars?

It's in 2011 inflation adjusted nominal dollars.

Er, 'inflation-adjusted' and 'nominal' are opposites, so that answer leaves me confused.

Looking at one of the prominent early data points, AAPL:

Some online sources peg its IPO valuation at about $1.3B at the time (ie, nominal 1980 dollars), rising to about $1.7B in the first day pop. According to WolframAlpha, [$1.3 billion 1980 dollars in 2011 dollars] is about $3.5B adjusted by the CPI - close to the NYTimes $3.4B chart number.

From that I conclude that the numbers have been adjusted to "2011 (real) dollars".

Sorry, brain fart. Yes, it's adjusted to 2011 (real) dollars.

It's just crazy how slow Firefox gets at the simplest samples from the D3 site.

Just try the "Evolution of Wealth & Health of Nations" in Chrome AND Firefox http://bost.ocks.org/mike/nations/

Is there any open source version of your port?

I'm evaluating D3 and other libs for charting. If you could provide more details on specific things that were slow, resource heavy or buggy, that would be very useful.

Fantastic job. I've seen quite a few impressive interactive charts over at the nytimes in the past year or so.

This [1] is my hideous destruction of the graphic (sorry) to overlay world population on the Internet with respect to time, data from some worldstats site[2]. I propose that the proper size of a tech IPO should be normalized (and upper bounded) against the possible global reach, which I think the % of world population on the internet does a good job of showing.

[1] http://i.imgur.com/HyeKf.gif [2] http://www.internetworldstats.com/emarketing.htm

I wish I could think of a reasonable way to normalize for the age of the company. A decade ago, there was no SOX and it didn't suck as much to be a public company, and it wasn't as easy (?) to raise $100MM from DST as an alternative. Facebook delayed its IPO for as long as possible and went to the private capital markets instead of to the public. Its valuation should be compared with companies at a similar point in their lifecycle.

Tech isn't only the internet. Microsoft and Apple both went public before the Internet and managed to do quite well without it.

Correlation != causation. Might as well graph it against the number of World Cup or Euro championships in that year and get a much higher correlation since the IPOs seem to all be in an even year.

It'd be much more interesting to look at perfomance vs sector and performance vs the index.

> Microsoft and Apple both went public before the Internet and managed to do quite well without it.

That's because Apple's and Microsoft's products don't (or didn't at the time the companies went public) rely solely on the internet. A big reason why Facebook has such a massive valuation is precisely because the population of the internet has increased so strongly.

Have you taken into account Facebook being the reason a lot of people joined the Internet?

I would find that a hard claim to believe.

I find it plausible. For anything more I'd want evidence, but all I can offer is anecdote. In the UK over the last few years most Android/not-iPhone adverts I saw made it very clear that the phone did Facebook - implying to me that Facebook was a major driver for people buying a smartphone.

You get my upvote for 'Correlation != causation'. I say that probably twice a week myself - i'm still looking for a way to solve for causation, when I figure it out, i'm heading to Vegas.

That is an interesting thought but China complicates things since a good chunck of new internet users are Chinese but Google and Facebook are blocked there.

You need to keep in mind advertising $$ for those internet users as well. As someone mentioned, China is a huge factor and Facebook isn't available. And the vast majority of advertising revenue a company like Facebook would make most certainly comes from the US and Europe, which is pretty saturated with internet users already. South America, Africa, and Asia are all going to be much harder to monetize on.

This is an excellent insight. One really has to consider the total available market when valuing a business. The simple and intuitive example is if there is only one dog food store in Duluth Minnesota, population circa 87,000 it might have a value of $X, but if there is only one dogfood store in New York City, population circa 8.25 million then you would have to get the store in NY a higher valuation than the one in Duluth, based simply on the number of potential customers.

That is one of the big differences between the dot.com bubble and today's valuations. In the 90's there were tons of people who only used their computers at work and didn't have one at home. That has flipped where lots have them at home and a large fraction have them in their pocket. So the valuation of a company that requires network access to be successful has to be proportionate to the network population.

100% agree, it should also be normalized against the value of dollar.

Interesting. The largest IPO in the original dotcom boom, above everything else on the chart except Google and Facebook, is a company I don't recall ever hearing of: Corvis Corporation.

They started in 1997 as "Nova Technologies", changed to "Corvis" in 1998, did an IPO in 2000, and promptly took a nosedive. In 2004 they became "Broadwing Corporation" (which I also don't recall ever hearing of), and in 2007 they were acquired by Level 3 Communications. It's amazing how little remains of them on the web. Except for some terse stock info, from which I got the above, all I can find are a couple of articles from 2000 talking about how everyone loves them.

The dot-com bubble was almost like a blister compared to the much larger telecom bubble.

Startups like Procket and Caspian Networks raised over $350M each, at multi-billion dollar valuations. In the public markets, Cisco's valuation reached $500B, now-bankrupt Nortel was $300B+, etc.

I found this interesting too. After doing some digging of my own I found this:

Corvis Corporation is the first company to make the intelligent all-optical network a reality. Our solutions enable telecommunications service providers to construct manageable all-optical networks that will accommodate the continuing growth of Internet, video, voice, and other data traffic

Source: http://web.archive.org/web/20011103200605/http://www.corvis....

First day value change seems misleading to me. I think it represents the value at which select shareholder can buy the initial shares issued by the company and not the value which they are first traded at the stock market.

Consider for example linkedin, which got "issued" at around 35$ per share. The first price to hit the stock market was around 70$. This is because the first shares are sold to a select few at the 35$ price so there can be volume at opening. Those people traded the shares at 70 right away.

We can conclude at least two things from this : 1-Banks will try to scam companies during IPO by issuing shares lower than market value so that early buyers can pocket instant profit 2-If you are not in the selected few, making big bucks during an IPO is much more difficult than what is represented in this graph

You are correct. The relevant citation is this. http://blogs.reuters.com/felix-salmon/2011/06/07/the-us-ipo-...

On average, banks take about 7% of IPO proceeds due to a Wall Street cartel of underwriters.

Its fun to see majority of IPOs in 80s were proper Engineering companies. I was just talking to someone who is a Civil Engineer (in his 50s) and he mentioned the same. He did comment that people in his generation built proper engineering companies. I don't understand why do people think Internet startups are trivial. One way we can say Internet has been game changer - Groupon brought easier advertising mechanism for local SMBs. LinkedIn connected professionals across the world. FB has created a virtual interconnected world. Of course there are duds like Zynga whose biggest achievement so far is pasting cows on the Internet. And multitude of other companies like AirBnB who are disrupting entirely new industries. These companies are definitely gamechangers.

However, I've seen lot of my colleagues from fields like Mechancal Engineering or BioMedical Engineering jumped on SoLoMo wagon and mostly working on some awful apps. I think Internet has diverted attention from real engineering to some "shitty" startups/solutions that are in search of problem, thanks to dot com bubble in 90s.

To be honest if I could go back and talk to myself ~2000 and I said that the most exciting , hottest Internet company in 2012 is a website that allows users to post text based messages to other users of the same website I would have been somewhat disappointed.

That's really cool. What do the size of the circles represent? Is it the same value as shown on the Y axis or something else?

And what do the colors of the circles mean?

The size of the circles is also the I.P.O. value -- the same value shown on the Y axis.

In the first few steps, it's not terribly useful. But once you switch to the logarithmic scale, it's nice to keep a reference to the true (read, "linear") relative value.

I feel like Page 5 doesn't really represent the information it's trying to convey, which is how well these companies did 3 years later.

Yeah, that information is really represented in the transition between the pages. So I just ended up clicking back and forth between 4 and 5 to get a feel for those kinds of patterns (like an animated weather map).

yes, i think it would have been better if the y scale was ratio of current to ipo valuation.

Actually that's misleading. If the radius of the circles go up linearly, the area/size of the circle goes up exponentially. So on the logarithmic scale, the height changes with the log of the value but the size goes up exponentially!

Edit: nevermind this, it's wrong.

Nope. As @arscan says, in this as in most bubble charts, the value determines the circle's area, not the radius. Otherwise it certainly would be misleading.

It looks like they scale it based on area, not radius... just by eyeballing it. Polynomial (not exponential) growth btw.

It looks like the size of the circle is the same as the value on the Y axis. The advantage of having higher values have bigger circles is that you can make all the crowded companies at the bottom of the Y axis can have smaller circles.

Title was: "Graphic: Compare all the Tech I.P.O.'s Since 1980 to Facebook's"

That seems strictly superior to me. Who's the editor/bot changing every submission titles back to the article title even when it removes (sometimes vital) context?

It wasn't me. I didn't realize that HN had the ability to swap 'em out on you.

Yeah, they'll change 'em sometimes if the title is too linkbaity or editorializing, usually to the title of the original article or blog post. Recently I've seen a few titles being changed to the source title even when the source title is completely unhelpful, leading me to wonder if someone hasn't tried to automate the process a little too effectively.

Bubbles don't just happen overnight. It's a slow loosening of caution as people convince themselves that a boom is unstoppable. I think people are still shell shocked from the '00s. A few more large and successful IPOs could change that. We won't see the next one coming until the current boom turns into a bubble. By then we'll all be too happy with how things are going to hit the brakes before we fall off the cliff.

I think it's different this time. People are aware that a bubble is an eventuality, so this time the game is knowing when to jump ship.

Sign #52 that you are in a bubble, people start using the phrase 'It's different this time'.


I'm about to set the range to various booms and busts to see what people were saying.

All generations had plenty of bubbles. Maybe you were too young to have seen a previous one last time around, but certainly people in charge weren't.

People consumed by a bubble mania find reasons to justify continuing even when an objective observer would say they need to stop. I don't know what would be different this time. Plenty of people saw the housing and tech crashes coming, but they were shouted down by the powerful hype machine that precedes every bust.

As an example, Canada's in a housing bubble right now, but when I tell my friends and family they should consider selling now and renting for a couple years they look at me like I'm wearing tinfoil.

The circles should be colored to indicate performance.

I definitely feel as though the color information is being wasted; the placement of the circles along the horizontal axes already conveys "period of time" pretty well.

I think this would be even more interesting if the "Value 3 years later" number was incorporated into the graphic somehow.

See page 5, especially the transition between 4 and 5.

Yeah, the real value of this chart from my perspective is the transition from slides 3-4 (watching things go up the first day) and then slides 4-5 (watching some stuff bubble up, and others fall off a cliff). So I found myself just cycling back and forth through slides 3-5 to get a feel for the data. The obvious cool thing to watch is the early-2000's dotcom waterfall.

Nice job. Could probably have represented things slightly better given what you are trying to get across (using colors & ball sizes), but cool nonetheless.

I actually think the transition from 3 to 5 (directly) is more interesting than the one from 4 to 5. Well, they're both interesting, so it's cool the visualization supports both.

Was about to say the same! It was fun. My impression is that on average, the same % of companies fall in 80s, 90s and 00. (of course during bubble there was much more companies on the market). Some to think that Yahoo made 3,500% three years later, not too shabby!

What I'd really like to see would be trails going off into the future -- rather than just a +/-% for the +3yr mark, why not just show the whole story for each company (perhaps only when you hover over it to avoid clutter). This would make the current color-coding much more useful -- each trail would be the color of the company's IPO year.

Witness the trends... and tell me the bubble isn't about to burst.

Facebook's IPO will be bigger because they put off their IPO as long as possible. In fact they got special dispensation from the SEC to delay the IPO even longer. So comparing FB in 2012 to Google in 2004 is not really useful.

Google IPO'd six years after they were founded, Facebook waited 8 years. The comparison is not that unwarranted I think.

The fact that they are bubbles makes that an obvious comparison. But from the IPO perspective, things aren't nearly as crazy as they were in the ~1999 timeframe. The sheer number of IPO's back then was nuts, and most of them ended up falling off a cliff. Sure, there are a couple of bigs ones now, but it doesn't seem like mass delusion as then.

What are you looking at to draw that conclusion?

Glad to see Google blowing everyone out of the water.

Actually Amazon and Yahoo and others have much more impressive "after 3 years" returns

Those are both terrible metrics. They're sensitive to IPO timing, for a start. Facebook is going public very (!) late, it's already achieved market dominance and very significant revenue. Contrast Microsoft, which went up before Windows 3.0, when they were mostly a software supplier to IBM (their other offerings were distant second place products to market leaders like Wordperfect or Lotus).

And "after 3 years" misses a lot of detail too. Apple looks pretty good in that metric. But their "after 12 years" numbers are a disaster. And of course "after 30 years" looks to be pretty fantastic for them.

I'd just like to point out that you are chastising them for going public after obtaining reliable revenues for several years. This used to be the minimum requirement for going public.

Who's chastising anyone? All I said was that they were "late", which is undeniably true. And I tend to agree with you ethically, though that "used to be" leans heavily on "used". Since the PC revolution in the 80's, almost no tech companies have gone public with multi-year profitability.

Note that, for better or for worse, relative performance after three years has a lot to do with how the tech market overall happened to be performing at that time.

For example, the companies that I.P.O.'d in 1996 and 1997 tend to do well using this particular timeframe, and the ones that I.P.O.'d in 2006 would be affected negatively.

Yup, it's just one metric, and "after 3 years" is pretty arbitrary.

Still interesting if you keep this in mind.

But I'd argue the most relevant time (for comparison of how well these companies built stable long-term value) is now, and it would be interesting to see numbers for % growth from IPO to now, or better yet, that measured per year (CAGR).

I do appreciate the bubbly themed chart.

I think this IPO will signal the beginning of the latest tech bubble reaching maximum capacity. I can't imagine how a valuation as big as this, held against a revenue stream that's nowhere near as tangible, could possibly create meaningful growth for shareholders. Unless Facebook decides to find more ingenious ways to sell the data we provide.

It will fertilize the valley and help more tangible start-ups spring up. If all those new millionaires don't move to foreign countries to avoid paying taxes, that is.

I notice bands in time with these IPOs. Very interesting. We're in the middle of another band it seems like.

You might be noticing the lack of IPOs following the 2001 bubble burst as well as the 2007-8 financial crisis. Since these are unrelated, these "bands" might not be a general trend.

Between slide 4 and 5 there's one tiny bubble that blows up into a big giant bubble and a small one on the left that sinks a little. They are Yahoo, up 3,590%, and Apple, down 25%. They should do a follow up to see how they are doing today.

Nice. Suggestion for next time is to have the names of the other large circles when you tap on them.

Works great on an iPad.

I was surprised it works great on my 3.2 android tablet with firefox as well. Everything works perfectly, great job.

It would be interesting to have a "where are they now?" slide to see how many of those companies are still around and still worth anything. And on a log scale apple, google and microsoft shouldn't blow everyone else too far out of the water.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact