Cash-Burning Facebook Borrows $100mm...Because It Couldn't Sell Stock? | Hacker News Reader

Cash-Burning Facebook Borrows $100mm...Because It Couldn't Sell Stock?

alleyinsider.com

22 pointsrockstar928 comments

Facebook is unique because, unlike other 'giant' silicon valley startups, they don't really offer anything new in terms of revenue. They are a sophisticated billboard. Google, on the other hand, built something new - targeting ads by keyword/content analysis.

Here is a thought experiment: If all of facebook's users came to your website tomorrow, chances are you could monetize them just as well as facebook. However, if you took all of google's users, you would not do as well because you lack their technology.

Facebook is like any other site on the net, just with more pageviews.

not any other site. Google can target better than anyone because they know what you are looking for and can provide advertisements for that thing. If I'm trying to find a track bike, and if I search google for track bikes, I get ads for places that sell track bike.

However, If I have in my profile that I ride bikes, bike companies can target me very specifically. Or if I say what bands I like and have listed where I live or go to school, advertisements for concerts for those bands (or bands I might like) can come up. Obviously a lot of sites online can do similar target style advertising, but IMO not as targeted as what facebook can pull of. And they can go horizontally with the company types that advertise, as profiles on facebook are very general. I can put a lot of my info up - for facebook, the more the better. Most sites don't have this broad ability. They are much more niche oriented.

>However, If I have in my profile that I ride bikes, bike companies can target me very specifically.

Sure, but does you having that item in your profile tell me, as an advertiser, anything about your intent to buy a bike? Buy it today? Are you buying it in the next two weeks while I have the money to run a campaign? The answer is of course a big no. Just knowing something about someone doesn't tell you about their current intent. Now, would showing someone like that bike ads over the course of a year help him decide on the make and model of the next bike he purchases? Probably it would. But that's a very expensive proposition and not very profitable. This is why Facebook has such low CPMs.

Now, suppose I had a way of saying "I'm looking to buy a bike"... a person like that would be much more valuable to an advertiser than someone who already rides bikes and has preferences set already for the type of bike/model that she likes. A customer that is new to bikes is A LOT easier to persuade to try out your brand than someone who's been riding bikes for years.

In the end, Google's intent-based advertising is so much more amazing since it caters to everyone's whims so perfectly and displays things to you that you really might want to buy and buy it now.

A lot of advertising isn't targeted towards getting you to buy things now, it's so that when you do go out and buy things, you'll recognize their product first and prefer it over the competitors that are also sitting on the shelves.

As a kid, I always wondered why companies bothered to run TV commercials. After all, you aren't likely to jump off the couch and buy a car this instant, no matter how much Ernie Boch tells you to. But a friend's mom explained that they don't care if you buy now: they advertise so that in a year or two, when you find yourself in need of a car, the first place that comes to mind will be Boch Toyota.

It's amazing how many people don't do any research on consumer purchases and instead rely on a simple recognition heuristic. Even people who should know better. The reason my startup's on GoDaddy is because when my cofounders bought the domain & rented the server (before I came on board), they just went with whatever they'd heard of.

Yep, I think those are called branding exercises. Only big companies engage in those types of advertising since they are not easily gauged for effectiveness. It's very hard to convince some smaller company that they should spend $5M on some blimp and slap their logo on the side or something crazy like that. Big companies (car companies spring to mind) have that kind of money and advertising budgets and they do run these kinds of branding campaigns all the time.

But the issues with these campaigns is that they're the first to go when recession hits. In times of recession, first thing you cut is branding advertising since you can't measure it and you can't track it.

Online banners, and especially google adwords, will fare better in slowdowns since a company must advertise if it is to survive and get some revenue in. They will probably pick something that's effective and something that they can track right away.

That logic may be true about bikes.

But what if my Facebook profile says I like the movies Godzilla, Jurassic Park, and The Bourne Identity? It's very easy to extrapolate a common movie "typeness" that someone likes from a list. Then you can target new movies in that genre very specifically to users who would be most likely to watch them.

For movies, books, tv shows, etc, and repeatable purchases, the cache of information at Facebook is far, far more valuable than two words at Google.

"For movies, books, tv shows, etc, and repeatable purchases, the cache of information at Facebook is far, far more valuable than two words at Google."

That data is valuable, but not more valuable than Google's data. It's not just two words. Google knows every movie you've searched for since the last time you cleared your cookies. Assuming Google links the cookie from AdSense with your general Google cookie, they know what sites you visit. They know tons of things about you, and that data is similar in value in isolation to Facebook's data. When you combine that data with intent, which you know from a given search query, you can display more valuable ads than Facebook could ever hope to.

If Facebook Chat gets traction and Facebook pushes its messaging as more of an email replacement, they'll have useful data they can use to determine intent, which they can use to better target ads and get more money per click.

Speaking of FB Chat, I used it for the first time yesterday when somebody messaged me, and it freaked me out. I don't know how many people enjoy having to chat with a browser window open. Sure, it's portable, but it's unappealing in many cases; and you're practically guaranteed that they're analyzing your chat logs.

I am pretty sure Google doesn't use your back search history in advertising targeting. They've said this many times.

If they did, here's my past few searches: jordan ritter, glowing resin bar, mac hard drive use visualization, ironport, joost commercial, photoshop light bar, wiki futurama, wiki lost, xserve g4 wiki

What does that list tell you about me or my buying habits? Now expand that list by a few thousand random search terms. Any more detail?

Those aren't "random" search terms. They tell quite a bit about the types of products you would be interested in, in this case hard drives and computers. Plus, data in aggregate is far more useful than data in isolation. Google knows what ads people with a similar search history have actually clicked on, and they can use that information to target ads at you. Google can rebuild most of the useful data from your Facebook profile (what kinds of books, movies, activities, etc. that you're interested in), and they also know what you're interested in right this moment. That is very valuable.

I didn't know that they don't actually use historical search data, but they have the data and it can be used in the future. I don't see how using past searches to target ads is so different from using current searches.

DoubleClick is now part of Google and DoubleClick has a terrible record with promises and privacy. In the 1990s, DoubleClick said that they wouldn't associate off-line profiles with on-line profiles. They lied.

DoubleClick and Google were very prolific with webbugs before their merger. (The latter is in the form of Google Analytics.) They're probably the most prolific now that they're combined. And they wouldn't do it unless they thought that the data was valuable.

All I know is that facebooks ad's are obnoxious enough that it was the catalyst for me finally installing adblocker plus on firefox.

I don't want to see ads for meeting attractive libidous neighborhood women on the sidebar of the website I use to keep in touch with friends and organize social gatherings with those same folks. It makes me feel like I'm using myspace or some porn site whose ads are specialized to the geographic region indicated by your connection....

use email then

the funny thing is, I'd much prefer that, but there are many people that for whatever reasons, are comfortable only using one particular variant of electronic communications, and if I want to reach them I have to accommodate that.

I have friends who aren't comfortable using phones (the lack of body language stresses them out), friends can't deal with textual communications because of the lack of explicit emotional tone leaves them confused, and friends who only check their email when they're expecting an email. And those are just some simple cases. In the context of that, using any single digital communication mechanism only allows me to communicate with a subset of the people I care about.

Anyways, my point was that the ads on facebook can be obnoxious enough that it spurs someone to install adblocking software

Google flailed around for a long time before finding a route to profitability. Before adwords, they didn't offer anything new in terms of revenue, either.

I'm not saying that it's a given that facebook will find a profitable business model...but with their user base, it's a much better bet than for most other sites.

(That said, $15Bn does seem a bit nuts.)

I don't see the point of comparing Facebook to Google. Even if Facebook's advertising is less targeted than Google's, so what? Facebook's revenues = pageviews / 1000 * cpm. Nothing about the effectiveness of Adsense says Facebook can't be a thriving business.

But the real reason you don't often see emerging private companies take on big debt-loads is that borrowing money is riskier than selling stock, and it also subordinates the existing equity holders.

I can't get my head around this sentence. I remember Henry Ford bought out all his original investors (when Ford Motor was "emerging") because he didn't want to lose control. (And the family still hasn't). Creditors only have power when you're in bankrupcy court. Shareholders have power until judgement day.

EDIT: Holy beans! When you buy one dollar of Ford Motor stock, you are buying 10 dollars of Ford Motor debt:

http://finance.yahoo.com/q/ks?s=F

  Market Cap (intraday)5:	16.20B
  Total Debt (mrq):	168.53B

Debt gets paid before equity. Meaning if they don't sell out or aren't able to pay off all their debt. The shareholders get screwed. That's it.

The title is a bit aggressive. It's likely that Facebook it couldn't sell stock at last fall's $15 billion valuation. Rather than do a "down round" of financing -- a round with a lower valuation -- they likely found it easier or cheaper to use debt.

Right. Debt financing is a much more common strategy for start-ups now. Sometimes the difference between a huge round of funding and a lesser one is just timing, and debt financing is a good way to give a buffer to time things perfectly.

In Facebook's case, who knows. It does suck to burn equity to buy servers. If Facebook is going to persist and continue to grow, paying off this loan shouldn't be difficult for them.

Debt financing isn't just about avoiding a "down round." It's also about maximizing the upside value for current stockholders.

There is some maximum possible business value of Facebook, no matter how much capital they take in. The founders hope to reach that value. If they issue more stock, even if they take in the same valuation for that stock, that final maximum value is being divided among more shares. Each shareholder's upside is limited as a result.

The lesson: -- equity is dilutive for growth companies, even if it's priced appropriately --

Debt also makes a lot of sense for variable costs like servers, where you know roughly how much revenue you'll get per additional dollar investment.

Akamai financed most of their infrastructure with debt, which was a huge drag on the stock price in 2002-2004 when people thought they were going to go out of business, but since has paid off very nicely for stockholders.

So how many servers does $100 million buy exactly? A bespoke built data center full of them I would have thought!

They're looking to buy 50,000 apparently

the question I have it how much of the new server farms are to accommodate growth and how much of it is them just struggling to have the current setup be resilient to the normal usage patterns that they see?

In the Startup Junkies series on iTunes, EarthClassMail borrowed the money to buy servers from a special bank made up of former venture capitalists who specialize in loaning money strictly for the purchase of physical hardware, also used as collateral. If Facebook already has almost half a billion dollars in investments, and feels this expansion will make them money, then they could pay off the loan without having to lower their valuation.

I don't know if Silicon Valley Bank is made up of former VCs, but they will lend to startups.

If you have real assets to back the loan (here, the physical servers they are buying), debt is a far better way to get money than equity. Sure, Facebook _could_ issue stock, but investors would expect a 20%+ return. Instead, Facebook issues some IOUs to a bank that expects a 10% return. The current investors don't get diluted and everyone goes home happy.

Buying physical assets with equity usually unnecessarily dilutive. Save your stockholders; issue debt instead.

It's impossible to know if this was a good deal for them without knowing the terms.

Also, Google actually did lease machines early on, which is essentially debt (done for accounting reasons, as I understand it).

Reply on news.ycombinator.com