Platforms are for Suckers: Why you shouldn't build your business on one. | Hacker News Reader

Platforms are for Suckers: Why you shouldn't build your business on one.

spencerfry.com

91 pointsdmitri198164 comments

I like everything about the post except for the word "platform." That is an overly broad word that does mean web browsers and the Internet itself and electricity. I prefer the phrase Proprietary Platform, because the issue is one of control.

If a single vendor gains control of the "market" for web browsers, as Microsoft did for a while, you are at their mercy. If there are multiple, competing web browsers and control of the APIs (e.g. HTML, CSS, JS) is decentralized, you are at the mercy of the whims of the marketplace, but no one person can wake up in the morning and decide to shut you out.

The concerns raised in this post are isomorphic to Apple's concerns about Flash. They don't mind developers using web standards to develop apps because no one vendor controls HTML, CSS, and JS. If developers write their apps in HTML, JS, and CSS, Adobe cannot flip a switch overnight and break every app on iPhone because their proprietary tools no longer emit a flavour of JS that runs on iPhone without errors.

So...

Proprietary Platforms are for Suckers.

I'd say the true target of this criticism is content-aware platforms, as opposed to content-neutral platforms, which I believe are what people often mean when they talk about "closed" and "open" platforms. The author favorably mentions targeting Windows and Mac OS, which certainly are proprietary, but they do not restrict the user of any content, essentially dumb pipes from hardware to content; these content-neutral platforms would be OK to build a business on. Platforms mentioned as bad were the iPhone, Twitter and Facebook, platforms where the vendor can and does restrict their users' access to content via arbitrary changes to their usage licenses, and such content-aware platforms are shaky grounds to build a business on.

Well, MacOS and Windows may be "content-neutral" in a certain sense but historically both vendors have gone into competition with developers... So from my perspective "proprietary" means sleeping with an 800lb. gorilla, and all we are debating is whether the gorilla wants to eat you or simply rolled over in their sleep and crushed you by accident.

http://www.panic.com/extras/audionstory/

The lesson I took away from Panic's Audion story is that they got gridlocked while trying to play two potential buyers (AOL and Apple), and Apple ended up moving forward with a less encumbered competitor. Please correct me if I am wrong, but it sounds like Panic wanted to get eaten by an 800lb. gorilla.

Oh? I think they wanted to get rich, but not to sell out:

In fact, I'd say that almost 5 minutes after the meeting Steve and I knew in our hearts that it wasn't time — that we didn't want to join Apple (yet). We maybe went through the motions of "deciding" on the flight back home, but I think we knew the truth. And the truth went something like this:

"This is our only chance to do Panic. We don't have kids, we're not married, we don't have huge obligations. We didn't invest our life savings into it, just a few hundred dollars. We don't even have life savings. We probably won't get this opportunity again in our lifetime — the full chance to take a complete risk, to experiment, quit our day jobs, start a business that certainly may fail, put our hearts into the soul of it, and try to make it fly — making the best possible Macintosh software we can without the threat of mortgages or the cost of braces or kids wondering why we're never home. And while there may be a time in our life where we crave some stability, or there may be a time in our life when things don't work out with Panic and we return to be a player in a larger, awesome team like Apple, that time is certainly not now. Panic's time is."

post scriptum: Whether Panic wanted to sell or not, the fact remains that building on a proprietary platform is perilous. Some people may still want to do so, but this is analogous to pointing out that BASE jumping is thrilling and deeply satisfying.

We can agree that Panic did consider a sale to AOL and Apple without taking anything away from the fact that a content-neutral proprietary platform still carries the risk that you will wake up one morning to discover that the vendor is giving away competition.

There's a huge difference between someone competing against you in the marketplace (yes, including with free offerings), and that same someone rejecting your iPhone app and preventing you from distributing it entirely. The risks are of different orders of magnitude.

If developing for Windows is an order of magnitude safer than developing for iPhone, how many orders of magnitude is developing for the web safer than developing for Windows?

I think the risks vary from platform to platform, market to market, and even possibly from executive making the decision to executive making the decision.

What are the risks of Apple competing with or even rejecting iDeco, an iPhone application for estimating decompression diving profiles? Minimal. Does that weaken the argument that building on a proprietary platform places you at Apple's mercy? I don't think so.

It's true that with iDeco they could reject it whereas with VPlanner desktop software they can only give away free decompression software. But this is like arguing the difference between living with a school principal who can expel you vs. a bully who can beat you up at recess.

If school looks good to you, go and enjoy yourself. The fundamental situation remains that when you build on a proprietary platform you expose yourself to risk of harm from the vendor. Manage that risk wisely and keep your eyes open.

> how many orders of magnitude is developing for the web safer than developing for Windows?

Depends on the primary source of your visitors/customers. If that's Google (as is the case with many web companies) then they could introduce a service much like yours and direct people there instead.

I think another important distinction is who is really dependent on whom. Are your success contingent on their them, and/or is their success contingent on you?

For example, Ventrilo was free to raise their rates because their success and profits really weren't dependent on the profits of the end users. They could jack their prices all over the place and most of their customers would have to pay if they wanted to keep their business going. They had lock in and they used it. They really don't care about how your company does, just that they extract maximum profit from you.

On the other hand, a company like Google has its success and profits dependent on your website being awesome. The more people who use the web and the more time they spend on it the more money Google will make. Google wants your site to succeed because it makes them money in the long run. They won't be as interested in jerking you around because they're not making money directly from you.

> Are your success contingent on their them, and/or is their success contingent on you?

This is the right question to ask. If A depends on B, but B doesn't depend on A, then B might shaft A. But if A and B are mutually dependent, both can assume it's unlikely the other with shaft them.

It still happens, though. See: Facebook and Zynga.

Or rather the advice should be not to build on top of a single platform. Building an API aggregator for example would be relatively safe since losing the functionality of a single API has reduced risk. And really in the case of browsers this is what we're doing already. If Microsoft decides to pull away from the standard there is pressure from developers since we are not at the mercy of IE.

It is funny that Windows is a proprietary platform and nobody (government) ever allowed them this kind of control/restriction.

Now Apple/Twitter/Facebook is going crazy and nobody does nothing...

The US vs. Microsoft case started in 1998 [1], but the investigation started all the way back in 1991. Given that the DOJ has just initiated an investigation into Apple over possible antitrust violations in online music sales [2], we can expect to maybe see a case in 2017 or so...

That said, there are many difference between the Windows platform and the iPhone. When the DOJ started investigating Microsoft in 1991 they did have a monopoly on PC operating systems, and (as the DOJ investigation found) was using this power to unfairly give Internet Explorer an edge.

The iPhone platform may command the lion's share of the attention in the smartphone market, it is nowhere near a monopoly [3, 4]. The same cannot be said about the iTunes Music Store [5]. It seems, however, that the investigation into Apple's practices are moving beyond music [6], so we'll see in 7 years time about "does nothing".

[1] http://en.wikipedia.org/wiki/United_States_v._Microsoft [2] http://topnews.us/content/221070-apple-under-doj-investigati... [3] http://mashable.com/2010/02/09/android-iphone-market-share/ [4] http://en.wikipedia.org/wiki/Smartphone [5] http://www.apple.com/pr/library/2008/04/03itunes.html [6] http://newsblog.thecmuwebsite.com/post/DOJs-Apple-investigat...

[Hope this wasn't too citation heavy, I tend to get a bit verbose when I am tired.]

I only write programs in my own custom programming language that run on my own in-house operating system. It only runs on computers that I design and manufacture myself. All the components are specially fabricated by me out of raw materials that are either synthesized in my lab or mined from my backyard. To spare you from the tyranny of the power company, it runs entirely off solar power, but it also comes with an emergency fusion reactor since the sun can just decide to go supernova any time it wants.

Do you have any safeguards in place incase atoms decide not to produce energy when fused.

If not, I have no trust in your platform

Google is certainly the biggest platform that people build on. How many businesses would die if their business had a catastrophic SEO failure tomorrow?

The point of platforms is that they make things easier, which is something that most businesses desperately need. You just have make sure that you understand the risks-- but they are oftentimes worth it. Zynga would never have built what they've built without Facebook. Yelp would be a tiny fraction of what it is without Google. Quickbooks would have a rough time without Windows.

That said, picking a really early platform has more risks (read: Twitter). Picking a platform run by a company that isn't really experienced being a platform (your Ventrilo example and, again, Twitter) has more risks.

But punting platforms as a concept is just dumb. 85% of venture-backed startups are dead in 3 years. More than that, likely, in the bootstrapped world. You need every advantage you can get.

Windows?

Spolsky has a good essay on that (and how they've lost): http://www.joelonsoftware.com/articles/APIWar.html

tl;dr version: Windows is a platform nobody wants to develop for. Windows developers have moved to the web (or to other platforms eg: Mac). The rest of the article justifies these two assertions.

I largely agree with the thesis, but can't seem to get my head around marketing on the open Internet without being tied body and soul to the Google platform.

Agreed. If anybody has some good ideas/links about this, please share.

Why are Bing and Yahoo not viable alternatives? Is it because your competition will out-market you using the better Google platform?

Why are Bing and Yahoo not viable alternatives?

Mostly because they have decisively lost the search market in the United States. I deal with one of the most non-technical niches imaginable, and Google gives me 10x the traffic that Bing does. (It is an interesting question of whether this is solely because Google is crushing Bing in market share, or if it is a reflection of adaptation of my business to meet Google's demands. That is largely academic, though. Its a factor of freaking ten.

Bing and Yahoo could both go dark tomorrow permanently and I might not even notice. The last time Google went dark, for 45 minutes, I started fielding customer support queries about how I "broke the Internet". And while that is amusing phrasing, I don't think it is false. Google is navigation on the open Internet. (Insert obligatory disclaimer: for English speaking Americans. It is a bit different in Japan. It is even more lopsided in the UK in Google's favor.)

Edit to add: Do not get me started on their PPC services. Its like Bill Gates said "We want this to have the technical chops of Google and the dedication to service of a convent!" and the memo got horribly, horribly misread.

Does Microsoft know that Bing lost? (Maybe from their POV, it's just waiting in the wings.)

Platforms are one of the best forms of leverage. I'd still use it but I would definitely want to have an exit strategy for those points the writer pointed out. You adapt to a platform for market share; you could also adapt to its changes or blaze a new trail in a different technological space.

It's even worse. I started an entertainment website in 1998, when the Internet had a competitive search and advertising landscape. I didn't rely on a single source for traffic or revenue.

Today, Google Search sends the vast majority of that site's traffic and Google Adsense pays me for that traffic. I wasn't tied when I started, but I sure am now.

But if competition among search engines led another search engine to displace Google, then your users would end up using that other search engine to find your site, right?

In my perfect world, there would be several search engines each with roughly equal marketshare. They would all be important, but none would wield the power that Google does today.

Amen. My startup's too dependent on a single platform right now, and we're being burnt by it.

What platforms are for is speedy cash. The access you'll get to large numbers of people is a great help. Just keep your efforts cheap and dirty and put the money in your pocket as soon as you get it.

I could not agree more.

I am also dependent on a single platform and am acutely aware of the fact that this is not a good long term (6 mo +) investment of my time. However, building for a popular platform is a good way to make some quick cash without having to worry about anything but the quality of your product (ie. getting discovered, marketing, payment processing, etc.)

Is a meta-platform possible? Leaving 3.3.1 aside, what about a company acting as a clearinghouse that takes care of all of the platform issues for you, submits your app, markets it, takes care of payment.

Perhaps this is possible with Javascript web apps for mobile devices?

This is silly. What was it Paul Buchheit called advice? Limited life experience + overgeneralization or something like that? That's exactly what this is.

Platforms have their positives and negatives. What you should do depends on the unique combination of your goals and market forces. Broad, sweeping generalizations like this are sad because they miss the real point, which is that you should carefully consider potential future ramifications of building on a platform. Not that you shouldn't do it, but that you need to give serious thought about what success might mean.

Actually, the problem is not in platforms, but in proprietary platforms. As another commenter stated, it's turtles all the way down, but most of those turtles are either non-proprietary (HTML, Javascript, PHP, Apache, IBM architecture) or irrelevant to the application (it doesn't matter whether a Web application runs on Linux, Windows or MacOS).

You always build on a stack of platforms, but the only one that matters is the one directly underneath your application. If it is proprietary, and/or you have no influence on the way it works, you might get into a trouble as described by Spencer.

There are some people who work well within platforms and some who work well outside of platforms. A platform, in many ways, levels the playing fields. It gives a bigger chance to us 'normals'.

See, those who do well outside of platforms are those funny, smart people that are always the center of attraction. They somehow have this magical ability to get people to gather around them - they become they center of attention wherever they go.

Those who do well within the platforms are the people who are willing to spend time studying the quirks of the platforms, and optimizing for it. Within a platform, the popular people who are usually the center of attention just by the force of their personality can get overshadowed by the meticulous researcher dude.

So often, you'll find that people who are already popular and famous say to work outside of the platform. Those who know they are not the type of person to draw people to themselves by magic will prefer to work within a platform.

I use Facebook Connect as the only way to register in my new venture site. I realize it is far less than ideal, however, we're just starting out and FB connect provides a really good way of getting genuine people and traffic. It cuts out spam by the ton and there is lesser resistance to 'sign up'.

I agree though. Building on top of platforms is one way of being bitten in the derriere at some point.

This may seem like an oversimplification but like the article says all web apps are built on browsers and all browsers are built on an OS and all OSes are built on hardware, all hardware is built on a chip set etc, etc, etc, till you get down to the fact that all of it runs on electricity.

It seems like it's turtles all the way down.

That is an oversimplification, in important respects. There are multiple competing options for most of those platforms, responsibility for them is diffuse, and change happens very gradually.

I mean, yes, technically speaking, I am dependent on electricity. But that doesn't mean there is a $8 an hour representative at the power company whose job is to shut down 24 businesses today to preserve the power company's vision for what proper use of electricity looks like, and whose decision is beyond repeal.

That is an oversimplification, in important respects.

It's not so much the fact of a platform. It's the nature of the ecosystem that lives on that platform. How open is it? Is it subject to whims and sudden change? Are these alternatives? Are there open standards? Is there competition between implementations?

> That is an oversimplification, in important respects. There are multiple competing options for most of those platforms, responsibility for them is diffuse, and change happens very gradually.

Also, at some point it's not platform, it's infrastructure. You could assume that your business will use credit cards and online payment since this is part of the infrastructure but when you are using the payment API of company MegaCorp to have those payments, you might get burned with their particular Platform.

> I mean, yes, technically speaking, I am dependent on electricity. But that doesn't mean there is a $8 an hour representative at the power company whose job is to shut down 24 businesses today to preserve the power company's vision for what proper use of electricity looks like, and whose decision is beyond repeal.

This isn't a good comparison. The thing is the electricity provider is much more regulated by laws, more stable and slower in changes compared to some other companies. Also -- it's infrastructure.

Technically, some "power company vision" is kept since you can't actually put any two wires to make a gizmo and sell it on the market as a product.

Another interesting factoid: home users have a limit on how much electricity you can consume. After a given limit you have to ask special permit from the electricity provider and probably pay extra. The trick is all these things are stipulated somewhere and public.

The issue seems to be reliability. Ventrillo was a bad situation because it could change one day and break anything built on top of it. With an OS and some hardware, I can be confident that a program written today will still run a year from now, because those have proven to be reliable platforms.

If I could only be as great a sucker as Zynga...

That said, it makes more sense, to me, to suggest that each business perform a cost/benefit analysis for every platform they use, and quantify risks.

1. For some business categories, such as games, you will be on a platform. Perform cost/benefit analysis for your market. Does flash make sense? Does ActiveX? What are your risks, and what are your benefits? (Some would suggest the game market itself is for suckers.)

2. Many platforms, such as Facebook, are glorified advertizing and lead generation. Do not create for these platforms, but rather advertize through it. The best example to me is PopCap Games, which directs people to their casual games through their flash games.

3. The further you wish to push the envelope, the more risk a platform entails. Porn? Be as independent as possible. Corporate apps? .NET is not your worry point.

> The granddaddy of all reasons, and the only one strong enough in itself to refute any contrary opinion . . .

Setting himself up for failure by claiming that the coming point brooks no rational argument . . .

> is that at any time Twitter, Facebook, etc., can make any change to their Terms of Service that their heart desires, leaving you high and dry.

And fails. Although this is a reason to be wary of platforms, sometimes to have a business at all you have to build it under the dominant platform. This may not be true of your level of risk tolerance, but for some people a risky business is better than no business at all. I'm sure that Zynga's owners feel this way after their recent partnership agreement with Facebook.

Aside from the admitted fail in building their own software, there's some good points in this article.

In the world outside software, avoiding platforms can be equally hard. If you sell lawnmowers, cheap meat, car parts, then you may depend on respectively WalMart, McDonalds, GM, which are all platforms in their own way.

I wonder if there will be 'leakage' from iPhone software towards Mac software as the screws tighten somewhat on the App store.

It's like building a house of cards. The most solid platforms are open standards, like HTML. Other platforms may be shakier, especially if they're controlled by a single individual or a single company who can change their mind at any time. If you build your business around open standards you're less likely to suffer from changes in terms of service of "platform providers".

See also: Platforms are for Suckers: Why you should build one, and sucker suckers into building their business on one.

Building on platforms isn't a problem. It's actually pretty cool being able to use all these wonderful APIs. The problems start when your business is depending on one of them. If what you offer is a product that work on top of 10 platforms, it wouldn't be a big deal if one of them will shut you donw.

I give credit to the writer as an entrepreneur, but one thing he missed for sure is: platforms were made for an adaptable market. Maybe it was just the nature of their operation that didn't fit into the equation

Commercial platforms are for suckers. Open platforms are huge wins.

Can we have examples of open platforms, please?

So far as easy to deploy webstack platforms go, Joyent's SmartPlatform seems to be open: http://www.joyent.com/technology/smartplatform/ Code on github to run your own instance: http://github.com/joyent/smart-platform/

This is why I stay away from Google's App Engine. I don't get why Google keeps it closed source. At least with EC2 I'm only renting server hardware.

How about Amazon S3/EC2 Services?

Amazon seems pretty solid, stable, and flexible to me. And many startups are using them to scale.

A good test is to ask yourself how hard it would be to switch to another provider. If it'd be very difficult and/or expensive, you should make sure that's a situation you're comfortable being in.

author's experience is creating an app around a "platform" developed by one person. you still have to pick the right platform to build for, you know.

As one of those people chiding you, I see now why you make the distinction between those two kinds of platforms.

I do think you have a legitimate point, but I'm not sure how to fix this problem. There's so much potential in web platforms...

I don't think you can fix it. You take risks all along the continuum - if you are standalone you have more control but you need to get user buy in. If you go to a busy popular platform then you risk falling into the sinking sands caused by someone elses control.

in both cases people will fail from things out with their control - its just easier to point at if its someone else changing the rules. (and easier to grumble about)

edit: like the guy above says - you need to tailor your risk. On someone else's platform - look for quick rewards and work on the basis its only a matter of time until it goes away

"If I have seen further it is only by standing on the shoulders of giants." -- Sir Isaac Newton

One of the suggestions is that this was intended as an insult to a rival, Hooke, who was short and hunchbacked. Although popular, this suggestion has been disputed:

http://en.wikipedia.org/wiki/Isaac_Newton#Fame

Ventrilo's popularity is a priori evidence of how fucking dumb gamers are.

Language! But that said, the popularity of MMOs, with game dynamics all too similar to slot machines and resort casinos isn't a good sign. (I say this as a former Eve Online addict.)

Reply on news.ycombinator.com