What Not To Build | Hacker News Reader

What Not To Build

halfanhour.blogspot.com

200 pointsvorador72 comments

This is, literally, incoherent. The iPhone is a "fad", because:

most of the attention being paid to it has to do with applications and games, not telephony.

A few paragraphs later, under the list of "dead technologies", we have... "telephony".

Meanwhile, on his list of hot technologies is: alternative interfaces (where the guy actually cites the iPhone as an example!), portability, location-specific apps, intelligent recommenders, and sensor networks. All of which are being explored on the iPhone platform.

Maybe he means the general glut of wholly worthless iPhone apps that merely translate some web app or some gimmick onto the device. As an iPhone developer, I agree that most of what is going on falls into the fad category. If that's what he is talking about, then he's right. Don't sink time into building such-and-such for the iPhone when you could be exploring any of the other worthwhile areas he outlines in the list.

If that is what he means, he could have said so. It would have been a much more coherent argument. Which is to say: +1. ;)

And since when providing 'fadware (for lack of a better term) is an inherently bad thing? People are having fun with theire device, for a couple of bucks. Are we geeks critisizing anyone for /that/? C'om on.

By "telephony," he is referring to "in particular, dedicated lines and switched services."

Note that he isn't saying that mobile devices are a fad, but that the iPhone in particular is a fad. He is recommending not to develop apps for it. As we have seen on countless articles on HN, betting on a single app's success is like playing the lottery.

Do you mean the first time he used the word "telephony", or the second time? ;)

This seems to be what he meant the second time. But then what did he mean the first time? What does he think the iPhone should be for, if not for nearly every single thing on his list of up-and-coming technologies?

Meanwhile, I can respect the argument that the iPhone's app store is so crowded or so oddly designed that no matter how good your app is it's unlikely to find an audience. And I can respect the argument that, in many cases, something like Android is a better choice of platform than iPhone because Apple might veto you at various points. These, however, are not arguments that were made by the original submitter.

And I have very little sympathy for the "oh, it's just a passing fad" argument. Our entire field may be a passing fad, every existing mobile platform is a passing fad, people used to claim that the web was a passing fad, and my entire life is, if not a fad, quite definitely passing. If you want to experiment with a portable, location-specific app that has an alternative interface, what should you do? Deploy on a platform with no installed base and make all three of your users very happy? Raise N million dollars to market the thing that you stapled together in your basement, taking on an enormous risk? Sit around theorizing until the perfect platform with a perfectly-sized userbase with perfect demographics is presented to you by your fairy godmother?

Certainly a platform that sits squarely in the middle of your vision for the future of technology shouldn't be dismissed out of hand. It might even be worth working with for a while. Technology careers are all about brachiation: You stick with one thing long enough to spy something else to grab onto.

Here's my opinion on fads -- you gotta take a risk BEFORE they're ultra-popular and make sure that some of the first apps available for them are yours.

Then, as the apps market becomes more saturated, either: (1) diversify by developing LOTS of apps or by congregating other apps under your brand, or (2) become a freelance developer for that platform, advertising your first-app-to-market experience.

If you're a latecomer to the market and didn't take that initial risk, your chances of success don't seem to be great -- the market is saturated, and you don't have the platform experience to attract the highest hourly wages. (And now, getting experience by making apps won't be as profitable since the market is saturated.)

The type of risk that you propose to take has a fundamental issue for startups. As a startup, you can't do anything to make a platform itself ultra-popular. Say, you develop a new hip location-based service for Palm Web OS, but Palm Web OS never passes a few million users. You can do nada about it.

But what is worse, during that time, you don't learn much about your own service, because your biggest problem is to get any users for it. When you then decide to port it to more successful platforms, you still don't know what works, what gets people excited about your service etc.

In a contrast, on an ultra-popular platform, if your service goes unrecognized, you can do something about it and you can learn from your efforts. You try out different marketing techniques, maybe they don't work, but all the time you are learning and getting feedback of your effors.

I agree, but if you wait until a platform gets to be ultra-popular before developing, the chance of great success is much lower. It's classic risk vs. reward.

Apps aren't that difficult to make, so taking early risks on platforms like Facebook (with already a large userbase) and the iPhone (backed by Apple) aren't that bad, even if the platform fails.

Incoherent is a good word:

"Don't build a framework ... It's advice the Perl Parrot and Radaku [sic] projects should probably have heeded ages ago, advice the Ruby on Rails people should keep in mind today."

Why does he cite open source projects as examples of things not to build in a startup context? The fact that Awesome Project X isn't (a) going to make lots of money right off the start, and (b) isn't currently the most popular project in its area doesn't mean it shouldn't be worked on at all...

Roll the clock back 9 years and "Don't build a search engine" would have been on the list. Surprised it is not there now.

To me it seems it is the exceptions to the common knowledge that so often seem to exceed the rest. Off the cuff: Japanese cars are cheap (answer: Lexus), markets are efficient (answer: Buffett), VC's need home runs (answer: Y Combinator?).

I wonder how is each item on the OP's list might be mistaken and why?

It's not there because at this point, it's self-evident. There may be a path to profitability for niche engines, or engines which do something quite different from the big players, but it seems ever more unlikely that a new direct rival to Google/Microsoft/Yahoo will arise.

I mean, I get your point. But there's a difference between the pre-Google world and now: before Google, search sucked, and everyone knew it. Now, search is awesome. It's not really a pain point.

People were not aware of the potential for search before Google. They thought the market for search was over-supplied with options. Evidence is in some 10-year-old Red Herrings that mention: Alta Vista, Lycos, Web Crawler, DogPile, mySimon, etc., etc..

I was hard-pressed to make Google find anything 10-years-old+ from the web. And advanced search was no apparent help. Perhaps fodder for a rival?

Clayton Christiansen might say that out of the niche of a rival which undermines Google's business model rises the Google-killer, or at least -queller.

Google themselves have identified ways to make search better. Why text based search? Why not explore searches in a more interactive way? How about searching based on your existing thoughts and ideas? How can we make search work if you don't have a browser?

You don't need to compete with all of Google to succeed in the search arena. It is sufficient to capture 0.01% of their market to make a couple of startup founders successful.

I remember saying, in the late '90s, shortly before I saw Google the first time, that it was probably pointless to build a search engine as Yahoo had pretty much said the last word on the subject. That was a reasonably common thought at the time and Google proved it was completely wrong.

EVERYTHING on the OP's list is potentially wrong. Maybe he should have said "if you're going to build a {whatever} then don't just copy another's feature set - make sure you solve some problem that the market leader(s) failed to do."

While reading it, I remember thinking that posterous attacked blogging/microblog, which falls under the first section of stuff not to do, and they did pretty well.

posterous is mini-blogging though, of which the only real competition that existed was tumblr.

micro-blogging and full blogging are pretty much over-competed though.

wait, theres a difference between mini and micro blogging?

By mini, it's probably meant milli; so yes, there's three orders of magnitude difference. Wait till we get to nano and pico blogging: your status in less than one character!

That's actually kind of interesting. Some sort of monitor that just determines (based on alpha waves, heart rate, whatever) whether you're happy or sad.

The result?

^ or V

Over time, your friends see this

^^^^^^^^^^^^^^^VVVVVV^^^VVVV^^^^

if they see too many VVVVVV, you get a phone call. Dude, is everything OK?

Mini blogging (ala Posterous/Tumblr) is suited for a paragraph or two of text, or an image, or a video. It fills probably the best position for blogging for the majority of people on the net - who want to get across a concise thought/message, without the dedicated size of a full blogging solution.

Many of these are good, but a couple are wrong. We've funded both a social network (http://dailybooth.com) and a travel site (still only soft launched) and they're both doing well so far.

In regard to social networking, I think the author means "Don't build a generic social network (Facebook, MySpace, Orkut, Friendster, etc)" We have enough of those, and I think Social Booth doesn't really fit in to that category.

All useful modern web apps have elements of "social network" in them.

Good point. Facebook can't be all things to all people, so I definitively think there's room for niched community sites. Facebook hasn't crushed Flickr, for example.

He's not really saying it's impossible to succeed at these things. Only that there are other ideas out there with far better conditions and chances.

I simply don't see the reasoning behind high competition = low chances. The whole idea of web startups is to be able to move fast around big, slow competition.

Then by all means, out-facebook Facebook. I'll be cheering for you.

And Posterous can be counted as a counter-example too, couldn't it?

I imagine having YC behind you makes a significant difference.

Could you be more specific, please? I too think YC backing can be a great advantage, but surely it's not impossible to »make something people want« without it.

I hate to be the one to defend Java, but he's so wrong about this one that it's laughable. He and I and you and all kinds of people use Java apps every single day; they just happen to be server-based.

I think the article has good general advice but stuff you should really take with a grain of salt. As another commenter said, lots of counter-examples. When I was reading the "don't build edutainment" bit, I couldn't help but think of Brain Age.

The list is funny. I kept expecting it to keep going and turn into a parody.

Don't build edutainment.

Don't build an iPhone application.

Don't build a website.

Don't build an application that runs on any platform that is already popular.

Don't build anything that your customers can already understand.

Don't build anything that's like anything that people currently find useful.

Don't build anything that you expect to actually work.

Don't build anything that might turn on you and force you to stake it through the heart before it devours your spunky sidekick.

Don't build anything with a computer.

Don't build anything.

:D :D

On the surface this list makes a lot of sense. Each category of things not to build comes with a couple of examples of giant competitors: Wikipedia, Wordpress, Expedia...

If I were to build a startup in one of these categories, these would be the examples I would think of in the bar-napkin stage. They are obvious. I don't think 280 North will see this post and think "jQuery! We forgot about jQuery!"

There is always room for something giant and disruptive. More commonly there is always room for something small and sustainable.

I think that the big take-away from this shouldn't be: "don't build an application in X market" it should be "build a niche application in X market".

There's no way that you can compete easily with Facebook. But maybe there is an opportunity to build a social network for people that specifically don't want to share their real identities. Maybe you can't compete with Word, but how about a word processor specifically designed for vision impaired users?

The internet is a huge marketplace. If you choose a good niche and cater to it then you'll more than likely find a enthusiastic audience.

Great suggestions and observations.

But I'll remind everyone, especially those considering a startup, that they're not rules. Most likely, this article will be horribly wrong on at least one point in five years. It could be wrong because of your idea.

Just an FYI, the article is 11 months old

(which I think makes the article more interesting, as we can see the results of his predictions)

And while I'm here, he isn't saying the iphone wasn't/isn't important, just that it probably isn't the the most fertile field to spend time building a business on.

I think these are generally somewhat true, but with a lot of glaring counter-examples. As the companies that control these spaces get larger, there will always be room for smaller, disruptive competition, especially in more domain-specific areas.

I'm reminded of an article from HN earlier this year wherein a Harvard Professor admitted telling Mark Zuckerberg he was chasing his tail trying to create a social network. (Anyone have this link?)

In 2004, a plain old social network probably wouldn't have worked, but a Social Network of Ivy League colleges might.

Some good points, but I disagree with the what not to build section. If you have a great idea for one of these that is truly innovative, build it. Just because you'll have large competitors shouldn't stop you. Now if your idea isn't that ground breaking, that's a different story

Think of it more as playing the odds. Yes Facebook will be surpassed one day, but we don't know when.

Odds of succeeding to make the next facebook 1/5000 Odds of succeeding in building recommendation system for something that currently has no recommendation system 1/2.

Thinking about both of these problems for a few days you can come up with tons of interesting angles and technical problems. Every once in a while a major player will be replaced one someone else, but every day someone else will have a sustainable business in a less competitive area.

Don't start a photo site- Flickr

Don't start a search engine- Google

Don't start a "social network"- Twitter

Don't start a dating site- OkCupid

Don't start a word processor- Zoho

Don't start a blogging site- Posterous

Don't start a storage site- Dropbox

Don't start a project manager- Basecamp

For each company in this list there are another 50 failed attempts at cracking the same market. Several of which probably had the same or superior technical merits but the wrong business approach, had personality issues, or had plain bad luck.

There are likely other less sexy product ideas where the expected return on investment is much larger because the ideas aren't as obvious or the fields addressed aren't your in your typical hacker domain and haven't gotten the loving care of good software development.

"For each company in this list there are another 50 failed attempts at cracking the same market" .

I'm pretty sure you can say the same thing about almost all successful products in any market. Whenever I look at such a product, I'm sure that if I search hard enough I'll find the (at lease) 50 failed ones.

I've found that in technology, it's a bad idea to use negations. If you say things that should "not be done" or "will never happen" or "is not possible", at some point in the future, someone will quote you and laugh at you.

So, basically ... everything is good enough and established enough. Don't try to improve or compete.

Yuck.

I'll employ a well known Internet meme to describe this article: EPIC FAIL.

Are the current technologies really the answer to our computing problems? Are Linux, Windows and OSX the "final" operating systems, Microsoft Word the "final" Word Processor, Facebook the "final" social network?

No, seriously. I believe we can do better than this. In fact, I believe that the current crop of technologies are just a teaser of something better to come.

To quote Paul Buchheit, all such 'advice' is just someone's limited life experience expressed as a set of over-generalizations. If it was really this simple, VCs would have success rates upwards of 90%.

All said, I can understand how someone who reads proposals for a living could come off sounding jaded. It's pretty easy to pass judgment, especially from the sidelines.

Don't listen to lists telling you what you should or should not build.

Any one of the points he lists could have been acceptable if explained, but so many are dismissive and not fully reasoned.

For instance, it blows me away that people think cloud computing is a fad, when I know for a fact that computing power-on-tap has changed the way we and most of my startup friends do business.

He will no doubt be proven right on some things and wrong on others.

In general though, if I had a big pot of money and had to decide on where to invest it, my decisions would likely fall along the same lines. Key here, as another commenter said, is not that success would be "impossible" but rather "improbable".

He really went off the deep-end in the last paragraph. Transportation has been getting much cheaper adjusted for inflation, regardless of whether one looks at bicycles, cars, subway tickets, train tickets, plane flights or cargo shipping. Any blind assertion that all these things will get more expensive is difficult to take seriously. Ditto for his comments about living in "closer communities" and individuals growing their own grapefruit.

I suppose if the advance of technology ground to a halt, it could happen. What's far more likely, though, is that manufacturing and transportation will both continue to get cheaper, communities will be even less bound by geography and people who grow their own grapefruit won't be doing so for economic reasons.

I must admit a lot of the points made me feel the exact opposite: if the general feeling is that "application X is done, finished", it seems to be a tremendous opportunity.

It means that people are already blinded to the true possibilities in that area. They can only see the status quo and not see beyond it. Anybody entering the field with some revolutionary way of doing things might be able to grab a huge market share. They would not have any competition, because it never occurred to anybody to even consider innovating in that area anymore.

(note the Jan 4th post date)

2 things not to build seem contradictory: Don't build a platform-specific app Don't build a new standard While I can see limited usefulness of yet another markup standard xqrhtml43, generic apps, as opposed to platform-specific, have to look to standards to work upon. What should be read, however, is when building a new standard, extend existing standards to meet your needs, if at all possible, and make sure there is some sort of 'reference implementation' (ie, yours).

I understand this analysis as a heuristic and not rules to follow by the letter.

Disruptive innovation can invalidate the analysis.

For instance, I suspect such type of analysis would have classified web search as a no go domain when Sergey Brin and Larry Page started working on the google web search algorithm. There were plenty search engines provided by the big players at the time.

The added value (or problem solving) criteria is a much better heuristic to determine where to go or not.

The comment I left on his site:

I think this is a pretty one dimensional view on the world of innovation. You say "dont build an operating system", but you also say "do build cloud computing technologies" - but what if the operation system was designed specifically for cloud computing, or for decision support systems?

Why not build a better auction website? In your world would businesses still be doing manual payrole because 'we shouldn't innovate'?

Seems rather silly to say not to build a java application. Surely what matters is what the application does. If your idea will fail purely because of being implemented in Java then I think it must have some other rather serious issues with it. I can only think here he means some specific unstated context, but it sounds rather naive and ignorant as written.

I found this point quite interesting and thought provoking:

'From the point of view of the internet, 'you' are a concept - the one thing in the whole system that isn't actually a part of the system. How to leverage that will be the stuff of genius and innovation'.

I will welcome any ideas as to how it can be leveraged.

Know your competition.

I think that's the shortest summary of a mixed-bag of an article. And knowing your competition is mostly a marketing challenge, especially where the competition is virtual, like the NULL competition of non-consumption/non-use (of your product/service).

Don't build something you can't buy qualified traffic for and have a positive ROI.

Do: build a better mousetrap.

It's the "better" and "mousetrap" that are the hard parts.

Actually the "build" isn't all that straight-forward either. A common observation is that ideas are easy, what really matters is execution.

(Running through my head is that mousetraps execute mice - not sure that's helpful or useful.)

The "a" part is also hard.

I like the "sensor data processing" one. Measuring, grouping and manipulating traffic whether on the web or on the street will give search much more dynamics.

imho not so great article, its still best to build all the things you use yourself. Third party scripts are a pain and take about the same time to check for errors, bugs and faults then writing some from scratch. I dont mean that you need to re-invent the wheel, but you can make a wheel that drives even better on your road..... instead of using a drupal to make your website....

> There is a range of applications we might call Kantian applications - they depend on time and place.

I quite enjoyed this.

Hell of a lot of words to pretty much just recover, "Build something people need."

Very insightful article about ideas that sound good but are actually a dead end.

gist: Build around a problem that exists.

He mentions E-Government... I have thought about this subject for great length. My ultimate conclusion, is while it may certainly be possible to make it happen today. The idea itself is most likely not what is best (at least in todays world) My first idea on the topic was to make an open source platform for a virtual currency. I imagine the currency could be distributed via interest free loans. Transparency would be important. The key would seem to be convincing people to accept the currency, as it would have no value of its own... The system could on its own evolve into an E-Government as regular government experiences troubles. For instance the US seems to be heading towards an inevitable fate due to our growing national debt. As an ideal, this seems like a good idea (though i believe an economy based on debt is a bad idea, but couldn't think of an alternate method) but most people are not informed, and easily tricked. My solution to this was a competency test before you are allowed a vote. Of course, the test can be manipulated to a groups interests (http://en.wikipedia.org/wiki/Literacy_test)

half an hour a post really shows....

does hn count as a destination site? cus i come here an awful lot.... also to webcomics websites.... i don't use a rss reader, i just rely on firefox's url bar autocompleting when i type a letter.

Reply on news.ycombinator.com