Predict Whether a Startup Will Succeed or Fail | Hacker News Reader

Predict Whether a Startup Will Succeed or Fail

xconomy.com

38 pointssanj16 comments

I'm troubled by this line:

* ARM processors are cheaper and worse [than Intel chips] *

Beyond the fact that they have a very different profile and are optimized for mobile, I don't think of ARM chips as "worse" in any sense. Their instruction set is beautiful where Intel's is... not.

That sentence continues but they’re getting faster with Moore’s Law. They're worse in terms of speed.

They are described as a threat to Intel in terms of Intel's market, the desktop, ie. they may invade the desktop from beneath, which would be a classic disruption.

Typical apples & oranges comparison. Arm chips have been optimized for an entirely different use case, the parameters are probably more closely related to stand-by power when idle and mips/watt.

To simply use the 'speed' metric when comparing a < 1W CPU with a > 30W cpu is dumb.

Arm chips will not enter the desktop market any time soon, the ARM architecture is for the moment solidly stuck in 32 bit address space and then there will be the problem of convincing a company the size of microsoft to bet on a different architecture.

This will happen in due course, but not in the next 3 to 5 years.

85% on a yes/no decision sounds more impressive than it actually is. The majority of startups fail so the baseline is somewhere above 50%. What would be more interesting is if it could calculate expected value at better than chance.

Right, simply predicting 'no' in every case could be right 80% of the time if you believe someone people's number on startup success.

On that note, I wish we had a better data set of attempted startups and general agreement on when a startup 'starts.'

> better is actually a con because it will invoke a competitive response.

Be stealthy, young startup, and non-threatening.

SINGLE-PAGE VIEW http://www.xconomy.com/seattle/2010/04/28/how-to-predict-whe...

I would be interested to see exactly what is meant by "worse." One of the common lean startup themes seems to be that a simpler product can be an advantage over incumbents who have succumbed to feature bloat. I surmise that simpler might equate to worse according to the metrics used in the analysis.

I think by "worse" it means sheer performance.

Think electric cars being generally underpowered and containing heavy batteries compared to their internal combustion engine counterparts which are powered by millions of years of natural geological processes distilled into a convenient liquid. Naturally, electrics are getting better over time. Ford and GM must have laughed for years at those silly beatniks and their wacko electric vehicles, but now GM is betting it's future on electrics.

37Signals' products deliberately do less than their competitors (less performance? the analogy doesn't really fit here) but are highly successful.

It can be sheer performance; going by the innovator's dilemma, in general it is whatever attributes are higher-ranked by the market (ie. customers). This makes sense, because that is what incumbents naturally worry about, and what they compete on. So if your startup is worse on those attributes, you don't threaten them, and they leave you alone.

So what's the point of your startup, if you're worse for customers? You're better for other people, who aren't customers of the incumbents, and who rank the attributes differently. There's lots of examples.

The coup de grâce is when both you and the incumbent improve over time, but the incumbent becomes better than it needs to be for its existing customers on those attributes, and your startup becomes good enough, but you also have those other attributes - then the incumbent's existing customers switch to you, and incumbent dies. It doesn't always go that way, but it does happen, and that's the pattern studied in the innovator's dilemma.

Incidentally, electric vehicles were popular at the dawn of the car; even edison was developing batteries for them. They've been overtaking the petrol car for a long time. But they have made inroads in other markets, such as those disability wheelchair carts (I'm not sure exactly why they're a better fit for electric, but they are certainly not a focus of incumbent car manufacturers - and maybe that's the important thing).

sorry, I can't help explaining this whenever it comes up, whether it's needed or not. It's such a cool observation.

No need for apologies, I find this kind of stuff fascinating.

"In 1906, Fred Marriott drove a steam powered vehicle built by the Stanley Brothers to an amazing speed of 127.659 MPH."

http://www.steamcar.co.uk/lsr_history.html

I think this is amazing considering the time it happened.

Henry Ford himself experimented with corn oil and ethanol as a fuel source for his vehicles in 1942:

http://en.wikipedia.org/wiki/Henry_Ford#Interest_in_material...

It's amazing that cars evolved to run on refined oil at all! I know I've gone completely off topic here ;-D

127.659 MPH that's terrifying. Talk about a steam punk.

I like Ford's plastic car... sounds like the soviet niva (sp?) made of plywood. Easy to ridicule, but cheap, economical, though not that safe. Unfortunately, once cars went mainstream, cool became the key attribute.

It's nice to observe that even the most famous inventors had many failures; you only need one success (probably, s/even/especially/, because more attempts --> more successes).

They are using Christensen's definition so "worse" means "worse by the standards of the market they are entering."

Those standards will vary widely according to what market you're talking about. For instance in the 70s and 80s, disk drives were judged entirely on size. So a smaller size of disk was necessarily worse. And this resulted in waves of companies wiped out. In mobile devices today, people care more about portability, ruggedness, and power consumption. So a smaller disk may actually be better. Therefore whether a smaller drive is better or worse depends entirely on the context set by the market it is entering.

In the examples you are thinking of, simpler frequently does equate to worse. If the existing companies have sales pitches that are geared towards lists of checkboxes on feature sets, then I guarantee that simpler equates to worse. And yes, this is an advantage for the startup. (Assuming that it can make itself lean enough to live on the scraps at the bottom of the market.)

The underlying dynamic is that existing companies find it hard to compete at the low end of their market for a variety of reasons. One of which is that companies tend to develop infrastructures that depend on having certain profit margins. Relative to how the company is structured, the bottom of the market is always unprofitable. Conversely companies always find the top end of their market very enticing.

However technology tends to improve at a faster rate than the market's needs increase. So the crappy cheap solution naturally grows up into a sufficient solution. The providers of the cheap solution happily eat up the market, and at every step the incumbents have a hard time defending it. And so incumbents get driven out of the top of the market.

This dynamic is frequently recognized while it is happening, but for deep organizational reasons it is really, really hard to change.

The main nugget:

But if [this model is] so successful, why haven’t more people—entrepreneurs and investors in particular—adopted disruption theory? Probably because models for predicting how companies will do are a dime a dozen

sounds like an argument for release early, release often

be crappier and cheaper and as you gain marketshare, improve

Explain Google and Apple then?

Google's main selling points were that (1) it was bloody fast, (2) the search results were better than everyone else's back then and (3) there were no banner ads, which helps (1) a lot.

Apple's main selling point, in its revival form, was that people are willing to pay more for a product that is a work of art.

Single-page view: http://www.xconomy.com/seattle/2010/04/28/how-to-predict-whe...

Reply on news.ycombinator.com