Friday, June 22, 2007

Three types of ideas - bad ones are often the best

Product ideas can be divided into three categories:
  1. Obviously good ideas that are very difficult to implement. Efficient cold-fusion, flying cars, and a lot of other sci-fi ideas fall into this group.
  2. Obviously "good" ideas that seem possible but haven't happened yet. Video phones and HDTV were in this category for a long time. I think this happens when people get excited about technology and overestimate the benefits (and possibly underestimate the cost). I just don't care that much about having a video phone.
  3. "Bad" ideas. Many of these ideas are truly bad, but some of them will in hindsight turn out to have been very good ideas. I put them in the same category because they are difficult to distinguish without the benefit of hindsight. Some examples are the personal computer ("why would anyone need a computer?"), Google ("there are already too many search engines, and besides, search engines don't make money"), and Blogger ("can't you just use Geocities, and besides, are there really that many people with something worth saying?"). More recent (and still controversial) examples include Facebook and Twitter.

I expect that a lot of people will argue with the fact that I've grouped "truly bad" ideas together with "seemed bad but were actually very good" ideas. Everyone thinks that they can tell which ideas are the good ones, but observation suggests that they can't. I do believe that some people are better at picking winners than others, but at best they have maybe 50% accuracy. I've seen a lot of very smart people (such as at Google) be very wrong about these things.

For example, I remember when the "upload videos and put them on the web" incarnation of Google video was first being developed. Nearly everyone inside of Google, including me, was very skeptical that anything worthwhile would ever be uploaded. The common prediction was that it would all be movies and porn. Of course there was some of that, but the skeptics were completely wrong about the lack of worthwhile content. Uploaded video is one of the most important developments of the past several years. Unfortunately Google Video was burdened with an incredibly bad upload process (it included installing a windows client to do the upload!), and YouTube, which was started AFTER Google Video launched, took over. I believe that mistake was partially due to the negative expectations.

Here's my point: The best product ideas are often found in the "bad ideas" category!

If you're a super-genius researcher looking to dedicate your life to a really important problem that you may never solve, then the category one "cold fusion" ideas may be a good option. If you've found some clever way to dramatically reduce the cost of implementing a category two idea (like Skype did to the video phone), then that might be a good option. However, the real "low hanging fruit" is likely to be found with the category 3 "bad ideas". You'll have to deal with annoying skeptics and haters who say that you are wasting your time, but sometimes you'll create something hugely important, or at least moderately successful (and they won't).

Realizing that good ideas and bad ideas are often nearly indistinguishable, there are a few more lessons to be learned:
  • Instead of endlessly debating whether an idea is good or not, we should find faster and cheaper ways of testing them. This is one of the reasons why open systems such as the Internet or market economies develop faster than closed systems, such as communism or big companies -- individuals and small groups are able to build new things without getting approval from anyone.
  • Your product idea is generally worthless, because to most people it's indistinguishable other bad ideas. Very few people will want to buy (or steal, or take for free) your idea because they already have their own bad ideas which they like better. (There are other reasons as well, but this is one of them. The way to make your idea valuable is to demonstrate that it's not bad by building a real product -- this also demonstrates your ability to execute.)
Obviously there is a lot more to say about these two points, and I haven't done enough to explain or justify them right now. Maybe in another post...


Michael said...

Yeah, Google video does have some great content. The "Google Tech Talks" series for instance. It's a great resource for someone who wants to learn an internet/computer related subject. The Steve Wozniak interviews are also great for young scientists and engineers, or anyone!

Ah yes, the low-hanging fruit. The first step is realizing it's there, and the second step is being ready to seize the opportunity when it arises. The second step seems to me to be the bottleneck.

secretGeek said...
This comment has been removed by the author.
secretGeek said...

good work Paul. I was nodding my head and agreeing with every sentence. thank you.

i came up with a product idea but thought it would never work, so i didn't implement it. luckily i blogged about it -- so when someone read about it, and succesfully implemented it, i was able to team up with them. now we have a product,

'bad ideas' can be great.

lb (

Unknown said...

How exactly was facebook ever considered a bad idea? It was an obviously good idea: use the social networking model then made hugely popular by friendster and apply it to colleges in which a physical social circle already exists.

Anil Philip said...

I smiled at this blog post. What I have developed is considered a "bad idea" by many at news.YC.

Unfortunately, executing till it reaches approval level takes cycles of iteration and learning, for which you need 'sponsorship' money (a.k.a. investment) and collaboration/cofounders. Easier said than done...

Ryan Baker said...

Yes, most "easy" good ideas are already out there, bouncing around one or more corporations (usually more), and among other individuals.

Adopting a test policy might work well for the easy ideas, but what about the ideas that aren't obviously good and are still hard? There are plenty there for which the payoff is well worth the cost.

I think another way of stating what you were saying is:

(1) There are things that for which the payoff is known to be large, and the cost is known to be large. Sometimes there is also the concern the cost may be infinite (impossible).

(2) There are things that the payoff is known to be average, possibly larger, and the cost is known or suspected to be large, possibly enough such that cost/benefit is not favorable.

(3) There are things that the payoff is unknown. Whether the cost is known or unknown becomes less important, since 0/10 = 0/1000.

Andrew said...

I'm living in Thailand where YouTube is banned.

Could you guys please, please, please make a video sharing site where the content from YouTube is filtered for offensive material so that it doesn't get banned.

Let YouTube be the freedom of speech video portal and put up another site for censored content.

Its driving me crazy not being able to get any information. I use Google Video but sometimes it just redirects to YouTube which is banned.

Nathan Braun said...

"But what if you don't like low-hanging fruit?"

There are enough opportunities for everyone now, dollars in bank accounts are only the beginning of where real value is added.

Let the beggars pick low hanging fruit. More value can and will be created in solving larger, "impossible" problems (without significantly impacting ones' lifestyle in a negative direction.)

All fruit is low now, anyway. The conglomerates will all be open sourced and communized in the blink of an eye.

Matthew Taylor said...

Another bad thing to do is to sit on a good idea for too long... Eventually someone will beat you to it!

Paolo Amoroso said...

Concerning ideas in categories #1 and #2, you might be interested in the PALEO-FUTURE blog.

RantyDave said...

A complete aside: Look how far we've come. In just a few short years we've moved from "windows native" being an important requirement to the point of near-obviousness, to the idea of installing a windows client as laughable and silly ... also to the point of near-obviousness.

Microsoft must be hating this.