One of my favorite phenomena over the past few years -- and one of the stories I like to tell most about why the internet is awesome -- is something I call "Prescient Markets". Marketplaces (or just producers in some cases) that take a large amount of the risk out of producing & selling products by only building what they know, in advance, that people will buy. The original, and still my favorite example is Threadless. In case you don't know it, Threadless is a t-shirt making/buying website where users upload the t-shirt designs. Then, other users vote on the designs. The highest-scored designs are then actually produced as shirts and put up for sale. "Winning" designers are given $2500 cash if their design is selected (in advance of any sales, mind you) plus $500 in Threadless gift certificates and $500 every time their design is reprinted. With this system, Threadless drastically reduces their editorial role in deciding what to produce, based on the assumption that a shirt with many votes and a high average score is likely to sell. (of course, it's no guarantee, as the people voting may not actually buy, or may not speak for real potential buyers, but it's a really good head start). Then, perhaps my favorite feature of Threadless is what happens when shirts go out of stock -- if you see an out of stock shirt that you like, you can sign up to order it in your size, and if enough people do that to justify a run, new shirts are printed and put up for sale. Another site that takes this approach is
One of my favorite phenomena over the past few years -- and one of the stories I like to tell most about why the internet is awesome -- is something I call "Prescient Markets". Marketplaces (or just producers in some cases) that take a large amount of the risk out of producing & selling products by only building what they know, in advance, that people will buy. The original, and still my favorite example is Threadless. In case you don't know it, Threadless is a t-shirt making/buying website where users upload the t-shirt designs. Then, other users vote on the designs. The highest-scored designs are then actually produced as shirts and put up for sale. "Winning" designers are given $2500 cash if their design is selected (in advance of any sales, mind you) plus $500 in Threadless gift certificates and $500 every time their design is reprinted. With this system, Threadless drastically reduces their editorial role in deciding what to produce, based on the assumption that a shirt with many votes and a high average score is likely to sell. (of course, it's no guarantee, as the people voting may not actually buy, or may not speak for real potential buyers, but it's a really good head start). Then, perhaps my favorite feature of Threadless is what happens when shirts go out of stock -- if you see an out of stock shirt that you like, you can sign up to order it in your size, and if enough people do that to justify a run, new shirts are printed and put up for sale. Another site that takes this approach is
Quirky
, which is basically "Threadless for stuff." Through a
, product ideas are honed, and ultimately a final product idea is developed. Quirky employs a presale process that proceeds to manufacturing once a certain threshold is met. Quirky's process is infinitely more complicated and collaborative than Threadless', and to be honest, I'm impressed that they've been able to successfully architect, describe, implement and manage their scheme. Revenue is shared with each product's "creator", plus a (potentially quite large) group of "influencers" who have contributed to the development of the product in certain ways. One really nice touch is that when you order a product from Quirky, the picture of the inventor and the names of all the influencers are on the outside of the box. Another, super high-profile example of a related model is
-- with a modest fundraising goal of $15,000, the creator Scott Wilson raised $942,578 in pre-sales. Last I checked he was on a plane to Beijing to scope out manufacturers. Kickstarter is slightly different than Threadless and Quirky because there really is zero risk for the producers -- they're not committing to build anything unless they get the money up front -- and I almost didn't include it for that reason. What I think is particularly interesting Quirky and Threadless is that they are using crowdsourcing and community to de-risk a product development bet. Naturally, I'm wondering how this kind of a process can be applied to other sectors. In particular, it's gotten me thinking about how it can be applied to the government technology buying/building market, which we focus on with our work at
, and cities are running out of money. So something's got to give. Somewhere within all this mess, I suspect that there's an opportunity for governments, vendors, startups, and civic hackers to make something good out of this idea.
Whenever you start a project (and I'm thinking about building websites and web applications), you are balancing two somewhat opposed goals: 1) get something working right away and 2) satisfy all your hopes and dreams. The first, I think, is a good instinct. The second is the real challenge -- it's your wildest hopes and dreams for a project which can ultimately stop you from just getting something working, now, that accomplishes the smallest essential essence. Paul Graham and Eric Ries describe this as the Minimum Viable Product -- I agree with that idea completely -- but sometimes in practice it feels like even "MVP" ends up being bigger & more complicated than it should be. I've worked on projects where the idea of "getting to MVP" looms large over the team -- I feel like when that starts to happen you're actually building more into your MVP than you should be. The beautiful moment in building a product is the first time when it actually serves some basic need, and does that in a way that's fun to use. And that moment can't come early enough. So, if you find yourself debating problems you don't yet have, or arguing the nuances of the perfect, most elegant data model, maybe the thing to do is stop completely and ask yourself if the most basic essence of what you're making has been built, and if it has, if it's fun to use. If it's not built yet, you should stop and build just the absolute simplest thing that works. Not for public consumption, necessarily, but for yourselves and your team. If it is built, then you should ask: is it fun to use? If it's not fun to use when it's at its most simple, it's only going to get harder to make it fun to use once it's more complex. Simple, and fun to use. I'm not saying I've successfully approached every project this way, but I try to, and will keep trying to.
This weekend, I built some shelves in my closet. It was pretty simple affair -- some pre-finished shelving boards, wooden corbels, and a rod for hanging things. What's funny is that the supplies for all this have been sitting on the floor in my office for about eight weeks now. Every time my wife asks me if I'm going to do the closet, I say "yeah, but there's a lot I need to think through, to figure out how I want it." I had big plans for super custom shelves, with beveled trim and all kinds of beautiful polish. And so while I thought about it, the pieces sat there for more and more weeks; meanwhile my clothes continued to pile up on the floor and become an undifferentiated mass because I had no shelves.
Finally, after more gentle prodding from Frannie, I took a few hours the other day and put up the shelves in the way that made the most immediate sense. It was quite simple in the end, but the change was dramatic. Now, instead of a big messy pile, I have shelves on three walls and everything is stacked neatly. My daily wardrobe has been refreshed as I've found shirts that have been on the bottom of the pile for weeks. Amazing! Some product is better than no product.
Now that the shelves are up, there are a few things that aren't quite right. Lucky for me, it's easy to put a few screws in the wall and move a shelf. So I did -- and I added another hanging rod when I realized I needed one. Simple -- a total of one more trip to the hardware store and 30 minutes of work. I'll surely make more changes in the future.
The point of this is not to talk about my closet and awesome t-shirt collection. This is about iteration and product development. And of course I'm thinking about it in terms of my real job, building things on the web. Especially when there' s a big team involved, it's really easy to get into a pattern of "think think think! argue! mull mull! get it perfect before we build anything!". When really, often times the best approach is to just build *something* and start using it, then go back and make revisions.
The old carpenter's adage is "measure twice, cut once". That makes sense when materials are expensive and decisions are permanent (i.e., you can make a rope shorter, but not longer). But with the web, like with closets, I prefer: measure, cut; measure, cut; measure, cut.
That's not to say I don't believe in planning; I do. I believe in working strategically from a strong thesis. But when it comes to building, less sooner is always better than more later.
This isn't a new idea! Of course not -- it has been written about extensively, and it's an idea I've been a believer in for years. But I'm still surprised by how easy it can be to fall back into a "measure twice, cut once" mentality.
Quirky
, which is basically "Threadless for stuff." Through a
, product ideas are honed, and ultimately a final product idea is developed. Quirky employs a presale process that proceeds to manufacturing once a certain threshold is met. Quirky's process is infinitely more complicated and collaborative than Threadless', and to be honest, I'm impressed that they've been able to successfully architect, describe, implement and manage their scheme. Revenue is shared with each product's "creator", plus a (potentially quite large) group of "influencers" who have contributed to the development of the product in certain ways. One really nice touch is that when you order a product from Quirky, the picture of the inventor and the names of all the influencers are on the outside of the box. Another, super high-profile example of a related model is
-- with a modest fundraising goal of $15,000, the creator Scott Wilson raised $942,578 in pre-sales. Last I checked he was on a plane to Beijing to scope out manufacturers. Kickstarter is slightly different than Threadless and Quirky because there really is zero risk for the producers -- they're not committing to build anything unless they get the money up front -- and I almost didn't include it for that reason. What I think is particularly interesting Quirky and Threadless is that they are using crowdsourcing and community to de-risk a product development bet. Naturally, I'm wondering how this kind of a process can be applied to other sectors. In particular, it's gotten me thinking about how it can be applied to the government technology buying/building market, which we focus on with our work at
, and cities are running out of money. So something's got to give. Somewhere within all this mess, I suspect that there's an opportunity for governments, vendors, startups, and civic hackers to make something good out of this idea.
Whenever you start a project (and I'm thinking about building websites and web applications), you are balancing two somewhat opposed goals: 1) get something working right away and 2) satisfy all your hopes and dreams. The first, I think, is a good instinct. The second is the real challenge -- it's your wildest hopes and dreams for a project which can ultimately stop you from just getting something working, now, that accomplishes the smallest essential essence. Paul Graham and Eric Ries describe this as the Minimum Viable Product -- I agree with that idea completely -- but sometimes in practice it feels like even "MVP" ends up being bigger & more complicated than it should be. I've worked on projects where the idea of "getting to MVP" looms large over the team -- I feel like when that starts to happen you're actually building more into your MVP than you should be. The beautiful moment in building a product is the first time when it actually serves some basic need, and does that in a way that's fun to use. And that moment can't come early enough. So, if you find yourself debating problems you don't yet have, or arguing the nuances of the perfect, most elegant data model, maybe the thing to do is stop completely and ask yourself if the most basic essence of what you're making has been built, and if it has, if it's fun to use. If it's not built yet, you should stop and build just the absolute simplest thing that works. Not for public consumption, necessarily, but for yourselves and your team. If it is built, then you should ask: is it fun to use? If it's not fun to use when it's at its most simple, it's only going to get harder to make it fun to use once it's more complex. Simple, and fun to use. I'm not saying I've successfully approached every project this way, but I try to, and will keep trying to.
This weekend, I built some shelves in my closet. It was pretty simple affair -- some pre-finished shelving boards, wooden corbels, and a rod for hanging things. What's funny is that the supplies for all this have been sitting on the floor in my office for about eight weeks now. Every time my wife asks me if I'm going to do the closet, I say "yeah, but there's a lot I need to think through, to figure out how I want it." I had big plans for super custom shelves, with beveled trim and all kinds of beautiful polish. And so while I thought about it, the pieces sat there for more and more weeks; meanwhile my clothes continued to pile up on the floor and become an undifferentiated mass because I had no shelves.
Finally, after more gentle prodding from Frannie, I took a few hours the other day and put up the shelves in the way that made the most immediate sense. It was quite simple in the end, but the change was dramatic. Now, instead of a big messy pile, I have shelves on three walls and everything is stacked neatly. My daily wardrobe has been refreshed as I've found shirts that have been on the bottom of the pile for weeks. Amazing! Some product is better than no product.
Now that the shelves are up, there are a few things that aren't quite right. Lucky for me, it's easy to put a few screws in the wall and move a shelf. So I did -- and I added another hanging rod when I realized I needed one. Simple -- a total of one more trip to the hardware store and 30 minutes of work. I'll surely make more changes in the future.
The point of this is not to talk about my closet and awesome t-shirt collection. This is about iteration and product development. And of course I'm thinking about it in terms of my real job, building things on the web. Especially when there' s a big team involved, it's really easy to get into a pattern of "think think think! argue! mull mull! get it perfect before we build anything!". When really, often times the best approach is to just build *something* and start using it, then go back and make revisions.
The old carpenter's adage is "measure twice, cut once". That makes sense when materials are expensive and decisions are permanent (i.e., you can make a rope shorter, but not longer). But with the web, like with closets, I prefer: measure, cut; measure, cut; measure, cut.
That's not to say I don't believe in planning; I do. I believe in working strategically from a strong thesis. But when it comes to building, less sooner is always better than more later.
This isn't a new idea! Of course not -- it has been written about extensively, and it's an idea I've been a believer in for years. But I'm still surprised by how easy it can be to fall back into a "measure twice, cut once" mentality.
