It feels good to be back in the startup game, rested and ready for the next Internet bubble. Since last fall, things have really accelerated: more people doing more interesting things online than in the past five years. There are a few differences this time, and I'd like to point out what I think will be at least a few of the important features of this new era. I'll also show where I'm putting my bets and spending my time.
It's practically impossible for any company to go public these days, which is why the average person hasn't heard about the Web 2.0 Gold Rush. Back in 1998, everyone and his housekeeper was a VC or a day trader. This time, the public markets don't reflect the frenzy. But it's there. Venture capitalists are now funding start-ups at practically 1998 levels. Because this bubble has less public hype, it may not pop the way the last one did. It's time, once again, to be an optimist. Let's dig in.
To me, it feels like Web 2.0 has already come and gone: the days of Delicious and MySpace grabbing headlines are over, everyone has figured out that JotSpot is lame, the blog sites are all pretty much terrible (including Typepad, the service I use and that you are reading now), so finally we can get over the hype and start building the next generation of ways to work and play. Social networks really can be useful, even if the design is so bad that they cause river blindness. Tens of millions of people use photo-sharing sites. Millions have RSS start pages. Google will soon offer an online word processor (Writely, which I'm using at the moment to write this), an online spreadsheet, and likely some uber-startpage where you can do all your Google stuff (you heard it here first, kids). Companies like YouOS let you store ALL your information online and forget about your hard drive. Every day, entrepreneurs announce new products and services, some of which really have a good chance to change our lives.
What's clear to me is that most of the Web 1.0 companies, and I include Google in this category, are now sitting ducks. The business plan at any successful web site at this point should be to aggressively put yourself out of business. Startups are sucking good people out of large companies faster than I can delete my spam. Consulting is back. New ideas are back. New companies can once again thrive.
I'm starting a company to address the friendability problem. I want people online to be more friendable, and eventually to find people for long-term relationships. So I'm starting a personals dating web site. In June, I found a partner, in July, we bought our Domain name, last September we had the team together and started building, and in 2007 we are launching the most advanced dating site ever built...
Why, you ask? Why would I start a dating web site now, with about 20 million people already registered to Match.com, EHarmony, and other established sites? Because these sites are going to be dead soon. A new generation of dating sites is sprouting quickly, and they are all free. Granted, most of them are terribly designed, but they all offer the same functionality as Match.com without charging a thing. Things are going to change, and I'm going to tell you exactly how they will change.
The free dating sites are out there and gathering members quickly. One thing you'll notice is that they all have Google ads running up and down their pages. And who exactly is advertising on those free dating sites? It won't surprise you to learn that almost all the ads on those sites are paid for by EHarmony, Match.com, Date.com, etc. (And ringtones - don't forget ringtones.) I'm sure these large sites are thrilled to have a set of "low end" sites handing people over to them to sign up for subscriptions. It makes sense - a small percentage of people will sign up for a pay site, hoping the pay site is safer and more reliable and that the people are more worth meeting.
So it would seem that there is room for the free sites and the pay sites. In fact, you can imagine the pay sites buying up the free sites and keeping them essentially as they are - a lead-generating front-end for the larger pay site. Yes, it would seem that way, unless something comes along to disrupt this equilibrium.
And that something is a company called LTR.com. We have a new model, one that isn't based on Google ads. A model we think will cause the nice people at Match.com to stand by and watch as we suck all their customers out, leaving a shell of a site with no people in it. Eventually, the people at Match.com will lose their jobs and close the site. Because our model is better for everyone.
Much of what we're doing in Web 2.0 is taking Web 1.0 for granted, and that's exactly what we should be doing. We assume all the databases have all their data. We assume we can use AJAX in the browsers. Now what? Now it's time to start seeing how useful a network can really be. I'll talk about this more in a minute, but let me start with some examples of what's new and cool as of this writing ...
A few things I've run into recently that I like ...
- More information makes things more findable
- why back-up your data online when you can create it there in the first place?
- FileMaker Pro in Ajax - Get in before Google buys them
- A cool tool that will tell you how anything corelates to anything else (God, won't the Daily Show have fun with this little toy!)
- Become a fonero and get free WiFi around the world.
- A very extendable OpenSource Wiki
- Much better information than Realtor.com (yuck)
- Helps you optimize the amount of money you pay for a given airline seat.
Sacred Cow Dung
- A good technology blog
The Agile Manifesto
- One way to think about programming
- Small standardized things can be useful
- The Music Genome Project
- What I listen to at work
In addition to reading all the Harry Potter books in French (it only takes me 6 months to read 1,000 pages), I'm reading a few books you might find interesting.
Small Pieces Loosely Joined, by David Weinberger
- Let's start to think differently about the Web
Ambient Findablity, by Peter Morville
- I think I've bought ten copies of this already and given it to people.
Use Cases: Patterns and Blueprints, by Gunnar Overgaard and Karin Palmkvist
- Must. Use. Use. Cases.
The Elements of User Experience, by Jesse James Garrett
- Jesse James Garrett's primer on the web design process
Don't Make Me Think: A Common Sense Approach to Web Usability, by Steve Krug
- Sites like VRBO and Craigslist need to read this
Futurize Your Enterprise
- Finally, relevant again
The Tipping Point
- The Internet is more about people than about technology.
Once again, I'm not going to say how fantastic Google is. How many times do I retype my search terms to try to find better results? What clues do I have as to which terms will work best? I often find myself in a sea of Nextag and .co.uk and eBay and PizzaRewardsOnline search results that don't have much to do with what I want to learn. (What? I can get a flu-shot on eBay? How convenient - maybe I should get twelve.)
If you wanted to learn about how much calcium you need, for example, what would you type into Google? Go ahead - give it a try. Type in "Dietary Calcium."
Did you get a link pointing to CalciumInfo.com, owned and operated by the people who make Tums? Did you learn how most people have a calcium deficiency? Could you be one of them? Answer: no, you could not. This kind of search result is called a false positive, also known as a time waster. (You could say the same thing about FOX News, but that's another article.)
Now go to Wikipedia, one of my favorite sites. Type in "Calcium." You'll get an article that again sounds very authoritative and means to help but unfortunately is dead wrong.
Wikipedia is wrong about a lot of things, but it has the ability to learn and get better over time. While the Wikipedia entry on Calcium was based on information from the stone age, the entry on Montenegro was updated on the same day that country became independent from Serbia (June 3, 2006). The Wikipedia has many hands, constantly trying to make it better. The difference between Calciuminfo.com and WikiPedia is that some day, someone will put better information about Calcium on the WikiPedia site.
WikiPedia is based on the philosophy of Karl Popper, who claimed we could never ever really know anything for sure, and the philosophy of Richard Feynman, who believed it was better to keep an open mind and keep challenging our assumptions, increasing our knowledge every day, and never think we're done learning.
Can you say the same thing about Google? Kind of. Google is meant to evolve, but it's evolving according to the needs of advertisers as much as to the needs of people doing searches. Try typing "dietary calcium" into the Google Health database, and you get the same uninformative links as you do with a regular Google search. Google wants to build a universe of services around your need for advertising, and they'll buy and try pretty much anything to do that. Google is designed to push more than pull. Google results are NOT always more relevant than Yahoo!s. Google changes over time, but sites like Calciuminfo.com are likely to come out ahead of sites that try to tell the truth about Calcium.
Which brings us to a little topic near and dear to my heart: the truth. It's about as easy to find the truth online as it is offline. And experts are rarely the best source of the truth. Experts are often good at explaining things in simple terms that are dead wrong. Walter Willett, chief of the Harvard School of Nutrition, has written many books and articles, given many speeches, and had many opinions about nutrition, most of which turned out to be stunningly wrong. These days, he's more careful about what he says. He's taken a learning approach to publishing, saying what he doesn't know as much as what he does. His page on calcium is a good example. He may not have all the answers, but he's going slowly and watching for real evidence. He's stopped perpetuating myths. He's less of an expert now, but he's right more often.
Not only is the web evolving, but web sites are evolving. A good web site is now built to evolve along with its owners, builders, contributors, and users.
Is your web site built to evolve? Does your site point out what it doesn't know as well as what it does know? Does your site say that there's more to learn? Does your site ask people to help make it better? If you ask me, that's what this Web 2.0 stuff is all about.
As Peter Morville says, "We're at an inflection point in the evolution of findability." I truly believe that. I'm interested in findability, because that's where I think the web is weakest. When you don't know what kind of door knob you need, search for information about door knobs. When you know which knob you need, you need to find one and get it.
Finding things online is difficult. Google and Yahoo! aren't designed to help you find specific things. You get a lot of false positives ("We have the custom orthotics you want at Overstock.com!") and false negatives (the great orthotic clinic you're looking for is just down the street, but it's not listed anywhere). So one definition of findability is "making it easy to learn the location of things without any false positives or false negatives."
Today, we use search engines to do research and stumble into interesting web sites. But no one has yet invented a findability engine. A few companies are trying.
Transparensee.com is a good example of this new breed of companies in the findability space. Transparensee tries to glean as much information as possible from the Web and make it as structured as possible. So, for example, you can use Transparensee to search for homes by specific attributes that you won't find at Realtor.com.
This is the same approach taken by Shopwiki.com, a site that tries to keep adding more information about products so it can keep getting smarter. It isn't very smart yet, but it has promise.
Globalspec.com is a B2B metacatalog, where you can look for fairly detailed items. The site searches a number of B2B catalogs, giving you access to information that Google doesn't track. Globalspec works by maintaining a detailed index of many B2B categories, like CMOS Camera Image Sensors, that don't exist anywhere else. An index like this is called a taxonomy; it's a necessary step for making things findable.
Riya is a consumer photo-sharing site that uses pattern matching to make things findable. I upload all my photos into Riya. Then I "train" it to recognize my neice Nina's face by circling her face in a few photographs. From there, Riya can find Nina in all my photos. Riya does this with remarkable accuracy, resulting in few false positives (thinking another little girl is Nina) and few false negatives (not finding Nina in a particular photo).
How good are search engines at helping me find something? Let me give you a complete example. I'm an amateur photographer. I'm looking for a lightweight tripod that I can take on trips and hiking. I'm willing to pay for the lightest tripod I can get, and I'm not interested in anything over 2 pounds (1 kilo). How do I find the right one? Let's break it down into steps.
1: Search. I type "Lightest tripod" into Google. The first few results are for Bogen tripods, none of which is under 2 pounds. Looking at the ads on the right side, I am immediately lured to an ad for the "Trek-Pod," a device I consider light and tall enough but not stable enough for my needs. A good detour, but not what I'm looking for. The third Google result links to an article that says "Not the lightest tripod, but a good value nonetheless." Oh, that's helpful. The next several links are to online forums, where I find a link to a Davis and Sanford "Traveler" tripod that weighs 2 pounds and costs only $20 but only goes up to about my waist. I find that Gitzo makes fantastic carbon-fiber tripods and explore the Gitzo web site, which is a heavy Flash-based experience that doesn't let me compare weights directly and is difficult to navigate (you don't want to know what happens when you hit the "back" button). Finally, I find the tripod I'm looking for: it only weighs .68 pound and does everything I want. Then I realize that the metadata for this product is wrong - the tripod weighs 3.8 pounds (I had to learn that from another source).
2: Compare. Another link sucks me into Amazon.com, where I can't do the one thing I want: compare tripods by weight. I can list the tripods by price, but not by weight. By accident, I learn about the Simmons Compact tripod, which weighs 1.5 pounds but doesn't reach very high. So I type "Compare tripods by weight" into Google, which takes me to a page on Epinions.com, where the tripods are not compared by weight. Back to my Google comparison search, I find tripods at PriceSort.com, which (surprise!) lets me sort tripods by price. Nothing lets me compare tripods by weight. Trying to compare takes me in circles. It seems that Gitzo makes the kind of tripods I want, but I don't know for sure.
3: Find. I go back to the Gitzo site. I can't compare by weight directly, but I can look at all the models and find the lightest one (this is called an exhaustive search, something I do routinely online, which is why I'm always so exhausted). Finally, I find it: the Gitzo G1028 Mountaineer Weekend. Comes in two heights (at 1.5 pounds and 1.7 pounds), is well made, and is exactly what I'm looking for. Now I want to find the best price. I start at Amazon.com, which has pretty much every model of Gitzo tripod except the Weekend. Now that I know what I want, ShopWiki helps me find it at Pictureline.com for $385. PriceGrabber gives me similar results. A search on eBay shows one at the same price, but of course the exact tripod I'm looking for could show up tomorrow for less.
4: Buy. I check the retailers online again and verify that $385 is the best price I'm going to get, so I buy the one on eBay. Why? Because it comes from a dealer (reliability), I get the extra positive feedback (reputation), and I'm already registered there (convenience). I should have my new tripod in a few days.
Is this about what you go through to find something? Could the experience be made, say, 800% better? I think it could. I think it's time we start thinking about findability.
Local findability depends on three main factors: the metadata, the functionality, and the interface.
More and more companies are realizing the importance of good, up-to-date metadata. Many companies publish catalogs on paper as well as online, so the metadata is in this tug-of-war between the paper publishers and the online group. In a paper catalog, you must trim the product descriptions to fit the page, and every page counts. Online, more information is better, pages can be practically any length, and detailed descriptions can come out of hiding any time the user asks for them. At this point, the paper people and the online people are starting to realize they need to work together and that their systems aren't very well designed for the versatility they now need.
The functionality is typically a bare-bones catalog site, designed to look more or less like Amazon.com. This will change a lot over the next few years. Catalogs are often one-size-fits-all. The paper catalog has a set table of contents. But that doesn't mean the web site should be a mirror image of the paper catalog. Paper catalogs may have their plus points, but they lack in functionality. Search is hard and ordering is very hard. Online catalogs, on the other hand, have the ability to treat every visitor as unique, or at least as part of a group.
For example, let's look at an interesting site called McMaster.com. This site is run by a fantastic company that really cares about its customers. They have insane customer service - it's hard to imagine how they can get that many specialty items shipped that quickly. You go to McMaster when you need a piece of hardware that's difficult to find and you need it now. Looking at the functionality of their site, you have to admit it's possible to find things there. But how easy is it? In some cases, the site works really well. In others, the approach isn't optimal. For example, protective clothing and office supplies are in the same category. You have to give these people credit - their site is different and easier to navigate than any of their competitors' sites. But - could it be better?
One way it could be better is to address the structure. They could make a site specifically for people in the electrical business, another site for people in the plumbing business, another one for people in construction, another for people in cabinet-making or machine shops. Customer segmentation is one of the most important ways sites can scale to accommodate a growing audience. I wrote an entire book on the subject of natural customer segmentation (Futurize Your Enterprise, Wiley, 1999). By orienting the site toward different groups, the granularity and taxonomy can change, and that makes things more findable. After all, electricians don't often need pumping and filtering items, and they only need a subset of several other categories. So no single way to cut the pie will work for everyone.
Another way to improve McMaster.com is to work on the interface, and this is where good design plus AJAX gives us the secret sauce. Examples of Web 2.0 interfaces are popping up everywhere. One interface I want to show you is at InstantDomainSearch.com - my current tool for finding domains. This is a cool little utility built by Beau Hartshorne. Just type any character into the window and you'll see the application access the entire domain-name database, one character at a time. You never have to hit return. Hit the backspace key, and you get the result of your new query in seconds. So hitting return and getting a new page view are out; having relevant information come to you instantly is in.
Analogous to hitting return is clicking the mouse button. Go to DontClick.it and play around for a while. I might have to wait for you to come back, because it's a pretty amazing experience. It's built in Flash, not AJAX, but you get the idea. Who needs to click a mouse and get a new page when - again - the information can come to you? That's one of the key ideas behind Web 2.0 - the information comes to you. The old click/page/forward/back metaphors are going to melt away, leaving online applications more suited to navigating the content.
So catalogs are becoming Applications. With new standards for metadata, functionality, and interfaces, we can see catalogs morphing into octopus-like living systems, rather than just paper publishing projects. A good example is GiantChair.com (a company I'm an investor in). GiantChair gives publishers a system for maintaining their metadata, so they can keep their catalogs up to date. Not only can it print a paper catalog, but it also provides a turn-key ecommerce web site based on the same information. Better yet, all of the publisher's current metadata lives right on its web site, so the information for each book can be picked up by distributors and retailers and fed right into their systems using an industry-standard format. This way, a spelling change or addition to the metadata flows automatically through the entire system, all the way to Amazon.com, where it changes by subscription rather than manually. This is just one example of how much more powerful catalogs will be in the future.
The other half of findability is global findability. Someone knows what she wants, and you have it, but she doesn't know you exist. It doesn't matter how well designed your catalog is if she can't find your catolog in the first place. In the same way that I found my new tripod from a Google search (eventually), you want both your products and your catalog itself to be findable across the entire Web. At the moment, the best way to make your site findable is to pay attention to what the search engines do and try to optimize your site's chances for showing up on various searches. You can also pay for adwords, ads, leads, and other kinds of traffic.
We all need to do this. We all need to pay attention to the user's perspective from the moment she has a need until the time the thing she's looking for arrives where she wants it to arrive. The best way is to build models of different kinds of customers and try to learn how they go about looking for things. Does a purchasing agent use the same tools and keywords as an engineer does to find the same thing? Probably not. By modeling customers and getting their input, we can try to guess all the different paths they may take to your content.
One way your customer can find your catalog is if it's already on her desk. In the paper world, that means having hundreds of pages of paper nearby to access whenever she needs something you offer. Online, it could mean making a small application that sits on her screen, waiting for her to expand it and start searching/finding stuff right from your database. Why ask for a site bookmark when you can offer an applet that gives her direct access to your database?
I think the findability space is wide open. I see opportunities everywhere. Because I'm only one person, I'm going to focus on building three new companies to help make people and things more findable. If I had a venture fund at my disposal, I'd start more.
Large Scale Dynamics
Some friends and I are starting a company called Large Scale Dynamics (LSD for short) to help companies with catalogs make their catalog items findable. We're going to focus on catalogs because A) catalogs at this point are very broken, and B) new technologies like AJAX will let catalogs finally become mainstream publishing platforms.
My first company, Studio Verso, worked with several catalog companies (we put Office Depot online). I've personally worked with a lot of large-scale sites, and so have the other people in the company. So we're going to try to help make catalogs work a lot better. In fact, we're interested in projects that have large datasets to represent, whether they're dealing in tickets, scientific data, classified ads, personals, real-estate listings, etc. I think it's time to look at how we can make the data in large datasets much more findable. (In fact, that's a common theme for all three of these companies, as you will soon learn.)
At Large Scale Dynamics, we are cooking up a lot of ideas for AJAX-driven databases. We have assembled a very talented team and are looking for the following kinds of people:
- AJAX/CSS specialists who can work on implementing client-side solutions. We're looking for senior AJAX developers as well as junior developers or interns.
- Good interactive designers and information architects. We have a fantastic senior designer, but we'll need more people soon. We're also looking for a summer design intern, to play with AJAX and design some cool demos.
- Clients with large-scale findability problems. The bigger the better. We're hard at work on a new site for our first client (see below), and we're hoping to start doing work for other clients some time during the summer.
If you're one of these people or know someone who is, please email email@example.com - thank you.
Web-Wide Findability Project
As I mentioned, a number of companies are trying to take the existing information on the Web and come up with ways to make it more meaningful. Transparensee is a good example for consumers. Silver Creek Systems is a good example for businesses with large catalogs. These companies are doing the best with what they have, but we're a long way from being able to compare tripods by weight, toys by safety rating, brownies by sugar content, party venues by capacity or availability, wine by varietal percentages, blue-tooth headsets by length, law firms by number of partners, offices to rent by price per square foot, consultants by practice area, flashlights by brightness, diamonds by carat weight, cars by gas mileage, speakers by distortion, etc. Sure, you can do a lot of this at various specialty review sites WITH THEIR OWN FORMATS, but you can't do this web-wide.
And, once you've compared, you can't find what you've decided you want. Once you know what you want, what do you want to see next? In many cases, you simply want to see a map of your area with ALL products that match shown on the map (again, no false positives, no false negatives), ranked according to whatever criteria you specify. You could specify a range of products (i.e., all Mercedes Benz 500 SL cars made from 2000 - 2003 with fewer than 80,000 miles) or a very specific product (all 2000 model year silver 500 SLs with black leather seats, black top, with fewer than 35,000 miles, no significant body work, one original owner, and the optional moon roof in the hard top). And again, you just want to see them on the map, rated in some way that makes sense to you (by price and distance, for example). Sure, you can do this on a few sites (kind of), but you can't do this web-wide.
Suppose you're looking for a law firm in your town that specializes in software intellectual property and patents. What you want is to specify all your parameters and then - again - see the results on a map according to your desired sort parameters. There are a few sites where you can try to do that, but you won't get full results. And you can't do this web-wide.
But, with a little luck and a lot of venture capital, that is about to change.
I've been trying to start this company for 15 years, even since before the Web. I've written about it extensively in past blogs and in Part IV of Futurize Your Enterprise. Tim Berners Lee has written about it. Bruce Sterling has written about it. Now it's time to do it. To write a successful business plan in this space, all you need to do is figure out how to productize and standardize the metadata of all goods and services in the world. Then get a world-class team together, raise some capital, and get going. So that's what I'm doing. I'm looking for world-class people to help in the effort. If you know what the word spime means, and you're really good at doing something, send me your resume. I'm looking especially for marketing people, technologists, security, managers, and others who feel they are at the top of their game. The company will likely be based in California but may have a New York office.
If you want to read more on these topics, you'll find some of the essays on the left interesting. Just click on the various months. For example, start with "Build Wikis, not Laptops."
Looking For a New Home
It's been fun using TypePad as my blog tool, but all things must come to an end. I think the company is poorly managed and unresponsive to user input. It may be great for some people, but I'm ready to move on. The problem is, it's tough to move everything (I like to write), and when you find a new tool you want to think you're going to be there for a long time. So if anyone has any good suggestions for a great wiki-style blogging tool, bring it on.
Here Now, Your Moment of Zen ...
This is for keri dae sutterfield