Here's my take on how to best launch (or relaunch) a search engine like Snap. I'm being paid to write here, but can say whatever I want — cool!

1. Have a great tool

When Google was launched back in 1998, the Googlers didn't rely on marketing campaigns really. What they did back then was to build the best tool possible, and then rely on word-of-mouth. Free marketing, first transported by geeks, then the mainstream press, and then everybody and their grandma — but it comes at the price of offering a tough solution to a tough problem.

In 1998, the problem was how to best rank search results, and the answer was to analyze the link structure of the web. Also in 1998, the problem was to realize that search was actually a problem — because many considered it solved and were now building portals. My number one tip to market a new search engine is to have it be a great search engine.

Once you build a great tool people like to use, how can you make sure it can compete with the big ones? By filling a niche, doing something that the others don't provide.

When Google moved into China earlier this year, there were many outcries of people who said — we don't want that to happen. We now need an alternative search engine. We need decentralized search, or a search engine that builds anti-censorship measurements right into its philosophy, right into its tools. Many people were ready to switch, but there was simply no serious competitor that did better in regards to self-censorship in China — in fact, most did worse than Google.

Web 2.0 offers many launches, but which ones will make it to 3.0? The ones that create solutions in areas where the big players either gave up, or likely won't go into. For example, it's unlikely Google, Yahoo or MSN will create a search engine specifically tailored to adult web sites, simply because they'll ruin their image… and this creates the opportunity for others to take that spot. Apparently, the big three also don't resist censorship… and this creates opportunity.

If Snap is significantly different from solutions offered by Google, Yahoo and others, and those differences appeal to some — they don't need to appeal to all, not even the majority — then it has a chance to launch well.

2. Create shareable content

When people like something, they want to share it with others. On the web that usually means sharing the link — you submit it to Digg, you post it on your blog, you email it to a friend (who emails it to more friends). In other words, the easier you make it to share your content, the more it gets the chance to spread. Here's what you can do to achieve that:

• Create neat URLs. When people want to send each other a link, there has to be a unique URL on the content they want to share. And that's not necessarily the case in many situations: a popup, a Flash application, a POST submit and other circumstances prevent the visitors from being able to easily share a URL. But even if there's a unique and visible URL on the content, it should look "linkable" and stable. (Usually, short descriptive URLs achieve that best — snap.com is a good basis for that.)
• Use Creative Commons. Creative Commons is a nice way to tag your content with a "some rights reserved — feel free to share" license. If you release your pages with that license, you make it explicit that you want others to take your content and run… and increase the chance they will.
• Create a REST API. REST basically means the server, based upon a parametrized URL, delivers a dynamic XML file to the developer, who then can go ahead and create new tools based on that data. A REST API for Snap would allow developers to build upon Snap data and functionality, which in turn helps to promote Snap itself (and might inspire an idea or two of what to add to Snap — did you know there was an unofficial Google API-based Alerts system before there was an official Google Alerts?).

3. Create something interactive for people to play with

From my blog, I just love to link to anything interactive. It can be silly or useful, but if you can play around with a website and create something uniquely related to your interest, then it becomes so much more. It becomes yours. I think one of the first things that people do when they try a new search engine is to enter something close to their heart.

What could that mean for marketing the Snap relaunch? Well, give people tools they can play around with to their heart's liking. Google Trends' search graphs feature is a nice recent example. Technorati had a similar feature before, and it also got people talking. It's one of the many approaches to render information in interesting, new ways. If Snap can give people the right interactive tools, then they'll start to have fun with Snap, come back to it, and talk about it.

4. Be radically real in your advertising

Shouldn't a commercial or something be part of a launch campaign? Print ads? Site ads? Yes, maybe, but then make it something real. Do you remember the colored balls jumping up and down a San Francisco street, advertising for some flat screen TV? I saw that linked on Boing Boing back then, and I loved that commercial… just the thought that somebody went out and believed so much in a crazy idea to make it real touches your heart! (OK, that video looked & sounded damn cool, too.)

5. Send out personal emails to bloggers

It's no secret that bloggers blog about stuff they think is interesting, so if you send an email to a blogger, or a couple of them, there's a chance they like what they see and write about it. I think there are only two errors you can make, and only one of them is significant: first, you can make the error to think you need an elaborate pitch to convince the blogger to use the stuff (you don't — it's the stuff you point to, not how you point to it, that matters to the blogger… bloggers act as filter for their readers), and second — and this is the significant error — you can come off as spammer by not personally addressing the blogger.

Part of a personal email, which can be just two or three lines long, should always be to address someone with their name (instead of the typical "Hello Sir" of Nigerian scam spam). I've once received an email from a well-known web company that went something like this: "We are sending this email out to top bloggers to …" That's right, it was non-personal, and bulk-emailish, and went on for many lines… and when I googled for parts of the email copy, as expected, I found the same "top bloggers" already attacking this company in their blogs…

6. Have some fun

There are lots of ways to have fun, even in the business of search engines. For example, you can put up special logos, hide Easter eggs in the search, or create "onebox" results on top of results. As long as the stuff is original — for example, you shouldn't only create holiday celebration logos, as it looks like you're copying Google — there's a chance people like it.

7. Serve the long tail of search

Serving the long tail of search can mean:

• Optimize for exotic searches… and aren't most searches "exotic" except for the person who's searching?
• Optimize for non-US countries. There are several countries where Google isn't top.
• Optimize for people with minority needs (is there a search engine specifically tailored to browsers for blind users, following the W3C accessibility guidelines?).
• Optimize for special devices.

If you optimize your tool for these groups or contexts, users will be happy to spread the word.

8. Be transparent

The more transparent a web site, the more we trust it — and usually, what we trust best we use the most. Being transparent means:

• Have a blog where you talk about pros and cons of your product (I'll only believe you talking about the pros after I heard you talk about the cons, too).
• Avoid wordings like "next-generation social aggregator for wifi-enabled web 2.0 surfers who …" etc., terminology which no one really gets.
• Respond to emails, respond to blog interview requests, respond to feedback (and demand it), get involved in blog comments, Slashdot comments, Digg comments, newsgroup posts…
• Let your team members become visible to the outside world (who's working at Snap? Where are they anyway? Can we have photos please? Who's the boss?)

9. Allow users to customize and socialize

There's two things Google, Yahoo and MSN search results don't have: skinning (being able to easily change the layout), and talking with others (being able to "chat" within specific search results or to comment on sites within results).

The new Google Co-op is a step to deliver a bit of socializing on web results, but it's too complicated to make it work right now as a user. Steps to make a search more social might improve certain aspects of it (for example, Digg made news more social, and digg.com's now one of the best or the best tech news aggregators).

Another thing a search engine could benefit from is skinning. You can use predefined templates, allow people to add their own CSS (I tried that once at a prototype search engine of mine — along with the ability to add comments to specific search results), or allow people to otherwise configure the layout of their search.

10. Give away free stuff

OK, free t-shirts are far from being original, but who doesn't like to get free stuff? I think if you can give away lots of goodies then people will at least step by to visit. Of course, you can't forever hand out free stuff to convince people to use your site (I think the last time we did that was in 1999), but if the tool you're promoting is great then people will come back another time. Which takes us straight to the last point…

11. Surprise people by delivering something "extra"

… and that's it!