here are some insights about Search from SurfCanyon CEO Mark Cramer

Q: Out of the many search categories, which 2 to 3 do you think hold the most potential for advertisers/marketers?

A: The categories that offer the most potential for advertisers are naturally those where the searcher has the greatest interest in making a purchase or conducting research prior to making a purchase. Shopping and travel are two obvious examples.

That being said, for any search category volume is going to be critical. Managing an advertising campaign has an incompressible fixed cost, so advertisers are only going to go where there’s enough volume to produce a positive return on their efforts. Irrespective of the degree to which the search experience is compelling, or the searchers motivated to purchase, this is often a very big challenge for anyone starting a new search engine.

Said another way, if a category is going to offer potential for advertisers it must first attract enough user volume to justify having advertisers investing their time and money. Therefore, you might simply look at categories that appear to attract the greatest user interest.

Q: From a business perspective, how difficult is it to actually get a viable engine?

It is very difficult. On the one hand, the components required to build a search engine have become much less expensive. Building and maintaining a comprehensive index of the internet is out of the range of all but the largest and/or most highly-funded companies, but many search engine may now rely on BOSS or SilkRoad. Servers and bandwidth are less and less expensive and many of the necessary software building blocks are open source. Depending on what sort of engineering is required, outsourcing development might also be a possibility.

Just because it’s built, however, does not mean that users will come. The first critical step is naturally having a search experience that not only delivers a lot of value, but is significantly differentiated enough from what’s already on the market to compel someone to alter their search habits. While it’s true that any new search engine is only a click away, capturing enough of the user’s consciousness to induce a change of habit is, to say the least, daunting.

That being said, if a search engine can a) keep costs low and b) capture enough mindshare to generate reasonable traffic, then becoming viable is simply a  matter of monetization, which can be done in any number of different ways.

Q: What kind of $ is needed for R&D before launch? SearchMe, for instance, reportedly had $44M and its CEO said it would take a total of $100M. Does that sound like it’s on target?

A: This depends on the category and what exactly the search engine is looking to accomplish. If, as noted above, the company wishes, or needs, to build its own index of the internet, then a tremendous amount of money is required. I can’t remember where I heard this, but I have a vague recollection of someone (Steve Ballmer?) suggesting that it would cost ~$300 million to build a comprehensive index of the entire internet. The number comes down significantly for vertical search engines, but it’ll still be substantial. Also, Cuil has purportedly developed a much less expensive method for indexing the entire internet, but even at 1/10th the cost it’s still out of the league of most everyone.

Developing the technology will of course depend on the nature of the new technology. Visual front-ends, for example, can be produced rather cheaply. taggalaxy.de is impressive and was built by a single individual. While I don’t have the details, RedZ, which has a SearchMe-like interface, is probably another good example. Search engines that require algorithmic work, especially if scale is required for network effects or optimization, will be faced with a considerably more expensive task.

Lastly, the most expensive element might be Marketing. Unless the search engine is attempting arbitrage (purchasing traffic for cheap and then converting it into higher-value traffic), it’s also potentially the most risky.

Even if there’s a vastly superior mousetrap, building brand recognition is costly and a considerable amount of Marketing dollars will likely have to be expended before there’s enough volume to monetize. It’s not surprising the Microsoft has allocated $100 million to promote Bing.

However, once again, if the company is able to control costs and build a more compelling search experience, regardless of the category, then perhaps they could grow the traffic organically to the point where it can be monetized enough to make the business viable.