(Business 2.0 Magazine) -- On the surface, everything seemed like it was going according to plan for Riya, a photo-storage startup based in Redwood City, Calif.

Since the site launched in March, users had uploaded 10 million photos, and the company had raised $19 million in venture capital from Leapfrog Ventures, Bay Partners and others.

But Riya founder and CEO Munjal Shah sensed trouble ahead - as well as an opportunity. For every user who uploaded a photo, 20 more were just using the site's search engine to find photos.

That kind of behavior wasn't accounted for in the original plan, which hoped to demonstrate its photo-organizing service on Riya's free website, and then license that technology to other photo websites.

"It's a trap that I see many Web 2.0 companies fall into," Shah says. "There's all this excitement about the tremendous growth in usage, things are going through the roof, so it's hard to see that it may lead to nowhere, that there isn't much of a business model there."

So Riya decided to abandon its original idea for a photo-storage business and focus on searching images on the Web instead. The site already had face-recognition technology that helped users sort through poorly organized personal photo collections. Now Riya would apply that same technology to searching through images found anywhere.

In the coming months, Riya will relaunch as a Web-image search engine, with a crucial difference: Point Riya to an image you're interested in, and it will find others that are similar.

"If your users are presenting you with a market that is much bigger than what you were originally going after, you have to listen," says Shah. "And if you are going to make that shift you also have to pull the trigger quickly."

Shah sees multiple, potentially moneymaking applications for Riya's photo search. For example, Riya could license its search technology to online dating sites, letting users select photos of potential matches they find attractive, and use Riya to find similar people - arguably an improvement over the time-consuming personality tests that now dominate most such sites.

Alternatively, Riya could run its own shopping search engine where clicking on an image of a dress or a pair of shoes would launch a search for similar items, with Riya earning e-commerce referral fees. Such fees have proven to be a lucrative revenue stream that Google, among others, has recently sought to tap.

Riya is hardly the first startup to switch gears. Over the years, Good Technology, based in Santa Clara, Calif., has morphed from a creator of MP3 player add-ons for handheld PDAs into a maker of e-mail pagers, and then to a provider of wireless e-mail software. Even Google (Charts) started out with a plan to license its search technology to others, rather than running its own search engine.

But what Riya shows is that the same phenomena that make it easier than ever to start a company also make it easier to change directions. The cost of everything from engineers to software to hardware is down significantly from a decade ago, which has made bets on a business model less expensive and easier to retool.

More recently, the photo-sharing website Flickr, which was bought by Yahoo (Charts) last year for an estimated $30 million, originally started out as a maker of online multiplayer videogames.

"In some sense, it's easier to get things right today than with the first crop of Internet companies in the late '90s, because it's not so expensive to get them wrong," says Allan Aaron, a partner with Technology Venture Partners, an Australian venture-capital firm that also invests in Silicon Valley.

"In the late '90s you would have spent all this money building a team and a technology, and if it didn't work out the first time, you just pulled the plug."

Whether the new Riya will be a strong business remains to be seen. What it does demonstrate is that today's new crop of startups have both the willingness and the capability to switch gears fast when they see a new opportunity. Tech startups can be built for so much less money today that they can afford to make mistakes.

Shah is savvy enough to see that his original vision for Riya was not headed for "escape velocity," as he describes it. He's willing to take another whack with Riya 2.0. And if that doesn't do it? He'll try something else.

"Change early and often," Shah says. "That's how we're doing it."

• Related: How to build a bulletproof startup