If a prototype is the first manifestation of your big idea, the beta transforms it into a product you might actually sell. The focus here is on usability and design. Your task is to create something so simple, so powerful, and so effective that people beg to become beta testers. As you amass comments and feedback, look for opportunities to simplify production and keep your cost structure lean - that's much harder to do after the product is released.

To develop a beta you'll need a T-1 broadband connection and enterprise-strength email -- providers like the Message Center offer hosted exchange service for about $10 a month per user. Also, invest in a feedback collection device -- SurveyMonkey's prices start at $20 per month to set up online questionnaires and compile user comments.

$$$ required: $500K to $1M

Objective: Build the core team that will carry you into the future.

You're ready to make some permanent (or semipermanent) hires. For most startups, the bulk of your hiring here will be technical -- people who can get your product to beta. Although it's shrewd to source talent globally, your core product-development team should be local. Enthusiasm is a precious commodity; look for people who are excited about your planned product. Hire the best you can afford, but as a general rule, Silicon Valley startups assume a burn-rate cost of $11,000 per person per month.

Objective: Let pros handle the admin so you can focus on the rest.

Accounting, payroll, and benefits administration aren't glamorous, but they're important parts of maximizing limited resources and keeping staff motivated. Early on, many startups tap one of the founders to keep the books using software like QuickBooks. That's fine when you have fewer than a dozen employees, but as you grow, a part-time bookkeeper or administrative assistant can help. Call in professionals for the heavy lifting, in the form of outsourced payroll and benefits administration services like Paycycle, Paychex or Ceridian. Staffing is the opposite: In a small company, each personality impacts the team, so avoid using outside recruiters and headhunters.

Objective: Solicit the comments you need to perfect the product.

Beta testing used to be a drawn-out ordeal, but for software, Web services, and online media companies, "agile development" has redefined the rules by emphasizing the release of fully functional products, asking end users for input, and addressing suggestions quickly to iron out bugs or add features. The lessons of agile development apply to other kinds of startups as well. Physical objects are obviously harder to alter, but the important thing is to solicit large amounts of user feedback (often by e-mail), respond to each comment, and incorporate changes quickly. Keep in mind that the factors that distinguish a successful product from a dud aren't always obvious. Guidance comes from highly granular research that measures the effectiveness of individual attributes and features.

Objective: Translate all that you've learned into a more realistic blueprint.

The product your company was created to sell may not be the thing you later unveil in the marketplace. "You're going to change your business three times by the time you're ready to launch," says angel investor Jeff Clavier. "Get used to it." You may be focused on the wrong market, or even the wrong product. Listen carefully, keep an open mind, and revise your projections and analysis accordingly. The assumption these days is that software and Web companies should break even after $20 million in investment; a hardware or consumer product company should do so with $30 million.

• Phase Four: Launch the product
• Searching for angel investors
• What a VC wants to see in you