Summarize the answers, present them to the candidate, and ask, "How would you address these answers?"

2.) Don't begin the process by visiting the candidate's Web sites. That's like wandering around the guy's office hoping to discover something interesting or useful. Most people wander a Web developer's portfolio because they're not sure what they want and are hoping something will jump out at them. This may be appropriate if your Web development focus is graphics design, but that is just a component to your Internet presence. Such unstructured Web site wanderings usually result in a collection of technical curiosities you like and therefore try to force fit into the solution to your problem. Stick to the objectives your managers gave you only have candidates show you Web sites that demonstrate how they would achieve those objectives.

3.) Hire the vertical specialist first. Let him/her put together the Web development plan and then assist in selecting the horizontal specialists. Expect the horizontal specialists to come and go, but hire a vertical specialist who will stay around as your Internet program manager.

4.) Avoid prima donnas. The term Webmaster was invented by people who were full of themselves. The Web is a swirling soup of everyday surprises with no one in charge. We are the masters of nothing, except how to be fast on our feet. If a candidate tells you they've been developing Web sites for years, ask them what they learned this week. Push a few questions where the only right answer is, "I don't know, but I can find out" and see if you get acronyms and arm-waving instead. Ask, "Can you teach my people to do what you do?" If the response sounds like a defense of the priesthood, you may not have a team player. And if your Web site is not adopted and supported by everyone in your company, you will not leverage this emerging technology anywhere near its potential. You may need to put up with prima donnas as horizontal specialists if they are creative geniuses, but your vertical specialist will need to battle them with strong leadership.

5.) Hire in-house specialists when possible. If you have a talented in-house graphics department producing your brochures, interview them to do the Web site graphics. The vertical specialist should be able to integrate them into the horizontal specialists forming the Web team, and they will come with a good understanding of your image, your customer base, and the pictures to convey one to the other. If you have a product champion in your manufacturing department dying to figure out how to put job status reporting on the Web, give her a seat at the table. The vertical specialist should be able to look over everyone's shoulder with guidelines of what works and what doesn't on the Web. Expect your in-house talent to show initiative and to put in some free overtime, but don't make them feel illegitimate by being entirely voluntary — allocate budgets and deadlines for their efforts, too.

6.) Match left-brained and right-brained workers to your tasks. If your company wants to field an extranet to support business-to-business commerce, you need efficient Web programmers, not the multimedia genius who knocked your socks off with the site he designed for the Rolling Stones. If you want to engage teenagers and promote consumer branding, you need the multimedia genius. Such left-brained and right-brained people should work together, but not to be converted into one another. Expanding the role of a successful multimedia genius into the dry Web programming needs of a Web site is often a way to snatch defeat out of the jaws of victory.

7.) Act fast to engage. Many jobs at your firm have a ramp-up period before employees can become effective. Because Web developers come trained at least in baseline functions and bring their own tools, you can give them a task immediately. I go to most interviews prepared with a few small ($200 to $500) tasks ready to assign to a subcontractor — design a splash page, get a given page listed on Alta Vista, produce a small investor relations page from an annual report. If the candidate appears promising, I assign the subcontract right then and there, 50 percent down and 50 percent upon satisfactory completion. My risk is a few hundred dollars with the deliverable helping a decision about thousands of dollars, and most candidates will put out $5 of effort for every dollar paid because they want to impress. But the biggest impact is on the candidates. They do not expect to leave a job interview with money in their pocket, to walk out part of our team, to see managers appear to make fast decisions in their favor. Web developers go to more than one interview, and this makes you stand out.

8.) Subcontract the entrepreneurial spirit. Ask vision questions about what makes your candidates get up in the morning. If the answers sound like founding a "dot com" may be part of their plans, you may be talking to a budding Internet entrepreneur. HR folks are skeptical of such candidates hanging in for the long haul, and indeed as employees they are often hatching plots to strike out on their own. The hype is out there, as is the lure of instant IPO riches. Because unemployed Web developers with experience are such a rarity, you can't allow talent like that to beat the pavement while your Web remains in the starting gate. If you see a mismatch as an employee, see if you can make them subcontractors, especially if in answer to your question, that is what they tell you they want. Subcontractors trade in results, not effort, which is in your favor if you can make the cost comparative to an employee.

In the beginning, your Web development project may be too undefined to ascribe a fixed-price subcontract to the whole thing, but you can almost always break it into smaller pieces. Is the first step a needs analysis, a written spec, a demonstration of a key technology component, a competitive survey? Each of those tasks can be subcontracted by defining it with measurable milestones; offer a subcontract price commensurate with what you would have paid an employee (including overhead and contingencies). Many managers resist this approach because they simply don't want to do the difficult work of stating clearly what they want. Forcing myself to do this, even with hourly employees, has improved my ability to meet deadlines and budgets.