*​I know this is a gross over-generalization, but bear with me.

As I look at the challenges of finding and building the right teams, it has become apparent that the entire recruiting problem is fundamentally broken*. The problem is threefold:​

1. There is no incentive for a person who is not right for the job to not apply - even if he/she knows they are not qualified. This means that the hiring company is deluged with applications and - as in any large sorting problem - shortcuts have to be taken and mistakes are made.
2. Companies traditionally don't train their employees how to conduct an interview or how to write a job description. There is a ​lot to be gained from an interview beyond whether you'd want to sit next to this person on a long flight. And writing a good job description not only helps the applicant better understand what they're getting in for, it helps anyone else (recruiter, HR, other interviewers) understand what they're looking for.
3. Applicants aren't given skills tests. In some industries certification exists to take the place, but certification is often a shortcut to assuming the right level of skill.

Since I see the most change happening ​with the third, and the biggest challenge in fixing the first, I'll take these in reverse order. 

Make them prove it

I think companies are starting to wake up to the skills problem. This is especially true in smaller companies where they cannot afford to engage in on-the-job training. (I'm going to focus on tech right now because I'm most familiar with it.) For developers, it's also problematic that the standards are changing. Now a good programmer will be able to learn new languages and frameworks, which is why it's important to focus on their ability to dissect problems. ​But, since there's no downside to claiming to know stuff you don't, you can still get an interview without having the skills. 

Some companies, like Interviewstreet,  are emerging to handle this problem by offering coding challenges and then utilizing the results to sort the quality of the programmers skills. Since most HR people can't determine a person's tech capabilities, and since it's often difficult to kill a phone screen in 5 minutes even if that's all it takes to know the person doesn't have the skills, automating this test creates a huge opportunity savings. (Think of the man hours and the disruption even with the phone interviews. This is time the developers are not developing.)

But it goes beyond ​just programmers. What about product managers or designers? One company that I know has a great process for interviewing product managers, where they ask you to interview a "customer" and then spend an hour with a developer and a designer walking through how you would solve their problem.  It's really the only way to know whether a person has the capability of active listening and distilling need from a laundry list of asks. Designers should be judged not just on their portfolio, but also on their ability to explain how the design met the business objective and the performance of their sites.

I've never run a restaurant, but I would hope that a chef doesn't get hired until she's cooked something that the manager has eaten.​

Be more descriptive

But are skills filters the only thing that you can do to reduce the number of inappropriate applicants? No. You can write a better job description.

In a lot of cases, the job description is cut and pasted or shoved into a format. It's seldom written with the intent of making it clear about what it will take to be successful in the job. ​In most cases, it's actually written to cast a fairly wide net. But where is the rule that says you can't ask for a detailed application that requires the person to - in their application - make their case for why their experience is applicable. You should write out all of those things that you'll be filtering for later. If you won't hire someone who can't work with pivot tables in excel, then say it (don't say, "Required", say, "if you can't demonstrate advanced knowledge of pivot tables don't apply").

 So, does that mean that we don't have to interview? Of course not. You're going to be working with the person, so you better get a full sense of them. Verbal communication skills and interpersonal skills are important, you need to see how they play out. But, as I say, interviews are expensive. So, they should be limited to the best choices and they should be used to really measure something. If the person will have to give presentations, why not send them a topic ahead of time and have them present on it. If the person will be pair programming, have them pair up with a developer or two. If the person will be changing bedpans, don't have them tell you about a time they did it, talk to them while they're doing it. 

There are a ton of theories about conducting effective interviews. That's fine, but companies ​should train people who are in positions in which they will be interviewing candidates (which could be almost anyone in an organization these days) on how to conduct an interview. What they should be looking for. How to structure their questions and note the responses. How do ask follow up questions. It's a waste of time and money for a company when the process falls off the wheels because the interviewers don't really know what they're doing. It's also unfair to the candidates to use them as crash test dummies. 

Pay for performance in the applicant management business

​Ok, now we get to the most counterintuitive suggestion that I have with the application process: get less applicants. People who are coming to you should be expecting to invest a little bit in the process if they want to show their level of interest. If the person can't be bothered to fill out a more information that explains how their experience matches up, or taking a test to assess their skills, will they really be excited to come work for you just because you interview them? (Granted when reach out to someone to get them into your company, you may want to take a slightly different tack. But you should still test their skills.)

​The other way to attack this is to make sure your applicant tracking system is oriented not toward efficiency, but toward efficacy. Look at them for their ability to easily integrate customized tests and to match skills to job requirements. Since recruiting is essentially a lead generation business, pay them on a cost per qualified lead basis. (i.e., pay for your applicant tracking system based on the number of applicants that result in interviews and charge your recruiters for your time that you spend on unqualified candidates.)

You should also encourage people who are interested in your company to apply generally. This shows you a level of interest. It also enables you to go out with a more specific questionnaire if a job comes up in the future. Or let's you build a job for them. This way you can cast your net wide, but haul it in efficiently.  ​

Bringing in people is a difficult task. There is no perfect way to do it. But right now, the process seems to be filled with a ton of waste. Surely, we can be better. 

​