The heading of this section is a slightly fancy way of saying "How To Pull This Off" or "How To Recognize A Good Job Analysis." Every job analysis should have the following stages. If you skip any of them, you do so at your peril.
1. Get Management Buy-In. This might be a broken record depending on how much of the site you've read to date, but repeat Renegade Psychology Mantra #1 after me:
No significant human capital project will succeed in any organization unless all the major stakeholders, AND THE MANAGERS OF ALL THE MAJOR STAKEHOLDERS, agree to support it. Why? Because almost all human capital projects require at least token participation from current employees and managers without additional compensation. A job analysis requires you to ask these stakeholders to take time out of their busy schedules for meetings (either virtual or face-to-face) to develop lists of critical knowledge, skills, and abilities (KSAs). It requires them to trust that you won't use the information you (or your consultant/s) collect against them (which is why trust and participation can be raised if you hire consultants).
In short, you'll be trying to gather knowledge and data from employees who don't report to you and don't owe you squat. If you've tried this before, you know this can be like herding cats. The best way to push through any potential resistance is to get the full commitment of the top bosses of their department -- and to make sure all critical project communications originate from their office. That way, any resistance places the participants in defiance of their managers (“Um…this notice came from Frank’s office?”) instead of pitting them against you ("I spit on your survey, HR swine!”). Believe me: when their department heads respond by saying "I find your lack of faith...disturbing", that resistance is much more likely to collapse.
2. Assemble a project steering committee to oversee things. This group should be composed of you, hiring managers, applicable HR personnel, and your third-party consultant (if applicable). Keep this team small: it will meet regularly to ensure the project is proceeding on schedule and according to plan.
3. Put Together A Realistic Plan Before you Start Collecting Information -- and share that plan with your stakeholders BEFORE you begin. Which jobs will you be looking at? Who will be on the team that steers the project (if applicable)? Who will have to participate in the project for it to succeed? What degrees of participation will be required? What points in the project will likely elicit resistance? How will you collect (gulp) survey data? How will you analyze the data? What is the desired outcome? What will you do with the results? How much will the project cost (in terms of people, infrastructure, money, and time)? This should all be established up front.
Consultants or no consultants? If no consultants, you'll have to develop those KSAs on your own by doing a lot of research -- which means meeting with job incumbents and managers to develop KSAs. (But easy on the managers: they probably aren't actually doing the job today, and they could influence the KSA lists and survey results in detrimental ways.) This stage will also will involve consulting existing job descriptions -- starting with the ones your company already has -- and resources accumulated from other organizations. Believe it or not, there are published books full of job classifications and KSA lists including the Dictionary of Occupational Titles. You can use a tool people refer to as the "Internet" to find them. The consultants you hired (if you were lucky enough to have the budget) probably utilized source containing information at least partially culled from these publications.
If you hire consultants, they will still have to conduct research within your organization to make sure your jobs closely resemble whatever data they already have on similar positions in other organizations. They will also put together a plan. If your hired guns proceed as though your position is familiar to them, feel free to ask them what resources they will rely on. They shouldn't be afraid to answer. If they haven't followed the preceding steps, they may have some 'splainin’ to do.
4. Let the employees know what’s up. The department heads (not you) should let all affected employees know that a job analysis will be taking place, and at what points they might be involved in the process. Be open and honest: you have nothing to hide, so there’s no need to be mysterious. Reassure them that their participation won’t affect their job status or prospects in any way whatsoever.
5. Begin the 'material' portion of the project by scoping out the job(s). There are a few established ways to collect the information you'll need to get a handle on the exact nature of each position after you've reviewed the existing literature. The most obvious one is -- gasp! -- watching *established* incumbents (6+ months tenure) on the job. Crazy, right?
But this is trickier than it sounds. (Most human capital projects are, even though some ignoramuses, aloof rich guys, or crappy managers suggest otherwise. Successful people sometimes poo-poo human capital issues because they operate under the mistaken belief that their achievements automatically make them experts on all aspects of business. Don't listen to them. You have a job to do, and the only one who will be blamed if it isn't done well is you.) You need to make sure you observe a broad, diverse sample of employees -- and observe them in as many different locations and environments as possible. This will supply you with the most thorough representation of what the job is like regardless of location, employee age, race, etc. Good job analysts do everything they can to ensure diversity at every stage of their project. Bad ones don't.
There are several ways to collect observations. One is by recording employees while they are on duty. Unfortunately, this solution is potentially expensive, potentially illegal, and potentially disturbing to incumbents - especially if you are foolish enough to videotape them without their knowledge. If you have a large enough staff, you can have internal employees perform the observations...but if the observers are in the same department and hold similar jobs to the people being observed, this strategy will almost certainly taint the data. If you possess the necessary skills and if the job class is small enough, you can perform the observations yourself or have your supporting HR staff record observations (in theory, at least some of them should have the skills). Finally, there's the best solution: hire outside trained observers, communicate to your incumbents about their pending observations, and get out of their way. Not to suggest that consultants are the be-all end-all, but very few companies carry enough staff to handle this phase without outside assistance.
[The shortcut method is to have incumbents and managers simply write down the tasks they perform on the job, compile the lists, gather a diverse group of incumbents, and have that focus group validate the list you compiled. This method is fraught with peril, simply because it introduces all kinds of bias, assumptions, and potential faking into responses. I've seen more than one situation where this type of internal list-building resulted in a list of tasks which, in retrospect, didn't make any sense...and resulted in the implementation of invalid pre-employment selection tests that set the company back by decades. People tend to exaggerate their jobs, which is why direct observations are far more useful. Use this method only as a last resort – for instance, when direct or recorded observations are simply impossible due to the nature of the job. And be prepared to question any outcome that looks funny.]
To finish the observation phase, make sure you have a conversation with a few chosen job experts to double-check your observations for accuracy. Then you can…
6. Use the data from the observations/interviews to create your job analysis survey. There are many survey design options, and the sheer scope and time involved can be daunting. Job Analysis Interviews are best suited for situations where the job covers a very small employee population. They also come in handy when employees cannot be easily or legally observed while on duty. The downside is they can be time-consuming if the employee population is large. And good luck completing them if the employee population is spread out across an entire country, or several countries. Job Analysis Questionnaires are more commonly used. The can be either tailored to a specific job, or prefabricated to be generalized to a wide variety of jobs. Questionnaires are best suited for job populations containing many incumbents that are spread out across a wide area. Compared to interviews, they take more time and money to develop. If your employee population is small, you’d probably spend more time and money developing one of these than you would spend actually using it.
So which should you use? [CONTINUED IN PART 2]