Picking up where I left off at the end of Why Most Hiring Processes Suck, the primary philosophical difference between our hiring process and most traditional hiring processes is that we want to first assess a candidate’s aptitude at performing the skills required for the job prior to conducting any interviews or making any job offers.
For some jobs this is easier than others. Developers, for example, can be evaluated by their prior work and/or by giving them a small sample project to complete. A warehouse worker can be evaluated by having them come in and pack orders for a day. In both of those examples, despite being completely different jobs, you can test almost the exact working conditions relatively quickly. Customer service is much more difficult, because their prior work isn’t publicly available, and because it can take many weeks of training before they have the knowledge to answer even some of the most basic questions. The onboarding time for a customer service employee is likely going to be longer than for any other position we hire. It requires a vast knowledge of how our website works on the front-end and back-end, how we process and ship orders, and at least a basic knowledge of auto-detailing.
Thankfully, we’re not alone when it comes to this problem. 37signals has been discussing this problem on their blog, in their books, and in their podcast for a few years now, and I’ve been paying close attention. They’ve grown rapidly, and, like us, have realized that support is one of the few things that doesn’t scale nicely – it grows linearly as you grow. They also confirmed my aforementioned theory that it’s much harder to “try before you buy” when it comes to customer support. This podcast on hiring with Jason & David gives a great summary of their thoughts and experiences. Over time they’ve iterated the process to place a strong emphasis on writing samples. This is pretty obviously reflected when you take a look at one of their applications.
We used this as a starting point for our application process. You’ll see many of their fundamental ideas peppered throughout our hiring process. I can’t thank them enough for sharing their experiences publicly. It’s one thing to know that you should assess an applicants skillset, it’s another thing completely to know how to do it in a way that might actually work. They’ve hired more customer support in the past year than we may ever hire, so even with their small sample size they’ve been able to learn things and draw conclusions that we may never have on our own. Even a sentence as simple as “2-3 months on-ramp” in their application gives us a guideline for how long it should take to train someone. We thought it might take a few months, but it’s affirming to know that other companies aren’t doing it in 2 weeks.
37signals is also relatively small. Their support team is roughly 10 people. IBM’s hiring process, no matter how robust and accurate, probably never would work for us. Just like ours would probably never work for them. We needed some practical guidelines that would work for us, and they provided that, just as (hopefully) some of the information I provide here will help a few other small businesses.
Our goals with the application were: to provide a clear description of the job, to assess the candidates writing skills, to assess the candidates attention to detail, and to act as a natural filter to ensure that the applications that we did review were serious applicants. We’ve got work to do growing our company. We didn’t want to sift through hundreds of random resumes.
Our application is still posted up on our website for reference. Looking at the application, let’s break down those goals in to a bit more detail.
When it comes to providing a clear description of the job, we listed out, in order or importance, the responsibilities that this person would assume. In an effort to be as transparent as possible, we provided examples of customer service questions, links to our Wholesale, Affiliate, and Product Reviews programs, and specified that this person would have to help out in the warehouse packing orders. If you weren’t interested in all of those things, we wanted you to stop right there.
If that sounded interesting, we next created a “You…” section (borrowed from Milk, Inc.) to give candidate a detailed description of what we were looking for. There was definitely some hesitation with posting some of that info (what if they used that information to fake being the exact person we were looking for?) but ultimately we decided that honesty was the best policy and had faith that our rigorous process would expose anyone who wasn’t genuine. Not that it’s impossible for that to still happen, but we thought that the pros of being upfront more than balanced out the cons.
I thought that the “Qualifications” sentence “Bachelor’s degree in business, marketing, management, or communications, or relevant job experience is highly preferred.” was very important in communicating what type of company we are. Degrees are nice, experience is nice, but ultimately we’re not going to eliminate a candidate for not having them. “Highly preferred” doesn’t mean “required.” It’s about finding the best candidate for the job. Period.
The “Perks” section was intended to give the applicant a taste of why it’s awesome to work at Pure Adapt. We viewed selling ourselves as just as important the candidate selling themselves to us. We’re looking for a mutual fit where both parties are excited about working together. At every step of our process we devoted a substantial amount of effort in selling Pure Adapt. There’s a reason why we launched the redesigned website along with the job application. We wanted to put our best foot forward, from showcasing our design and development skills, to expressing in our writing why we’re a different employer than most.
To assess the candidates writing skills, we asked them to forgo the traditional cover letter and instead write a cover letter that explained three specific points unique to this job application. Each point required a decent amount of thought on their end. They probably never had to think of those things before, so we hoped that we were getting a completely unique and thoughtful response. The second point, “Why you want to work at Pure Adapt and not somewhere else” also required that the applicant actually spend a few minutes researching our company before writing.
The writing samples sure acted as a filter. If someone just sent a generic cover letter, or if they missed any of our required points, they were eliminated. If someone can’t come up with a decent reason why they want to work for us after reading our site and the application, we really don’t want to waste our time seriously considering them. There were also a lot of other built-in filters: we specified the email address to send the app to, the subject line, the due date, and to include both their resume and cover letter. If any of those things weren’t exactly perfect, they were eliminated.
Promoting the Listing
The most important filter of all, as we later found out, was requiring that applicants visit our website and email in their application. This was by design for sure, but we didn’t realize just how much of a deterrent that is for people. I’m over-generalizing here, but my impression is that most unemployed professionals take the “quantity over quality” approach that the 37signals guys mention in their podcast. By disabling the feature on our listings that allowed a one-click application we likely lost out on hundreds and hundreds of applicants. Our ratio of views to applicants on job boards was hundreds and hundreds to one. I’d imagine that ratio would have been much much lower had we made it easier to apply.
Again, this was by design. We weeded out the lazy people.
To promote the listing, we did four things:
- Posted on our blogs, our social media accounts, and obviously on the Pure Adapt website.
- Posted on the career center job search sites of all of the local colleges: Siena, St. Rose, SUNY Albany, RPI, Sage, and Union.
- Posted the job on LinkedIn.
- Posted the job on Craigslist.
Craigslist is obviously the most questionable. We did it because we were looking for someone local, it’s free, it was easy to just link people to the app on our site, and because other local business owners that we knew had good results with CL.
The only money that we spent to promote the listing was on LinkedIn. Job postings were 50% off during the month of October. So the $95 listing that we bought only cost $47.50. That should be a signal to anyone doing hiring that October isn’t the best month to find applicants. But it was what we had to work with, so we did the best we could.
Somewhat predictably based upon the difficulty of our application and the time of year, we didn’t receive a ton of applicants, and many of the ones we did receive were downright horrible and didn’t come close to meeting our criteria. However, in the end, we got what we wanted – a small group of talented, qualified candidates. In the next post I’ll explain how we went from the applications to making the hire.