At the end of last year I decided to run an experiment, applying for several hundred jobs on Seek to test which resumes and cover letters got the best response.
To do this I created a new generic resume under a different name and paid a couple of professionals to edit it for me – one on Fiverr (cost US$5) and one local consultant (cost $80). I then created versions of each using different fonts for testing purposes. It was supposed to be a quick holiday project, but the results were pretty interesting and as a result I ended up extending the project to gather more data, which I now want to share with you below.
(The results also prompted me to trial changing my gender and my name too to see if that affected response rate .. spoiler: yep, it certainly did. But those are stories for next time)
These were the success rates for 300 applications per combination:
1. Spending $5 got me 50% more interviews.
The Fiverr resume was actually the winner here and a pretty solid return on investment when you consider the extra interviews achieved. This one was definitely the most colourful / modern of the three and while it lacked detail compared to the heavy consultant edition, it got to the point quickly.
In a world where people spend about 5 seconds per document to decide your fate, this is important.
2. Despite the content being identical, font choice did make a difference.
Arial and Times New Roman were reasonably close overall, with no significant differences – though subjectively it seemed Times New Roman only really suited the long, bland consultant resume. Using Comic Sans on the other hand basically ruined any chance of getting to interview, which was definitely funny but also kind of strange.
You could argue that font choice maybe correlates with personality, or that choosing the same Arial 11pt as everyone else demonstrates an understanding of the rules and how to follow them. But really – let’s take a step back for a moment. It’s ridiculous we ask applicants to send us an essay as the first point of contact in this day and age in the first place, and even more so that we judge them on it like a 5th grade English assignment when the content often has little or no relation to the workplace.
3. When time is limited, we use shallow, cosmetic data to make hard decisions.
This short experiment was intended as a bit of fun but when you dig deeper it reflects a lot of potential problems. It’s amusing to laugh at the ineptness of Comic Sans, but when you start to consider other factors as potentially similar sources of unconscious bias in selecting people to interview, things suddenly get very interesting.
I’ll be posting my findings on changing my name, gender and race in the coming weeks.
If you’re searching for a job, the takeaway is these small details matter – regardless of how qualified you are for a position. If you’re a company looking to hire .. maybe have a think about how your method screens people out, and why. The fact is, a lot of the time we make these decisions based on shallow, cosmetic reasons which are unfair to all concerned.
Should you reward the candidate who pays a professional to design a catchy resume for them? Does it reflect good preparation and commitment or a bribe to get around the system?
It’s sad to see the optimism and excitement that should accompany a job application turn to disillusionment on all sides when confronted with bullshit systems. It’s a major reason we started Covocate, and the fuel that keeps us going today.
Question of the day: have you ever applied for a job you thought you were a perfect fit for and not even made it to interview? Why do you think it happened?
I changed my name and qualifications for the purpose of this experiment (otherwise they may see my Linkedin and compromise results)
Outside of font, nothing else changed: I used the same resumes and cover letters each time.
I stuck to general admin positions for this trial for consistency purposes.
When I got the calls to interview I declined based on accepting roles elsewhere so as not to unduly waste people’s time.
Yes, it took a fair amount of time to apply for all these. We have several clever systems that made the process of identifying and applying for these jobs far easier.. I couldn’t have collected this amount of data without them.
It’s true that it’s generally a good idea to customise docs per position and address the selection criteria where possible. However, this was the same for each trial and therefore shouldn’t compromise comparative results.
I haven’t shared the resumes or names I used as it’s potentially sensitive information for the companies who did / didn’t contact me for interviews.
Likewise, I won’t reveal which Fiverr seller or consultant I used for this project.
The picture has nothing to do with the story, I’m just sick of terrible stock photos. (Tabi @ sanjamatsuri in Asakusa 2015 if you’re wondering..!).