Diversity & Inclusion? Start Hiring Based on Merit

0

Save time on candidates who look great on paper and less in vivo and recruit based on how they perform the task.

Ditching CVs as part of the the recruitment process is a much anticipated event and has been on many’s wish list for a long time. Regardless of the many solutions out there today, at this point, the CV is still king.
Even more advanced solutions in the recruitment process such as video interviews and DISC assessments are not an indication for a successful hire.
But, slowly slowly, we are getting there…and one of those solutions helping us get there is Vervoe. Finally we can save time on those candidates that look really great on paper, but not so good “in vivo”…

Hi Omer, what prompted you to start Vervoe?

Every single day wonderful people are excluded from the job application process for reasons other than whether they can actually do the job they applied for. Whether it’s their gender, their age, their degree (or lack thereof), their years of experience or something else on their résumé. Most of these things have very little to do with someone’s ability to do the job.

This is a problem I am very passionate about because I experienced it personally.

I grew up in Tel Aviv and, after high school, I served in the military and then worked at two of the hottest startups at the time. Then I moved to Melbourne and tried to get a job, but I couldn’t get an interview anywhere. Nobody valued my experience. Nobody could even pronounce my name. All that mattered was that I didn’t have a degree. Being screened out without being given a chance to prove what I could do felt wrong.   

Much later on, when I was leading a big team, I saw the same problem as a hiring manager. So did my co-founder, David Weinberg, who spent seven years with Juniper in the Bay Area. The common method of hiring – résumé screening followed by interviewing – favored people who look good on paper or are good at being interviewed. Meanwhile, capable, qualified and passionate people were being excluded.

The common method of hiring – résumé screening followed by interviewing – favored people who look good on paper or are good at being interviewed. Meanwhile, capable, qualified and passionate people were being excluded.

While we were discussing this problem we learned about the “auditions” Matt Mullenweg was doing at Automattic, the company that invented WordPress. Instead of résumés and interviews, Automattic invited candidates to work with the company on a trial basis. We immediately loved it.  

We wanted to create a way for companies to “audition” candidates using technology. At a fundamental level the focus is about seeing how people do work that is relevant to the job in the right context. But we wanted to make that possible at speed, at scale and based on data. We got to work and shortly after we quit our jobs and raised money. The rest is history.

Job Dashboard

Ok, so what does Vervoe actually do?

Vervoe is a skills assessment platform. It puts candidates in scenarios they would normally face on the job and asks them to complete job-related tasks. It could be doing calculations in Excel, editing a sales deck, writing code, designing a website or responding to a difficult customer.

Vervoe uses multiple question and answer formats, including text, multiple choice, video, imagery, audio, documents, spreadsheets and slides, code compilation, question banks and more.

Once candidates complete their skills assessments, Vervoe uses machine learning to automatically grade their responses. So instead of reviewing candidates in a random order and deciding who to interview based on each candidate’s background, recruiters and hiring managers can see a list of candidates ranked based on how well they performed. Their skills have already been validated.

How does the machine learning work?

Our machine learning models learn on three levels.

First, observe how candidates complete tasks. We collect many, many data points from how candidates interact with our platform. We’ve been doing this for a while at huge scale and it allowed us to create a learning set.

This means that all our skills assessments are instantly auttogradable. We don’t need to train our models on 100 employees from each company to generate a learning set.

Our skills assessments are instantly auttogradable. We don’t need to train our models on 100 employees from each company to generate a learning set.

This allows our clients to use assessments that are context-dependant and unique to them, and candidate responses can be automatically graded with a high degree of accuracy.  

But this is only the first level.

Then we look at the quality of candidate answers, comparing them to the “suggested answers” for that particular question, as well as how that question is graded by other companies.

Finally, we learn from each employer based on how they grade. We then recalibrate the scores based on their preferences.

Candidate card

Why go to all this trouble to make it possible to automatically grade responses to bespoke assessments? Why not let companies choose from a list of pre-prepared assessments?

We believe very strongly that everything should be context-dependant. A graphic designer at a series B startup is a fundamentally different job to a graphic designer at Amazon. They require different skills. So it doesn’t make sense to ask the same questions.  

Until now companies have had to choose between the following trade offs:

  • Traditional methods that are not highly predictive like phone screening and interviewing.
  • Very rudimentary testing that can be automated.
  • Tailored assessments companies need to grade themselves.

But we made it possible to have the best of both worlds: assessments that are completely tailored to each company’s requirements and are instantly autogradable.

Where do the assessments come from?

That’s the second place we use machine learning.

Some companies know exactly how they want to test a certain skills. However, most only know which skills they care about but don’t necessarily know the best way to test if someone possesses that skill.

So we built a library of skills assessments written by industry experts. Initially we let our customers choose an assessment that suited them. Then we made it possible to customize each assessment.

After a while we reached the conclusion that it needs to be even easier than that.

So now we simply ask recruiters to describe their perfect hire. They’ll type in something like “graphic designer who knows Sketch nad has a growth mindset”. We then use natural language processing to understand which skills they care about and automatically assemble a bespoke assessment.

So now we simply ask recruiters to describe their perfect hire. They’ll type in something like “graphic designer who knows Sketch nad has a growth mindset”. We then use natural language processing to understand which skills they care about and automatically assemble a bespoke assessment.

We have over 80,000 questions and tasks in our library and we have so much data about them. We know which questions are popular, which ones employers rely on most, which ones candidates like or dislike, how long each question takes to answer, and so on.

So we can assemble a near-perfect assessment that tests what a company cares about and is optimized for candidate adoption.

My Jobs

What roles does Vervoe typically test for?

The variety is huge, but at a high level we do a lot knowledge worker-type roles, so roles that require thinking like design, engineering, sales, marketing, customer support, finance, graduates and so on. We do far less blue collar roles, although we’re seeing more demand for those type of roles as well.

What kind of companies does Vervoe serve?

The two segments in which we’ve had the most traction are large companies and major staffing firms.

For companies hiring directly, we typically sell to the head of talent acquisition and we do much better with companies that have a dedicated talent acquisition function. That usually means companies with at least a few hundred employees, if not thousands.

Staffing firms have really surprised us. We thought they’d be late adopters and, as a result, didn’t focus on them early on. But we were wrong. We’re doing so much with staffing firms today.

For example, we work with Kforce, a US company that places candidates with Fortune 500 companies. We helped them effectively double their placement rates because, while their competitors are submitting résumés, they are sharing candidate cards with responses to tasks. So they’re putting forward candidates whose skills have already been validated, and their clients love it.

What kind of traction are you seeing now?

Since our commercial launch nearly 12 months ago, we’ve been growing at nearly 30% month on month. It’s been crazy. A number of things clicked at once.

We also learned a lot about ourselves and our key drivers of success. For example, we discovered that our most important success metric is activation. Early on ee weren’t good enough at getting companies to use the product straight away, but the ones that did loved it and always expanded. So we addressed that immediately.

So far, our conversion rate from pilot to expansion is 100%. Every company that has piloted with us has expanded.

What have been some of the biggest challenges?

Both David and I are industry outsiders, so early on we weren’t good enough at explaining our value proposition. It took us a little while to understand how the language of talent leaders.

The other thing is that we weren’t fully prepared to transact in the enterprise because we thought we’d get traction with smaller companies. We had to adjust really quickly to the things large companies care about like security, insurance and so on.

What’s the thing you’re most proud of?

Without a doubt it’s the voice we’ve been able to give to candidates who would otherwise be overlooked. Time and again our clients tell us “I just hired someone I normally would have screened out”. We survey every single candidate and read all their responses. It’s incredible. Our candidate Net Promoter Score is consistently above 60 and that’s because we give them an opportunity to showcase their talent, which is all they want.

Thank you, Omer, for your time!


Omer is the co-founder and CEO of Vervoe, an AI-powered skills assessment platform that helps companies hire the very best by focusing on who can do the job, not who looks good on paper. For additional info feel free to contact him directly: omer@vervoe.com.

Leave A Reply

Your email address will not be published.