Workshape’s Hung Lee on the State of Hiring Developers
This is the sixth episode of HackerRank Radio, our podcast for engineering leaders interested in solving developers’ toughest problems today: Hiring the right developers. Hosted by Vivek Ravisankar (CEO & Co-founder, HackerRank). You can subscribe to us on iTunes and Google Play.
Hung, why do you guys keep calling me for jobs I don't want to do?
When Hung Lee's developer friend posed this question, he didn't have an answer. But this sparked a turn in the tide of Hung's career trajectory.
The reality is, this question hit the head on the nail for one of the biggest problems faced by in-demand talent. How do you separate signal from the noise that developers are constantly bombarded with? What kind of candidate experience is this, anyway, for our most sought-after talent?
Why don't developers express their interests up front before going through the ringer of phone screens...only to realize the opportunity is far from a good fit?
One question led to another, and pretty soon, Hung launched Workshape.io, a talent matching service that helps companies connect with talent based on interest as opposed to keywords on a resume.
Naturally, this philosophy on hiring developers fits perfectly with HackerRank's mission to match every developer based on their skills---instead of resumes, pedigree, or connections.
HackerRank's CEO Vivek Ravisankar sat down with Hung to discuss:
- The biggest challenges of hiring developers
- What the best companies are doing right
- What's in store for the future of tech hiring?
Listen to the episode, or scroll below to skim the transcript.
- The biggest problem with hiring developers today
- The fallacy of passive vs. active candidates
- If Elon Musk called
- Is there really a talent shortage?
- What is at the root of pedigree-first hiring?
- What the successful companies are doing today to hire developers?
Vivek: Welcome to HackerRank Radio, Episode 6. I'm Vivek, co-founder and CEO of HackerRank. Each month we'll talk to top engineering leaders about solving one of the toughest problems today - hiring the right developers. Today we have Hung Lee founder of Workshape.io, a matching service for tech talent that focuses on candidate’s interest first before unlocking the conversation. Welcome to the podcast Hung. Thank you so much for your time and jumping on this. This is our Episode 6.
Hung: I think we share interest in this area. We want to talk about how do we hire developers, what do we need to do, how do we create a situation where developers want to talk to us and all that kind of stuff.
Vivek: Awesome. Maybe you can talk a little bit about how you got started with Workshape.
Hung: Yes. Worshape started through a conversation. Myself and one of my developer friends, I was a recruiter at the time, so I knew him because I was head of talent. As his recruit a friend, I was basically responsible for all of his problems. And he would say, "Hung, why do you guys always keep calling me for jobs I don't want to do anymore?" And then I thought to myself, "You know what? That's a very interesting thing you said that because that's obviously the prevalent experience that a lot of in-demand talent actually has, where they get constantly bombarded by recruiting messaging from recruiters pitching jobs that they don't want to do.
Workshape is really an attempt to cut the noise in that sort of initial part of that conversation. What we wanted to do at Workshape was to collect information about what a developer inherently wants, we visualize it to make it very easy for the recruiter to understand what that looks like, and then we match it to companies that have a compatible job. By doing that we hope that we can reduce the number of irrelevant conversations and of course increase the relevancy of the conversations that do occur.
Vivek: That's very interesting. So I did see the graph on your homepage - the graph that had architecture, front end skills, documentation, code quality, and all of those things. It was a really cool graph. I'm probably going to use that for a developer report as well. But I thought that was more to do with the skills of the developer, but it seems like it's more to indicate the developer’s interests.
Hung: That's exactly right.
Vivek: And so how does a developer indicate interest? Do you say that I am really, really interested on a scale of 1 to 10 in this area? Is that how a developer indicate interest on your site?
Hung: Yeah. Another thing that we wanted to do is to make it very easy for the developer to communicate what it is he or she is most passionate to do. And so we experimented with numbers of ways to do it. We just said, "Look, we cannot have people fill forms in anymore. This developer is not an active job seeker in the usual case, he's not going to tolerate sort of UX. It's going to be extremely difficult."
So we give them a very simple drag and drop interface to say, "Look, if you had the ideal scenario, what does it look like in terms of your time distribution?" So we think that Workshape is really a visual description of time distributed over tasks. So let's say I come into the system, and you know, you're a developer, of course, you are, let's say you hate to do documentation. I mean that's an honest thing. You don't like doing it. You really don't want to job that requires you to do it.
Then great, tell us that you don't want to do it there then and we will then match you to a company that does not require you to do that part of the job. So we want to match developer sentiment to employer need with a bit of honesty to say, "You know what? No job is ideal. No candidate is ideal. There's compromises that you need to make. But at the same time, if you tell us what it is that you're most interested in, we will only match you to companies that have that job. We won't expose you to the wider market where you end up having to deal with pitches."
The recruiters, of course, who are very talented salespeople often have the ability to persuade you to take an interest in a job that perhaps is not inherently in your interest. Before you know it, you find yourself in a recruiting process that actually you don't really want that kind of work. So we say, "Okay, forget about that. We'll only expose jobs that you've told us you're interested in."
Vivek: Maybe you can walk us through, or audience a little bit about when did you get started and sort of your traction so far in terms of the number of developers who've signed up, the profiles and the number of companies and also possibly the number of matches that you have been able to make.
Hung: One of our problems was that we wanted to create a kind of an opt-in service. We never wanted to create a massive kind of marketplace that scrape data and all the rest of it. We wanted everyone kind of come into this with deliberation and conscientiousness. So we always had this massive skill problem. Like how do we encourage developers to sign up to job discovery site when in fact obviously a software developer might not feel that they need this? They have a lot of opportunities coming at them all the time. So massive challenge for us.
But the way we positioned Workshape was to say, "You know what? We understand that developers generally don't behave like job seekers. They don't exhibit job seeking behaviors, in my language. So they don't have a resume to go, they're not going to apply for jobs and not going to do all of these things that a traditional job seeker might do."
So our pitch should then be to say, "Look, we accept that you currently probably in a great job, but if you're able to tell us what is ideal, then we'll collect that information and we'll just connect you to a company, but when that opportunity arises. It may not be immediate. It could be next week. It could be six months down the line. But if you do get contacted on Workshape, it will be because that job has matched what it is that you said that you wanted to do."
So instead of thinking of the dynamics of active versus passive, this is a very common question that we...I'm sure you get this as well. "Oh, Vivek, how many active candidates and how many...?" And we're saying, "You know what? This binary model of understanding people's behavior is probably not very accurate.
Vivek: Yeah. Because at some level everybody is sort of looking for a job, right?
Hung: Yeah, absolutely. It's spectrum.
Vivek: Yeah. I don't think I am interviewing, but yes, I'm pretty sure like most people in our company, I mean it's not that they don't like HackerRank, but it's...And I actually encourage people as well not very openly in all hands session or something, but "Hey, you should go and look out and see what's open but don't take the offer."
But at some level, I don't think it's binary like you said. There's always at some level people are always contacting others and you're always taking up like some sort of a coffee meeting or a lunch meeting or anything. And it's happening, so it's really not a binary one.
Hung: Right. If you and I cast our minds, I mean we kind of move beyond the period of actually thinking of job seekers, but let's say you remember there was once upon a time when you're an employee and then you think about the amount of time you're actually doing job-seeking behavior. And even when you were active, it was probably no more than 10, 20 hours a year.
What are you doing the rest of that time? Are you getting on with your day? But it doesn't mean that you're completely a nut. Even now you probably not a nut, Vivek. If Elon Musk came to you and call you up tomorrow and said, "Hey, Vivek, I love what you do in HackerRank, but can you help me send people to Mars?" and offered amazing opportunity, you'll probably take the call at least.
Vivek: Yea, I'll probably take the call, but I'm not sure if I'll go. But yes, I'll definitely take the call.
Hung: But this is exactly what I mean is if is the job is close to ideal, I think most people would have a conversation. And if you're an employer, you know what? You would not stand in the way of your staff if that circumstance occurred.
Again, it wasn't a strategy at Workshape to think like this. We didn't sort of have this insight right from the beginning, but as we were kind of building an aspirational matching product, we realized that actually most people are probably interested in a conversation and they want to keep the dialogue with the right kind of company, right kind of job opportunity, and they can be activated when the time is right. I think that is actually something that's the wider market really need to get a little bit better at.
Vivek: Yeah. That's what we discussed at some level. People are always, not always I would say, but at some spectrum, at some degree, people are looking for another opportunity. I would just at least like take a meeting and see how it goes. Because I do that all the time on the other side where I go and ask them "Hey, can we meet for coffee? Can we meet for lunch and do this?" So clearly it's going to happen on the other side as well.
Hung: You know what? Let me just add a final point to this. The real mistakes I think companies make when they're trying to recruit software engineers or in fact, anybody in the high demand with this kind of limited supply is that they're only using this binary model of active or passive. And they're either focusing purely on active candidates so they immediately eliminate 99% of the possible people they could talk to and they're just interacting with those active guys who of course are also interacting with everyone else.
So it's a complete race to close these people and it's a chaotic, high-stress types scenario for everyone involved. And they end up implementing a process that is just a way to push these people through the flow. When, in fact, if you started talking to some of the 99%, you'd realize that actually there's lots of people there that you could potentially recruit if you had the right approach to it.
It's actually very similar also to sales. I mean, of course, you know this now yourself Vivek that basically...Imagine buyers of HackerRank as a service. If you focus only on the active buyers that are obviously signaling they'll buy at HackerRank, then you'll have a very small market.
But of course, you don't do that. You do a lot of sales nurturing, you do a lot of marketing, you do a lot of inbound stuff because you want to build the brand. Eventually, a buyer who's going to buy a service, their timing is going to be right for that conversation to occur. I think recruiters need to work exactly in that way, especially for the high demand folks.
We can't treat developers as if they're graduates who are desperate for work or highly motivated job seekers. They not highly motivated job seekers. They're highly demand, highly skilled people who are busy getting on with their day. You've got to figure out a way how to talk to those folks.
Vivek: Yeah. What is your thought on this whole idea of...there are so many different articles that talk about there is a lack of supply, which is we don't have enough developers right now and hence we are struggling to hire. Is it because...okay, I don't want to bias, I don't want to give an option. What is your thought on that?
Hung: I've got a number of thoughts on it. Actually, number one that probably is an absolute imbalance that the very highest level there probably is a kind of a shortage of talent, I would say. Although that is not the main reason why people struggle to recruit.
I think the second reason is because people are focused only on the hyperactive. And the hyperactive who are in the software developer community is a very small number of people. They are active only for a very short period of time before they are off the market. So we're ignoring a lot percentage to be available resource pool. But then there's other reasons.
In recruitment in general, we have a number of biases which kind of reduces the available talent pool, especially in tech, we do a lot of what I call pedigree hiring - where this guy is got to be good because he worked at Facebook or he can't be good because he didn't work at Facebook. Stuff like this.
Then you think, "Right, firstly, how do you know this? Secondly, you are kind of saying this because perhaps you don't trust your ability to assess whether this person is good or bad, therefore you push it to the previous employer to say, 'look, Facebook must have put this guy through X, Y Testing, therefore, you must be amazing.'" It's kind of not trusting your ability to make a decision.
This is where I think your product or products like it have such a wonderful contribution to make to the decision making because you can help recruiters and hiring managers actually. It's not just the people who are doing the recruiting role. It's usually the hiring managers that have the pedigree, hiring bias. It's helping those folks to potentially say yes to candidates who are perfectly capable but perhaps don't have a big brand name that they recognize.
Vivek: And have you seen it happen at Workshape - which is like a developer who probably did not graduate from a great school or worked at one of these well-known companies actually ended up being hired and being very, very successful? Have you had stories like that at Workshape so far?
Hung: Absolutely. And in fact, again, we've accidentally had this insight because we didn't discover it this way. But discovery is important. So when pedigree hiring is an issue is when basically the employer has the ability to search the database.
Now, when you search a big database, of course, I'm going to see all the big names I like and automatically give them an A rating even though it just is a resume. How would I know this person is a good developer? I have no clue. You know what? He worked for a company I recognize therefore he must be good.
On our platform, there's no such. Immediately that's eliminated. Discovery is purely based on the shape of the match and some other inputs that we collect on either side. So we've provided a way for I think the 99% of brands to compete on an equal playing field with the mega brands out there because developers firstly will discover those opportunities in that way and companies will discover developers in a completely new way without thinking about this pedigree.
Yes, it's organized that way and we do see conversations occur that I think would not have occurred had that initial kind of search be met, which kind of lends itself to this sort of bias.
Vivek: Yeah, that totally makes sense. The reason why we even started this company was very similar to that, which is resumes have a very poor signal to your skills. I think it's there even on your home page. "No resumes required," or something on those lines on resumes. And what we have does a very poor signal and we focus a lot on skills to do it.
Hung: Let me say this. It's crazy. It's crazy how much dependency the entire recruitment market has on resumes. If you think about all of the systems that are there literally to process these documents and all of the processes, human processes that we put in because of the existence of these processes, it's like, oh my God. We literally have built an entire way of doing it based on what is essentially marketing literature. How can we expect to do a good job of matching people to opportunity when the artefacts that we use to exchange information cannot be trusted? It's a mess.
Vivek: It is like a really, really, really hard problem, which is that in a very idealistic way you would have a standardized system where you would quantify if you are level eight, it means this, if you're a level seven, it means this, if you're level six, it means this for every skill. And the entire world agrees to that because there's no standard. And then for each developer in the world, you could sort of put them in these buckets or in the skill levels and then match them to the right jobs, which is very similar to what you're saying.
Not every job requires you to have all the skills on 10 out of 10. There are certain strengths that you want in a particular developer for a particular type of job. And as long as you're able to identify that right skill sets and in the right proportion, then you should be able to do it well.
I think that's what I see a lot of companies, whether it's Workshape or HackerRank, there are a lot of other people who are trying to solve the recruiting problem as well, moving in the direction of "Can we get better signals?" That's really the goal.
If resumes are a poor signal, whether that's collecting data from developers to say, "hey, what do you really want?" or whether you can actually showcase your skills, or whether you can have people interview them saying, "Here are this person skills and strengths and weaknesses," then that's a very interesting shift. Although I believe that it's going to take a much longer time to do that.
Still, you find that job boards...and I don't have anything against job boards. Couple of them are investors as well so I love them. You still have people on job boards and resumes and connections and others still occupying the majority of how you do the searches. Right?
Hung: Yeah, and that's okay. I mean, sort of the way I've seen it is, perfect is not achievable and probably not desirable either. There's always going to be some messiness in connecting people to opportunities. If you like the dream of creating the perfect system, I think it'd actually be a dystopian thing. What we need to do as entrepreneurs and as people are trying to push the model forward a little bit is just to provide a little bit more information, a little bit better signal.
At the end of the day, I think recruiting especially when you're talking about permanent hiring, getting someone into the team, that remains I think a very contentious intuition. I would say that it'd be very difficult to objectively measure someone suitability without also factoring in members of the team, the project they're working on, the context, the environment. All of those things contribute to whether this person is good hire or not.
What we can do technology vendors is to help people make better decisions. That I think is good enough. I think you said no one trust resumes. That's the reason why a lot of recruiting is so painful because we don't trust the damn resume.
That's why we want this guy, you know, he needs to do a six-hour technical test. Actually, we're not happy with that, he needs to then do whiteboard exercise and meet with the CTO. We are not happy with that, we need to put him through a reference checking process. Still, we're not happy with that. We hired a guy and we have some sort of notice period of four weeks because we still don't trust any of this.
So all of these things is because we don't have confidence that this person is suitable. And the way I see it is, look, if we can just provide a little bit more than what is currently out there that's pushing the model forward, that's going to help their decisions be made and there will be incremental progress as a result.
Vivek: When have you seen this shift most pronounced where there's transforming from a resume based evaluation to a more skilled based one in your experience at Workshape?
Hung: There are tech companies that are doing it. I have to say that I think that how they've innovated is almost as a result of necessity also, but they've recognized that "okay, I need to hire a scale, I need to go from 50 engineers to 500 engineers." You know what? You cannot do that with the traditional recruiting pipeline based on applicants. And applicants typically you apply with a resume.
So instead they've started to reject resume completely and they've started to do much more event-driven type recruitment. They're pushing the assessments further back into their throat. So if you think about the recruitment pipeline as it is, the resume is really designed and so you should really make the system based on this resume at the beginning - more or less at the beginning of the pipeline.
The smart companies are recognizing because resumes are such a bad signal, you cannot use it in this way. What you can do is at least understand this person is a developer, get him involved in the pipeline in some way, and then you have a different way to assess as you go through.
Now, this could be a technical solution you guys have got. It could simply be, hey, listen, we get this guy in the room a little bit, we get him in a meeting, we do hackathon, or we do something which basically has him interact with our technical team. And then through more organic dialogue, you'll be able to surface out a few things that at least say, "well, this person's out or possible." And then you take him further into the process as you go that way.
So I think companies that I've been impressed with are the ones that understand that you can't process software developers and assume you can do that at scale. You actually need to have an iterative approach to recruiting them. So in some respects, I think crew does need to learn a little bit from how software engineers operate so that the modus operandi of recruiter is typically a waterfall process. Recruiting is typically waterfall. And yet developers, of course, has been agile now for a long time.
I think the way to recruit developers at scale is to adopt the agile method or at least the agile principles. And I don't mean sort of doing standups and doing the sort of symbolic stuff. I mean, literally reengineering the candidate pipeline, reengineering your workflow so that you're not processing in the same way.
Like I said, there's companies that have aggressively done this. Zalando is a good example. I think you guys use it maybe or they use you, but 2,000 engineers in 12 months, okay, you know what? You can't pipe these guys to do it this way. You need to have a totally different way to do it. And typically you modularize your recruiting pipeline and you start parallel processing parts that used to be sequential, before you know it you're starting to do things in like an object-oriented way, right? You're starting to recruit a lot more as you might do software development. Those are the companies and those are the examples that I think are going to win that particular battle.
Vivek: Thank you so much. For the listeners, if you have any specific questions or topics, tweet to @HackerRank. Thank you so much.