HRExaminer Radio

HRExaminer Radio is a weekly show devoted to Recruiting and Recruiting Technology airing live on Friday’s at 11AM Pacific

HRExaminer Radio

Guest: Dominic Barton, COO, North America & Asia Pacific, Broadbean
Episode: 150
Air Date: February 22, 2016


Dominic Barton is the Chief Operating Officer of North America & Asia-Pacific at Broadbean Technology.


Audio MP3





Begin transcript

John Sumser: Good afternoon, and welcome to HR Examiner Radio. I’m your host, John Sumser, and we are coming to you live from beautiful downtown Occidental, California. It’s five miles over the mountains to the ocean, five miles north to the Russian River, and the downtown has two Italian restaurants and a pub. I’ve got to tell you everything’s green here. It looks like Ireland. Today, we’re going to be talking with Dominic Barton, who is the COO for North America and Asian-Pacific of Broadbean.

Broadbean is an amazing company that began its life doing job posting distribution, and has become a recruiting technology process monitoring and sourcing automation company. It’s an extraordinary story. We’ve talked previously with Kelly Robinson, their CEO, and Dominic Barton is the genius behind their recruiting process control product called BDAS. How are you, Dominic?


Dominic Barton: I’m well, John. Very well indeed, living the dream.


John Sumser: Living the dream. Well, yeah. You are in Orange County, where it never rains and all of the people are beautiful. It must be quite a big change from England.


Dominic Barton: Well, I was originally born and raised in Dundee, Scotland, so eighteen years there, then twenty years in London, England, and came to California in 2010. So yeah, I saw a lot of rain before I came to the sun, that’s for sure.


John Sumser: Yeah, that’s like me and snow. I had my share of snow. I remember the last shovel of snow that I shoveled, and I headed to California never to see the stuff again. It’s a place for weather refugees, I think. Tell me about how you ended up being the COO of a company in California that does recruitment advertising distribution and process control for recruiting. How did that happen?


Dominic Barton: Gosh. Well, I started off working in banking, then tech, e-commerce, media, utilities, went on to online dating, but before coming to recruitment. I also spent the first ten years of my career working in divisions of large organizations, and then the last ten working with entrepreneurs and small companies backed by large organizations. I have to say I prefer the latter. I like working with entrepreneurs.

Back in 2005, I joined the MGT, which is a big media company in the UK, and started to do a lot of work with turnarounds with some of their online media businesses. I did that for about five years, and then I said to my boss, “Look, I think I’ve earned my spurs. Can I have a good news story please?” He said, “Well, we have this fantastic little business that’s just started in Newport Beach. It’s got a massive market share in the UK, but it’s only just started in America. How would you like to go over to Southern California to spend time working with Kelly Robinson to try and grow Broadbean?”

Now in the last role, I’d spent a bunch of time in central Ukraine outsourcing our technology there. From one minute central Ukraine, to the next Southern California. You never can tell what’s coming down the line. Very excited. We came over here for a six month secondment. That was myself, my wife, my dog, and my son, and we’ve been here ever since. Five years.


John Sumser: That’s an amazing story. Now, you really took to this business. You were a turnaround guy, and so you were really good at finding the flaws in the operating process in organizations, and I believe that your strength has been coming in, rationalizing the operations, finding those weak spots, and trimming things down until the operation could make money. Is that fair?


Dominic Barton: That’s a piece of it. A bigger piece of it I think is more about making sure you’ve got alignment in your teams. I remember I had a boss once, a guy called Stephen Jones whose focus was on execution. He was a great believer that a well executed lesser strategy will beat a poorly executed great strategy, and that the key to delivering any strategy is to have everyone aligned, on board, and pulling in the same direction, and that internal people who are not bought in will kill a strategy way more effectively than a competitor ever could.

A lot of that operational piece that I see is about having clarity of purpose and making sure everybody who’s on the team is aligned and pulling in the same direction. It sounds simple, but when you look at many failed organizations, that lack of internal alignment and people all pulling the same direction is often the issue. That’s been what I’ve seen.


John Sumser: Well, now that’s something that takes a lot of focus and a lot of attention, a good memory, and a clear sense of where you’re headed, because figuring out who’s aligned with what and the difference between what they say and what they do is beyond a lot of people. It’s really beyond a lot of people. Where did you learn how to do that?


Dominic Barton: Stephen was key in that. He taught me that. I also worked with a guy called Toby Strauss, an ex-McKinsey guy. He really did teach me the, if you like, the discipline and the intellectual rigor that McKinsey bring to business. I learned an awful lot of that from working with people like Toby Strauss, I would say. He’s was an amazing guy. You’d go in for meetings with him, and he had a sign on his wall behind his desk with that … Do you know that George Bernard Shaw quote about the unreasonable man?


John Sumser: I don’t.


Dominic Barton: He basically said … You’d go in for a meeting and above his desk was this sign that said, “The reasonable man adapts himself to the world. The unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man.” That was the context of every meeting. So, there was no counsel’s a bit unreasonable Toby. The sign was there. I guess he taught me as well it was quite important to be unreasonable sometimes.


John Sumser: Well that’s good. The hardest part about being a great manager is that sometimes you have to enforce rigid lines, and that’s … I’ve seen you in action, and what I know about watching you work is that you are able to enforce rigid lines without losing everybody in the process. When I do something like that, I might cause a few more casualties than you do because I’m less disciplined about how to deliver the message of the unreasonable request.


Dominic Barton: I think also there has to be a precursor to that, which is everybody figuring out the right answer together. Once you’ve figured it and once it’s decided and once it’s locked in, that’s when you can’t let anybody out. Everybody’s got to stay on the same track until it’s done. That’s for me, but it doesn’t mean there’s an absence of figuring out what the right answer is. There’s definitely a period for that, but it can’t be an iterative constant process. Once you’ve locked down what you’re going to do, then you’ve got to have a good bash at executing on it without anyone saying, “I’m just not too sure about this.” That just takes the energy from it, takes the focus from it. You’ve got to lock down and go, is my sense on that.


John Sumser: Well, that’s pretty interesting. I think that works really well. It also seems like it is the opposite of a lot of current software development practice. I’m of the impression iteration is how most software’s developed these days, and so there’s an interesting tension between those two things. You have spearheaded the effort at Broadbean to build something that I think you’re still calling BDAS, which is a recruitment optimization visualization and data management tool. I don’t think anybody will understand what that meant, so will you mind telling me what BDAS is?


Dominic Barton: Yeah. Basically we were talking to our clients about analytics, and the clients consistently had the same three problems, at least we came across they kept having these same three problems. The first issue is that everyone wanted to sell them analytics but they only wanted to sell them analytics in their particular part of the recruitment value chain. Recruiters didn’t really need any more insight into any one part of the candidate journey. What they needed was insight across the entire end-to-end candidate journey. But, not many people wanted to provide that, because it’s really hard, and the reason it’s hard is because you have to work with other peoples’ database schemas and other peoples’ file formats, which is not the easiest thing in the world.

Now, the nicest thing for Broadbean as you know is we post over I think two and a half million jobs a month to six thousand different job board feeds in over a hundred countries. We see a data feed from a ATS that looks remarkably straightforward and simple to us compared to some of the data feeds we have to normalize in recruitment over the last ten years for some of the job boards that we’ve worked with.

It turns out then in the last ten years we’d spent a whole bunch of time building up this expertise about normalizing recruitment data feeds without realizing that’s what we were doing, when we were just building job board feeds to post jobs. That was the first thing that our product was trying to do, which is to say, okay, if you have data in multiple places, can we pull that data together into a single holistic view of all the data? That seemed to be a real problem for clients both in terms of could we do it and how much effort was it taking them every month to do it, to pull the data from all the systems together.

The second problem people seem to have is that the sheer volume, complexity, and diversity of the data meant that a lot of the clients struggled to either find the story in the data, agree that story with their fellow stakeholders, or to share that consistently throughout the organization. Again, that was the second thing that we were working on: How do we find these great stories in recruitment data, and how do we help people tell them? We’re very blessed to work with some great people like Chris Hoyt, who basically shared with us the stories he was trying to tell within his business, and then we helped in effect systemize those stories to help everybody tell the stories as well as the best people in the business. We try to make it easier to tell those great stories.

The third issue that everyone seemed to have with their data was that there was always somebody in the organization who would point at their data and they would say, “That’s not right.” Of course, as soon as you undermine one part of your data suite you’ve undermined the confidence in the whole suite, and your ability to achieve change based on data has just gone down to zero. We saw that there was a need to have I guess what we call a forensic approach to record accuracy, which basically means that you get very granular up front about business rules, about business definitions, so that there’s never a circumstance where somebody points to a number and says, “That’s not right.”

Typically, the reason why that number’s not right isn’t because the machine can’t count, it’s because different divisions or different stakeholders are using different business rules to underpin the data that they’re actually looking at. Then you get into these awful situations where you have a one hour meeting on a particular topic, fifty minutes of which is spent debating what the right numbers are, and the last ten minutes of which is spent figuring out what we’re actually going to do about it, yeah? What we’re trying to do is turn that on its head whereby regardless what you’re perspective is on an issue, whatever data you need to address that issue, you’ve got it and you’ve talked it through within ten minutes, which means the next fifty minutes is actually about figuring out what you’re going to do next.

The product’s basically developed to solve those three problems. Single holistic view of moddable sources of data, consistent story throughout the organization, and then just this forensic approach to record accuracy so that you’re just never, ever wrong when you put a number up on the screen. That seemed to be a problem, or a set of problems, a lot of clients were having to the point where funny enough our best clients are the ones who’ve tried to do this stuff themselves first and then buy the product, more so the ones who are just getting used to analytics. Some of these things I’m talking about, they resonate a lot more if you’ve been through an analytics project, much more so than if you’re just about to start one for the first time, I think.


John Sumser: Are you saying that the best way to become a great analytics operation is to screw up the first project?


Dominic Barton: You know, I think Chris Hoyt told that story best, which is the one where he puts the video of, I think it’s A Few Good Men with Jack Nicholson saying, “You want the truth? You can’t handle the truth!” You just change that out for, “You want the data? You can’t handle the data!” His experience was that if you can put an awful lot of time into getting people all the data that they say they want, but the reality is they really don’t want to spend time looking at that data, and the adoption of that heavily data-driven reports is actually pretty low in the organization. What people really want is the story in the data, and the story is often buried. It takes a while to actually boil all the data out before you realize that actually people aren’t interested in the reports and the data. What they really want is the story that underpins it.


John Sumser: That’s interesting. I think the piece that’s missing from your description of the product is the remarkable level of attention to detail that you’ve put in to the design of the graphic interface. In my view, when you use the product, the graphic interface gives you the ability to dive into granular data that’s always right, and you don’t have to take a linear narrative through the data. You can ask questions in the way that a review team asks questions in a staff meeting, which is one idea prompts another idea, and it doesn’t go in a straightforward, logical way. Would you mind talking a little bit about the process of developing the graphical interface?


Dominic Barton: Yeah. What that comes down is what we call conversational analytics. What we mean by conversational analytics is that if an issue is worthy of serious consideration by an organization, then it probably isn’t going to be decided by a single person in an ivory tower. What mostly happens is those big decisions are a negotiated settlement between stakeholders with different agendas. What you can’t do in that environment is design your analytics to be consumed on your own. You have to design them to be consumed as part of a conversation. To refer to your staff meeting, what we do is we go and listen to people having these conversations, and we’ll say to someone, “Okay. What is the topic that you guys are trying to get to the bottom of?” Have that whole conversation as if you had perfect data. We listen to people having that conversation as if they had perfect data, we go away, we come up with the visualizations, and we come back with the perfect way of visualizing that particular conversation.

Now, the thing about conversation, to your point, is that conversations have a first level question, a second level question, and sometimes a third level question. Your analytics needs to allow you to very quickly get to the second and the third level questions. Now, you typically don’t get a fourth. Usually people don’t have the time to get that level of detail, but you need to be able to handle the first challenge and the second challenge that come with these things. Again, we design these reports around very specific conversations.

Now, the second piece of that is the biggest barrier, clarity on these stories, is information density. What you have to do is to be able to take out the bits that genuinely aren’t going to move a lever. The challenge process we put on clients is to say, “Okay, I know you find this an interesting piece of data, but if that number was higher than you wanted it to be, what would you do to change it?” It sounds simple, but there’s a number of pieces of data there where there is no plan. There is no lever that you would pull to actually change it. It’s just interesting, and those data points don’t get in the reports.

Now, what that allows you to do, if you can be that disciplined whereby the only data points that go in your reports are the ones where there is a demonstrable lever you can pull to change behavior or help make a decision, if you can do that then it basically frees up a whole pile of real estate to make it easier to tell the story. Again, if you’ve got a report that’s only being used for one particular conversation, one particular purpose, you don’t have to put a bunch of other pieces of information so you can use that same report for another purpose. Now, what that allows us to do is to create bespoke visualizations that are right for that particular conversation.

What we’re not talking about is an ad-hoc reporting tool whereby you can see how any two measures correlate or that type of thing. It’s very specifically, “Okay. What does the conversation between a hiring manager and a recruiter look like?” If they met once a week, what would they talk about? What would be the start to the end of their entire conversation? Or, if those guys are going to meet to say this is a particularly hard to hire job, how would we prove something was a hard to hire job? What data points would they need? How could we figure out what levers you could pull to change behavior on a hard to hire job? Does that kind of bring it to life a little bit?


John Sumser: A little bit. A little bit. I think what is fascinating for me is the way that the visualization helps examine the detail at every level of the process. I’m going to try to explain a little bit of what it looks like. It looks like a rainbow colored funnel that shows each step of the recruiting process with statistics about each phase of the recruiting process as the funnel narrows to the bottom, and then each one of those bands is clickable to open up the second layer of the conversation, which is what’s the overall performance inside of that band, and then the third layer underneath that, which is an examination of transactions or sources. What you get is an astonishing way of accessing all of the data in the recruiting system to explain performance and demonstrate the relationship between the transactional data about performance and the achievement of the goals in each of the segments of the funnel.

It’s such an enormous undertaking to try to map all of those variables in to a visual story that I think when you see this thing, it is stunning in its simplicity. It doesn’t look nearly as hard to understand in the flesh as it does in conversation. It’s a very interesting paradox. This picture really is worth thousands of words, and it’s an astonishing and beautiful thing to see. I’m your biggest fan here.


Dominic Barton: We’re proud of those visualizations, but really all they do is capture the existing processes which some of the best people in the industry already use to describe particular events or particular scenarios. We’ve been very lucky in being able to work with some of the best people in the industry to actually figure out how do they talk about the issues. We don’t sit in an ivory tower and try to think, “What’s the best way to describe these issues?” We literally listen to the real best practice practitioners that people travel to go and see at conferences. We listen to them talk about these issues, and then we figure out how to put a visualization that matches their story. But it’s their story, and we’re just trying to visualize it for them.


John Sumser: All I can say is you’ve hit the state of the art, and it’s a thing of beauty. Now you’re sitting in a world where you can effectively take all sorts of data sources and make the recruiting process easier to manage and more effective. At the same time, the recruiting industry is notorious for its rapid evolution at the front ends and its rapid adoption of new toys and tools and techniques. You must have an interesting view, because you have to be considering integrating it into your tool, of what’s coming and what’s it mean. What are you seeing out there?


Dominic Barton: I think the overall theme is one of convergence. I think the people that we talk to don’t really want their analytics product sitting in a place called analytics, their ATS in a separate place, their career site in a separate place. You’re seeing a convergence of all these pieces coming together whereby the analytics feeds the workflow tools, and the workflow tools have got analytics actually built into them so you’re getting recommendations already coming through. Before you post the job, the analytics is coming in and saying, “Did you know you already have this number of candidates who’ve expressed an interest for this particular type of job already waiting in your resume database?” And our semantic searches said, “Here’s the top ten of them who are most likely going to be candidates.” That whole convergence … so yes, there’s so many companies in the recruitment space looking at all the different aspects that you see in recruitment. More and more for us, demand seems to be, bring all this together to a single workflow, to a single way of our recruiters engaging with all the benefit of all these systems without us having to relearn six different systems. That integration and convergence I think is the theme that we’re seeing more than anything else.

As much as anything, I think that recruiters and their recruitment functions, they all had a full time job before they decided to start implementing software. The life for the recruiters is much more about, “Give me one neck to choke. If all things go wrong, I want one place to go where I can squeeze if this isn’t working.” What I don’t want as a recruiter is to have seven different systems everywhere and to be trying to figure out where the handshake between these seven systems isn’t working. There’s a lot of innovation out there, but one of the themes we’re seeing is, “Okay, how do you give me a dashboard to pull this together?” Then, “How do we pull the learning from say market data systems, how do we take that and stick it right alongside our ATS data in the workflow tools?” If that makes sense.


John Sumser: It absolutely does. That’s exactly what I see coming along as well and that … What I find interesting is so far everybody that I’ve talked to, everywhere I’ve looked, the idea is that an integrated system involves software that does integration. As I understand what you guys have done, you’ve just said, “Let’s just integrate it at the dashboard level and make the data make sense at the dashboard level and not have a particular opinion about which system is better and which system is worse. We just take the data, and make the data tell the story.” Is that a fair …


Dominic Barton: It is. It’s about being data agnostic, but I think if you step back a little bit, normally the way good companies work is you start with a strategy. Once you have a strategy then you say, “Okay, what’s our key performance indicators to show progress again for that?” You put in place your key performance indicators, and then you have your workflow tools that help you execute on that strategy. That’s a very solid way of doing it.

But what we find a lot of times in recruitment is it starts with a very large, considerable spend on some workflow tools, then on top of that we’ll put some analytics, and out of it comes a strategy. But it starts with the workflow tools. What we’re trying to do is to turn that around. Basically, you don’t have a lot of choice. It’s very difficult to do it the other way because every time you unplug a workflow system and plug in a different one like changing ATS, the reporting that you get is completely different. The workflow that you get is completely different. The integration points are completely different. You don’t have a lot of choice.

What we’re trying to do is say, “Okay, let’s put in place an analytics platform at least that is completely independent of the workflow tools underneath it.” If you pull out one ATS and put in another ATS, just as an example, it shouldn’t change the key performance indicators of the business, and the people in the business shouldn’t even know from an analytics perspective that one ATS has gone out and another one’s been plugged in. You can only do that if you’re completely data agnostic and you can take any data feed and map it on to whatever business rules the client wants. I think more and more, that’s where this piece is going, same with the workflow tools, to be fair, if you start plugging in some of the more niche workflow tools to your overall platform for recruiter workflow.


John Sumser: This has been an amazing conversation. We’ve whipped through the time, and it’s been deep and informative. If there was one thing that you wanted somebody listening to this to take away from your story and the work that you’ve done, what would that be?


Dominic Barton: I think it’s probably something around the quality of peoples’ data today. It’s probably not as good as you think, and you’re going to need some help actually getting it to the point where you’re comfortable with the quality of that data there. You can have the best analytics tools in the world, but if the underlying quality of data isn’t there, that’s where the focus needs to be. Our sense is people need help with that. It’s not the sort of thing they can just do with a purge on recruiters filling out forms properly every six months or so. That’s the conversation for another day, John.


John Sumser: Thank you so much for taking the time, Dominic. This was a great conversation. I really enjoyed talking with you, and I hope everybody in the audience got as much out of the conversation as I did. We will see you next time. Dominic, thanks for taking the time to be here.

Have a great afternoon, everybody. Bye-bye now.


End transcript

Tagged with:  
Read previous post:
photo of stair case in office building in feature article published April 8, 2016 v7.14
HRExaminer v7.14

Dwane Lay has lived through his share of HRIS implementations. In this week’s feature article you’ll learn that Data Is...