HR Tech Weekly Logo

Hosts Stacey Harris and John Sumser discuss important news and topics in recruiting and HR technology. Listen live every Thursday or catch up on full episodes with transcriptions here.

HR Tech Weekly

Episode: 223
Air Date: June 27, 2019




Important: Our transcripts at HRExaminer are AI-powered (and fairly accurate) but there are still instances where the robots get confused and make errors. Please expect some inaccuracies as you read through the text of this conversation. Thank you for your understanding.

John Sumser
Stacey Harris


00:00:14:08 – 00:00:19:28
Good morning and welcome to HR Tech Weekly One Step Closer with Stacey Harris and John Sumser.

00:00:20:18 – 00:00:33:02
Stacey. You know if this was the 1940s you’d have that hard luggage and there’d be big stamps from all of the countries you visited in the last two months, you are an intrepid world traveler.

00:00:34:10 – 00:00:37:25
I am, but I want to say you know compared to many of our colleagues, this is the first time I’ve been back to back to back. And it has definitely been a lot of traveling for me personally. We know a lot of our friends travel that much on a regular basis so my thoughts go out to them because this is a hard month. I love seeing. I mean as an Asia Singapore. All through Europe and multiple ends of the United States in the last two months. I love seeing everything but I am really really glad to be home right now. To say my backyard looks like the best thing in the world to me. At this point so I will I will take a sunshine and ninety four degree weather in North Carolina is not bad at all today.

00:01:10:12 – 00:01:12:18
And so and how about you. Are you home?

00:01:12:18 – 00:01:30:24
I am home. I am doing the research for this year’s report and it means I think I’ve been complaining about this for a while. I am I am in demo land, I have done a hundred and four demos and I have 20 more before I’m done.

00:01:31:24 – 00:01:33:15
When’s the deadline. Is it next week.

00:01:33:15 – 00:01:34:23
Are you happy that?

00:01:35:06 – 00:01:46:06
The deadline was tomorrow. So that means there would be hard, I’m not going to do anything more, Deadline was tomorrow, so that means Monday Tuesday filled right up.

00:01:46:08 – 00:01:51:02
Right. Exactly yes.

00:01:51:06 – 00:02:23:10
We have the same issue with the survey. I mean shifting over done it. And I would get like five more people think could we just get in left and then we wait for a couple more days then we get to five or there’s always someone who wants to just get it under the wire and you try as much to accommodate we we end up this year with as many as we had last year a little less head and then click but just as many overall responses to the survey so yeah it’s been a good. Everybody wants to get into into the research that’s here for both what you’re doing what we’re doing. Everyone else it seems to be a good year for research. I think people are excited about the topic.

00:02:23:21 – 00:02:50:09
Yeah. Well it’s evolving fairly quickly. I think you’ve you imagined that there’s sort of two bookends to the overall set of research projects. You have one of them. I have the other you have the what’s really happening in big tech across the board in exquisite detail. And I’ve got the what’s happening at the very front end of things and frontier where the marketing know.

00:02:50:22 – 00:03:03:04
And so. So between the two between the two reports it’s quite a comprehensive view of a very fractured and complicated marketplace. Now that’s a very good way of describing

00:03:03:13 – 00:03:37:03
A fractured and complicated. I think you know the one thing that I have learned in the last year doing as much traveling as I’m doing this year is that you know our view from the state is fractured and complicated when you add the global complexity which I’ve always looked globally. But when you step outside of the U.S. centric global view right that complexity becomes tenfold because language becomes much more difficult. So the idea of what is artificial intelligence what is an algorithm what is engagement what is. Performance management changes dramatically from country to country in language to language.

00:03:37:05 – 00:03:46:04
So it gets even more complex so. So yes it’s very hard to sort of put all of this into a box that that can make sense for anyone who’s trying to make decisions right now.

00:03:46:12 – 00:04:18:29
I was I was speaking with a company yesterday and this is this is kind of an example of how broad the range is when she was a company yesterday whose theory about jobs and how you figure out whether or not somebody will be good at the job is that there are about 2000 jobs and the world should see them all as sort of the same job. And then then then with these standards what a job is then you can figure out who should fit and who shouldn’t.

00:04:19:00 – 00:04:57:04
And those people will see it everywhere and at the other extreme or more sophisticated whereas like Google learning class whose idea is that there’s no such thing as a job there are loose top level categories. But technology permeates these jobs over time for a variety of reasons and what you want to look at is how that complexity affects your particular workplace and you couldn’t have two different views about what the raw nature of reality is to all right.

00:04:57:09 – 00:05:01:13
One thinks it’s fixed to the other things it’s atomic.

00:05:01:21 – 00:05:33:17
And I think and the interesting thing on that is I don’t know that there’s a right or wrong in that either right. I think for some organizations the simplification down to 2000 job types gives them room to improve. Right. And for other organizations that would be so constricting would actually probably having a negative impact on their company if they’ve got to look at it. I think the other conversation here is you know all of this effort I guess and to sort of categorizing and thinking about the idea of definitions is valuable in its own way but it is definitely hard to figure out what’s right.

00:05:33:24 – 00:05:39:26
You intentionally write in a type of failure or a company or an investment.

00:05:40:12 – 00:06:13:11
Yeah it’s hard not to see that spectrum as a measure of how much time and energy you want to put into your people right. So at the. Everything is fixed end of the spectrum. Well that’s right and everything’s a Lego block. And people are Lego blocks and you can build stuff with Lego blocks. And at the other end of the spectrum is people were complex and varied and diverse and you put people together in all sorts of combinations and get all sorts of things done.

00:06:13:13 – 00:06:29:01
And the hardest thing is predicting what happens. And so so one is a we really want to understand people the view and the other is a we really want to understand and restructure the company.

00:06:29:21 – 00:06:56:25
And I will tell you I myself included many people love playing with their Lego blocks. Today I don’t think that’s a bad way to look at things. In some sense the other way is how you create a character in your cartoon called Sparky. You can write a lot of different things that it comes up to become what you would call a toy story character. Right obviously my mind was weak it’s put out a lot. I was with a lot of kids over the weekend.

00:06:56:29 – 00:07:14:06
Forgive me for the egos you remind me this is this is an argument we’ve had since the day we met. Does benchmarking matter right. That’s OK. You put your question there. Does this work you know who gets them.

00:07:14:20 – 00:07:25:19
We have. We’ve argued over that for instance the very first day we met nature. I don’t know that we’ve answered that question yet. The more good is we get an answer. We don’t have a polarization.

00:07:27:21 – 00:08:11:06
Well we have a lot of other stuff going on this week that I think will continue to spark the argument I want and I do want to get to some of the conversations because this conversation we’re having about sort of the spectrum of technology and where everything fits is being played out quite a bit in the news this week and in the type of announcements I think that are being made. We’ve got a lot of investments going on this week work for software got an additional investment in there technology if anybody knows them their workforce management time and attendance application. We also have degreed raising 75 million going from that technical workforce management side to someone as sort of futuristic thinking about learning and development as degreed degreed race five million investment capital

00:08:11:15 – 00:08:45:22
Which added to what they really have to think right. It’s like one hundred forty million or something like that. They’ve stated. And we also saw another investment this week a six point three million and mentor click which is a mentoring and engagement platform. So again couldn’t be any farther on the spectrums of different types of applications and their focus on your organization but lots of investment going on as well. If we get time we also this week. The global payroll complexity index come out from North Gate Renzo and the GP my group. Which is the global payroll international organization I always love this report it comes out by annually.

00:08:45:22 – 00:09:22:17
It’s one of probably I don’t always give vendor reports or sometimes can be I think a little bit self-serving in many cases and all of them have their different reasons for being put together but I always appreciate this one because I think it does give a level of awareness to anyone who’s working in payroll around the complexity of changes happening country by country and so they’ve done a really interesting job. We can talk about that and then I think before we get done today we really need to talk about what’s going on in the artificial intelligence space John because you’ve been doing a lot of demos and there’s a lot of stuff going on prior on which is an artificial intelligence enterprise tool for workload raised 20 million dollars in the last two weeks.

00:09:22:21 – 00:09:57:13
Someone from the IBM Watson team previously came there and founded that organizations we’re talking about. An Amazon launches their personalized services a fully managed a power recommendation service that you can now sort of basically build into any of the applications you’re building an Amazon Web Services which makes the case I think that you can reuse or repeat something like personalized recommendations and an algorithm. And so lots of different ways people are thinking about investing money and technology and taking time to think about the technology as well as workforce technology.

00:09:57:14 – 00:10:29:28
What do you think about this Amazon piece. That’s the first one I thought of when I was reading this for you. Is this reasonable can someone actually take the Web services the tools that Amazon has created to personalize buying a book or you know going into their Web site and creating something tie that into a little package you give it tease you can put it on your developed application or your developed toolset that you’re creating on their web on their cloud environment and reuse it as quickly as you would an API package for connecting an integration of two softwares.

00:10:29:28 – 00:10:31:06
What do you think about something like that

00:10:31:24 – 00:10:49:27
So visual you would. This is what the big players are all trying to do in general with a they’re trying to build repeatable processes. So when you look at you look at why does. Why does Google build tools that the chess games

00:10:51:20 – 00:11:21:20
become Joe Masters. The reason they do it is because they see the can come to understand how machine learning works in ways that can accelerate the machine learning activities of their clients while having a smaller requirement for data than than you do for the big big projects. And you know the the the there’s there’s a theme running through all of A.I.

00:11:21:21 – 00:11:55:24
which is that bias gets embedded in the algorithms themselves based on what they were designed to do and that that you you sort of get outcomes from a guy that has ghosts. The original purpose of the tool and so the Amazon recommendation service will have that consequence and you might ask yourself when the last time was that you bought something that Amazon recommended.

00:11:55:26 – 00:12:43:06
It doesn’t really work in a one to one way there. There’s something about it that’s interesting but it’s not for me very successful really. They’re standards of what’s successful are something that you might want to know maybe maybe if if it works one out of 30 times it’s a huge boost to their revenue I don’t know. They may have a very very very low bar for success. And so. So if that’s the case and you use the personalization engine and you can tolerate that level of success rate it certainly is the case that there are a lot of companies of age or tech who are spending their time working on recommendation tools.

00:12:43:06 – 00:12:49:02
Yeah without a doubt. Personalization has been definitely a theme I’ve heard all through the events this year. Yeah.

00:12:49:09 – 00:13:29:27
Yeah. Yeah. And so and so. So this will be tempting but personalization of an experience for a human being at work or the choosing of work or in the choosing of people with whom to network. That’s very different from buying a book. Right. And so. So what you’d have to do if you were a person thinking about using the Amazon services for each or functions is you’d have to you’d have to have a good grasp of whether or not you thought the Amazon tool could handle the complexity of human relationships.

00:13:29:28 – 00:13:42:07
You know it’s interesting you say how often does Amazon actually recommendations work in that low threshold conversation because you know in a marketing and buying world 20 30 percent can like you said make it big.

00:13:42:07 – 00:13:43:21
Billions of dollars a difference

00:13:45:09 – 00:13:48:24
20 percent marketing. You could be Jeff Bezos.

00:13:48:26 – 00:14:08:17
Exactly. But in the in the human world and in the world we’re talking about. Can you imagine getting 20 percent of your hires. Right. Right. That would that would kill a company. I mean it’s bad enough that I think I think that the actual numbers like 50 percent or something like that right now. That sort of thought out.

00:14:08:19 – 00:14:13:16
Yeah yeah. The actual number is ridiculously low. This could take it lower.

00:14:14:06 – 00:14:48:24
Yeah it is particularly well in that you know that that was a monster threw me on this one. I mean I love the idea of services that can be wrapped up and sort of bundled that makes the world a much more usable place for people who are non-scientific I know you read through some of that language in this conversation and a lot of it have to do with why why should just be people who understand python and understand programming languages who get to be data scientists can we give the data to people who have it on a sort of more regular basis through tools like this through do the acquisitions that we just saw last week a looker and Tableau and some of the largest firms as well.

00:14:48:24 – 00:15:21:01
I think the answer comes down to you can give it to them. But if they don’t understand at least some level of the underlying biases risk challenges with that data that or those what those tools are trained on from a data set perspective then you’ve to ransom some real risks and these aren’t developing things where I’m just personally making that vision. What these tools are for is developing software which then we’ll go out to thousands millions of additional people in different ways right and be used.

00:15:21:06 – 00:15:35:00
Yeah I think it’s a really interesting thing to see. How does something like this play out when they’re saying that this will reduce the need for training the datasets and that to me is probably the scariest thing because training data sets are what we talk about all the time.

00:15:35:27 – 00:15:51:09
Yeah. Well so the silver lining there’s an underlying view of the world that Amazon has which is that everything is an object in the supply chain. That’s the that is the Amazon point of view. Everything is an object in the supply chain.

00:15:51:09 – 00:16:05:15
And what you want to be really careful of is treating people with the idea that you can move people around because they have an array of characteristics is frightening

00:16:07:02 – 00:16:07:27

00:16:08:08 – 00:16:45:08
I would be it’ll be something to keep an eye on because I think we’ll start to see software that’s built with some of these tools. And you know in their comment here is that they’re that they are doing this in select regions so they’re they’re rolling it out with different I think capabilities with the idea that it is personalized by region or by area as well. I’d have to read deeper to understand that a little bit more. But I’m I’m it’s still it still scares me quite a bit to understand this might be something that would be added to an application at some level but we also have someone like prey to prey on raised twenty million dollars and they’re automating enterprise workloads with artificial intelligence.

00:16:45:09 – 00:17:24:04
Now this is a North Carolina Organization here in Raleigh my personal hometown. But I don’t know anything about him. You know I had just a store this conversation come up on my newsfeeds I was like Well this is really interesting. Twenty million dollars get a lot of money for a series a funding. And it looks like what they’re doing is they are basically aggregating and we’ve seen this done before but they’re aggregating your work environment. They see they fill the marketplace space between emerging technologies that are difficult and expensive to implement and those that are designed for simple roles based services. There’s somewhere between sort of an RPA robotic process automation and sort of a role workflow tool but they gather up your your work environments and then sort of automate that at some level.

00:17:24:07 – 00:17:35:29
Is this another place where we’re going to see more artificial intelligence. And is this more likely to come out with better I guess outcomes. Do you think then decision making type of artificial intelligence

00:17:37:10 – 00:18:11:21
So this kind of thing with prior does is it collects. I don’t know where the limits are right. I don’t know if it actually investigates e-mail for instance but it collects data from around the organization in a variety of sources and then purports to be able to tell you anything that you want to know. And the problem is that most companies don’t have all of the data that they wish they had. Right. And so what you’re liable to see was something like prior is that it doesn’t really get you anything more than you already have.

00:18:11:21 – 00:18:23:17
Right. The thing that the place where businesses really have problems with data is not understanding the data that they have. There are insights in the data that you already have that are missing.

00:18:23:17 – 00:18:34:19
But it’s not really the case that any technology available currently can give you enough insight about the future so that it makes a huge difference.

00:18:36:06 – 00:19:10:13
But the tools may make incremental differences but they don’t make great differences. What makes big breakthrough differences is having a comprehensive set of data. And so what you run into is something like this is that in order to make it work really well you have to go do all the stuff that you have been too busy to do up till now. Right. Right. And so it good in theory and I really like it. Good theory it’s really good in theory but it may not be as good in practice.

00:19:10:22 – 00:19:43:06
And which is very similar to how we think about technology right. People are like well if I invest in the technology it will provide me with the reports and the insights that I need right. Well then that’s not the case. We found out we’re on 30 plus years of enterprise applications being built inside organizations are deployed inside organizations. And what we’re quickly finding across the board when you look at the data set is that the number one issue within a year or two of implementing any system even today is reporting and the reporting is an issue most of the time because people aren’t putting the data into the system.

00:19:43:06 – 00:19:57:10
And so it still comes back to you need to track the information you need to do the hard work of change management. You need to do the hard work of making sure people understand why they’re using that tool whether it’s paper based or technology and. And then the technology can do the job right.

00:19:57:22 – 00:20:33:05
All right. Well then you need to layer on top of that. What was starting to be clear is that these are complex measurement regimes that we’re talking about and what people do when they find complex measurement regimes is they start figuring out how to game it. And so the stuff that’s been in the news this week is about I think it’s New York City where the crime rate dropped and dropped and dropped and grow and grow up because they were measuring it very very intensely. And it turned out that what was happening was that policing was going increasingly off the books right.

00:20:33:05 – 00:20:38:10
So the best the best way to have a low crime rate is to not report crimes. And

00:20:38:10 – 00:20:48:09
And so if you can solve the crime problem on the streets and some more physical way and you don’t have to report it.

00:20:49:27 – 00:21:08:21
Well that’s what happens. So so the demand for a decreasing crime rate based on measurement statistics produces this crazy response as people try to meet their teepees to keep yours. A guys. And so I think you start to see that to these things. Well

00:21:08:21 – 00:21:36:04
Well I think the gaming conversation is actually really really valuable here I am. Well it wasn’t like I read something which dealt with putting out some great new resource that someone had. But they were saying that this system can’t be gamed you know it’s just a list. Well any last or anything that has order or some sort of search capability to it can be gamed because you’re trying to be the first person or the first thing that someone sort of goes to or excessive. Right. You know am I wrong in thinking that.

00:21:36:04 – 00:22:08:07
Are there areas where we’re just putting data out there that it can’t be gamed at all and in some sense oh I just I just read this article this morning about people who are issuing phony news releases because these hedge fund investment algorithms scan all the news releases. Right. And so you could move stock prices up and down with fake news releases now. Right. So no no there is no such thing as that system McCabe reading that.

00:22:08:14 – 00:22:18:26
Well that remains a really good human beings are really good at particularly American human beings are really good at understanding what the rules are and then figuring out how to beat them.

00:22:19:27 – 00:23:13:00
Yes we really we really are fascinating. You know it’s my fault. Some of the gaming market because my son is in his 20s but heavily involved. Like most I think kids his age are. And yes he is still considering a kid because he’s my son I get it. He’s 20 is an adult definitely other older son tells me. But the idea of the amount of investment that goes into gaming mechanics right is that the psychology and the marketing and the data analysis in that world rival anything I’ve seen and what we’re seeing in this based of age our technology or enterprise technologies that industry knows the player and the human being at a level that I just haven’t seen you know in a lot of other technologies at some level and I don’t I think a lot of us are ignoring how necessary it is to the human psyche right.

00:23:13:00 – 00:23:16:16
In some cases to feel like we’re winning at something all the time right.

00:23:16:20 – 00:23:51:14
Or if it’s necessary that the human psyche. But it certainly is a distinct to the human psyche a better way of putting it. Right. Because it’s a doper being released and one way of thinking about the gaming industry is that it’s all about generating reliable local being released. That’s why I give it. But this larger thing about how do you beat the accounting system or how do you appear to have your EEO numbers right.

00:23:51:14 – 00:24:28:03
Well you really don’t have the intention of doing or how do you meet the letter of the homeowners association regulations while doing exactly what you want to do. It is a kind of a time honored American approach. I imagine it’s true in other countries. Well we’re having a set of rules doesn’t necessarily mean that they go forward. And that’s more of the kind of gaming of the system. I think they’re running into and they are figuring out how to beat the game for your advantage is is pretty common to people to try

00:24:29:00 – 00:24:51:11
You know whether you’re talking to a certain basic level or the highest level you know this gaming conversation I think is something that if it’s taken into consideration you sure you shouldn’t build around it to some extent right or build in some safety checks for it right. Is that feasible I mean in the work you’re doing with companies you know can they sort of build some roadblocks or some double checks for gaming of the system in some way.

00:24:51:11 – 00:25:13:26
I don’t think so. I don’t think so. So there’s a great article that I happen to have written on the H.R. examiners a couple weeks ago about how to game or retention system and it’s what are the things that you need to do if you want your management to treat you like you’re a flight risk. Give you a faster track raises and promotions

00:25:16:05 – 00:25:17:04
that raise red.

00:25:17:05 – 00:25:21:29
But doesn’t it feel like you want to leave. I don’t want you it but I want to think you’re a flight risk right.

00:25:22:09 – 00:25:52:18
Yeah certainly. Those are the things that I don’t understand how you would control in that particular case somebody changing their LinkedIn profile so that it would trigger the Monitor that you have that looks at whether that LinkedIn profiles are being changed because they want to signal to you that they need a raise. Right. And there are 20 things you can do where you start to cause the automated system to look at whether or not you’re a flight risk.

00:25:52:19 – 00:25:58:16
You think that you are and give the company the opportunity to decide what to do.

00:25:59:25 – 00:26:37:15
And I guess some people would say. But if you’re doing those things you’re thinking that way you might be right more of a flight risk anyway. Maybe there’s some some validity in that value sort of process right. More subtle versus direct way of saying I believe I need a raise. But then the question what kind of for those who don’t know how to play the game right. And that’s the other piece of this which is so then this gets very very jaded along with almost everything else that we’ve seen in the business market in some cases which is who knows how to play the game versus who doesn’t know how to play the game. So I tell my kids all the time is that you know a lot of what’s going on in the technology space is you will have to be part of the process or you will be out of the process right.

00:26:38:03 – 00:27:08:28
Yes. And there may be a kind of a diversity and privilege issue here. I don’t I don’t know those things are really hard for me to see because I see that sort of the pinnacle of privilege but the idea that the system exists for me to play you. Rather than being a set of constraints that are immovable that that way of seeing where the workplace is like I think that can be taught and transmitted.

00:27:08:29 – 00:27:28:22
And I think that almost everybody would be better off if they understood that where they are is in a world where it’s important to want to optimize your financial output. But I don’t know if we should do so. I think that’s a question that bedevils parents and has for maybe over time.

00:27:29:01 – 00:28:04:12
I would agree with that. Well we have rushed through our 30 minutes once again gone in a hundred different directions John from a different discussion about artificial intelligence to gaming systems to where the market is heading from an investment perspective. We didn’t just talk about payroll but I do want to just re-emphasize that if anybody hasn’t had a chance to see it. If you’re looking for the complexity of various payroll system this is just my answer because I use their report on a failure basis when looking at complexity indexes I don’t know if anyone else can does it. The G PMI an MP and get around the global payroll complexity report came out just a few weeks ago.

00:28:04:12 – 00:28:08:08
I think it’s worth downloading, it’s been a good week this week. Lots of good stuff going on.

00:28:08:25 – 00:28:30:17
Yeah. Yeah I think I think we should underline that the investment in AI continues. It appears to be moving away from algorithmic hiring and into some of the rest of the HR conceptual workload and there are really interesting things happening that work with software.

00:28:31:04 – 00:28:54:09
Well hopefully we will get a chance to talk a little bit more about some of the stuff that you’re finding in all of your demos once you’ve got a chance to think about it pull it all together in some semblance or form over the next month or so. I know we’re cleaning our data at Sierra Cedar so probably in the next couple of weeks we’ll be able to share at least some early tidbits from some of this stuff. But boy it’s just it’s a great time to be doing research in this space. Thank you for having the conversation as always John.

00:28:54:18 – 00:28:59:11
Yeah. Thank you Stacey, it’s been a great conversation and thanks everybody for listening in.

00:28:59:12 – 00:29:06:29
This has been HR Tech Weekly, One Step Closer with Stacey Harris and John Sumser and we really appreciate you listening. See you next week.

00:29:06:29 – 00:29:20:07
Bye bye.


Read previous post:
What Doesn’t Quite Work: Data Ages Like Fish

“There are no good ways for tracking the current relevant quality of the data under management in an organization. You...