Ask Us Anything: Using Labor Market Data to Drive Workforce Strategies
This Q&A Digest has been derived from the Ask Us Anything session on "Using Labor Market Data to Drive Workforce Strategies" held on December 1, 2020, with Karin Kimbrough, chief economist at LinkedIn; Matt Sigelman, chief executive officer of analytics software company Burning Glass Technologies; and Nikhil Patel, partner at management consultant McKinsey & Company.
- The workforce system has many stakeholders, and there is often a disconnect between them. It’s important for researchers and data analysts to build tools that proactively connect system stakeholders and drive toward greater coordination in the labor market. Stakeholders include job seekers, workforce boards, nonprofits, employers, local governments, and researchers. Given the diversity of these stakeholders, it is also important to understand the varied, nuanced needs of these users of data analytics.
- The emerging theme for the workforce system is to look skill to skill rather than job to job. Workforce boards should ask how they can build on a person’s existing skills to help one secure a job with living wages and mobility. Data analysts should use the same mindset to build a model of skill adjacency that brings new efficiency to training and skills maintenance.
- Demand for skills is changing. Even jobs that have been slower to change (such as insurance claims clerks) are seeing the desired requirements change as much as 60 percent. Meanwhile, demand for so-called "traditional skills" is falling.
The Atlanta Fed’s Center for Workforce and Economic Opportunity offers a number of data tools and publications to help you track unemployment, reemployment, and other potential policy and practice suggestions while you manage recovery from the pandemic.
Federal Reserve Bank
Unemployment Claims Monitor provides data on initial and continued claims for unemployment insurance as well as claimants' demographic data. The tool also contains data on claimants in the Short-Time Compensation (Workshare), Extended Benefits, State Additional Benefits, Federal Employee, and Ex-Service Member programs.
Opportunity Occupations Monitor tracks trends in jobs that offer salaries of at least the U.S. annual median wage (adjusted for local cost of living differences) for which employers do not require a bachelor's degree—opportunity occupations—in states and metro areas.
Workforce Currents includes articles on various workforce topics addressing research, policy, and practice.
Exploring a Skills-Based Approach to Occupational Mobility from the Federal Reserve Banks of Philadelphia and Cleveland examines how transferable skills could pave the way for lower-wage workers to move up to higher-paying positions and help meet talent needs of employers.
Center for Workforce and Economic Opportunity Events describes upcoming events and includes registration links for Ask Us Anything webinar sessions.
Resources from our Panelists:
LinkedIn’s Workforce Insights offers insights about the U.S. economy and the current job outlook.
LinkedIn’s Economic Graph analyzes labor markets and recommends policy solutions to prepare for jobs of the future.
LinkedIn’s Career Explorer tool allows job seekers to see how their skills match available job titles in a specific location and highlights additional capabilities a job seeker may need.
Burning Glass Research uses job data to provide perspectives on labor market trends.
Rework America Alliance seeks to help unemployed and low-wage workers find lasting careers as we emerge from the COVID-19 crisis.
Event Q & A
Workforce Investment Boards
How can data elevate the quality of job coaches and connect Workforce Investment Boards (WIBs) to other important stakeholders?
The ability to translate data and understand its practical value can help career coaches provide valuable advice to job seekers. The data should help coaches work backward, looking at a community’s areas of opportunity to design career pathways for workers. Having data that reflect the specific skills needed for local opportunities in real time provides insight not only on the requirements for high-value work but also on the competencies needed to bridge the gap between occupations. Matt Sigelman of Burning Glass commented that skills are the distance between talent and opportunity. Burning Glass’s Workforce Command Center helps WIBs track displacement and demand and identify the most efficient learning paths. WIBs should reorient their training to look at repurposing existing skills rather than training from the ground up with new sets of skills. Especially at this time, it is important to capitalize on the skills people already possess.
How do Workforce Investment Boards remain relevant during the pandemic?
COVID-19 has raised pressure to reach people more effectively. Convening remotely has reduced logistical challenges these groups faced before the pandemic. Uncertain staff availability or the need to open and close offices often made evening or weekend events difficult to plan. As remote work has alleviated those challenges, partners have been able to serve more workers at their events. Remote meetings have also given the boards a chance to connect with employers and build new relationships throughout a region, since travel and commute time has been eliminated.
WIBs, with their dual mission of serving both job seekers and employers, are an important connection between industry and the workforce. Burning Glass data track the breathtaking pace of skills changes that have accelerated during the pandemic. On one hand, that change heightens the need for a facilitator/ intermediary, and WIBs can serve a critical clearinghouse function. On the other hand, that dynamism speaks to the imperative for WIBs to have access to detailed, real-time, local information in order to make effective connections and be perceived as relevant.
In September, the Center for Workforce and Economic Opportunity hosted an Ask Us Anything webinar with Midwest Urban Strategies, a workforce consortium with 13 members. For more insight on how these workforce boards are facing challenges from COVID-19, watch the webinar or check out the event Q&A.
Soft skills are not categorized or measured in the same way as technical skills. What should workforce development stakeholders do to focus on opportunity skills or job readiness training?
Unlike technical skills, opportunity skills, or soft skills, can be tailored to fit the discreet needs of a target job. Career coaches can help job seekers tailor nontechnical skills for a specific job. For instance, resumes can highlight communication skills related to the job: Does the job seeker have experience communicating internally and externally? Is the job candidate comfortable with technical language? Showing the value of such skills to an industry will help job seekers market themselves and help employers understand what workers can bring to their company.
It's also worth reconsidering the type of skills that are considered fundamental. A recent Burning Glass report charted The New Foundational Skills of the Digital Economy—skills that are in broad demand across occupations and levels of educational attainment. In addition to people skills or social intelligence, today’s foundational skills include an array of business and digital expertise as well.
Transferable skills are going to be vital in the post COVID-19 economy. How are data changing to help job seekers understand the capabilities they have versus the ones they need to develop?
Burning Glass Technologies and LinkedIn are tracking how skills are changing within professions. Burning Glass data reveal the fastest-changing professions and includes a model of skill adjacency—that is, the relationship between jobs based on the presence of related and transferable skills—that it first developed for the World Economic Forum for the Reskilling Revolution report it undertook with the Boston Consulting Group. LinkedIn has tools to identify skills in high demand by metro area. LinkedIn created assessment exercises that allow users to take quizzes and earn badges in specific skills. Many of these skills are tied to digital literacy, a competency for which there is a growing need. Additionally, LinkedIn Learning has created Learning Paths, a compilation of video courses for in-demand jobs that can be completed online. The career roles covered by these videos range from data analyst to customer service specialist. In October, LinkedIn launched an interactive tool that allows job seekers to see how their skills match to available job titles in a specific location. The Career Explorer tool highlights additional skills a job seeker may need to make a transition and will also feature LinkedIn Learning courses to help attain them.
Are employers changing the way they post jobs to emphasize skills that are in demand? Are hiring practices being updated?
As an increase in remote work allows location flexibility, we do see that employers are also being more flexible about job requirements. Rather than focusing on formal educational attainment, employers have started to look at other skills. Private employers want to expand their pool of candidates, but in many cases, they can’t articulate the skills they are seeking. Insightful data can help identify these skills. People working in research and engagement can help employers understand their needs by reporting the skills an employer is struggling to hire as well as the skills workers are actually using in their roles.
If employers want a diverse workforce from a skills standpoint, then they may need to change some hiring practices. For example, they may need to be willing to accept candidates that have the skills needed for a job, even though some desired education may be lacking. Being explicit about desired competencies and not leaving job seekers or workforce boards to guess about them is a crucial way to promote skills-based hiring.
Do programs make workforce data available to high school students preparing for post-secondary career decisions?
Data from LinkedIn is available to anyone with a profile along with Workforce Insights reports and the Career Explorer tool (which is hosted on Github, for which a LinkedIn profile is not necessary). Data from Burning Glass Technologies is not publicly available, but many reports that use it are free to the public. A number of K-12 organizations (school districts, for example) use Burning Glass data to plan student guidance programs. Burning Glass, working with the ExcelinEd Foundation, recently launched CredentialsMatter.org, a site that includes a 50-state analysis of career and technical education programs.
While these reports and tools might be helpful to high-schoolers, it is important to engage career counselors and workforce boards who help students make career decisions.
Is there an inventory (or tool) of feeder to lifeboat to opportunity jobs?
The term “lifeboat job” refers to positions that are immediately available and require little to no retraining. The term was coined by the CEO of Burning Glass Technologies in the report Filling the Lifeboats: Getting America Back to Work after the Pandemic. The report explains that these jobs–examples of which are shipping clerks and personal care aides—can “serve as ‘lifeboats’ for those out of work” and help unemployed people gain skills that could lead to higher-paying positions in the future. A few resources from the Federal Reserve embrace this idea and approach. You can read Exploring a Skills-Based Approach to Occupational Mobility, a report by the Federal Reserve Banks of Cleveland and Philadelphia. The Opportunity Occupations Monitor from the Atlanta Fed’s Center for Workforce and Economic Opportunity shows jobs that pay above the national median wage but do not require a bachelor’s degree. The tool is location specific.
Is there an expectation for unemployment data to become more granular (aggregated by census tract) and be better incorporated into state labor market bureaus?
The unemployment data is based on a national sample survey. For that reason, it is difficult for it to become more granular. Other measures from the Census Bureau’s American Community Survey show unemployment at a more local level, though the data are not reported as frequently. The Job Data Exchange from the U.S. Chamber of Commerce Foundation provides tools to help identify available jobs and needed skills.
In addition to the analytics used to determine trends in hiring, what kind of "on the ground" data gathering is being done to understand community needs?
There is a need to convert employers into investors of talent rather than investors in consumers. Data providers should validate their insights with local employers to get the “on the ground” reality check alongside community organizations, training providers, and others who have a lens into the needs of job seekers locally. Convening these stakeholders on a quarterly or biannual basis can help ensure that training programs are relevant and the right credentials and skills are promoted so that job placements meet local needs.
How can tools like LinkedIn benefit lower-income, less-educated workers?
While LinkedIn does generally have a white-collar clientele, there are a growing number of first-line jobs posted on the site. First-line jobs do not require a bachelor’s degree but may require an associate degree. LinkedIn also offers NowCasting reports to provide information on hiring in specific areas, which can help WIBs focus on specific companies or sectors that are experiencing growth. For instance, a recent report on the Atlanta area showed openings for Amazon and Home Depot, which offer a mix of white- and blue-collar positions.
The Career Explorer tool can help job seekers research new careers by examining data on skills overlap and historic job changes. The tool is geography-specific, and people can use their experience to determine what jobs may fit well for them in the future.
Is there any recommended research on recent college graduates in the job market?
Burning Glass recently published a report called The Permanent Detour, which discusses underemployment among recent graduates. The research report found that 43 percent of recent college graduates wound up in jobs that didn’t require their degree. The report also found that those who start out being underemployed are far more likely to remain so a decade or more later, and it also noted significantly higher rates of underemployment for women.
Additionally, the Federal Reserve Bank of New York has completed useful research that allows users to:
- compare the unemployment rate of recent college graduates with that of other workers;
- track trends in the types of jobs held by those who are underemployed; and
- gauge the earnings of recent college graduates against those of workers holding only a high school diploma.
Stuart Andreason: I want to really welcome everyone to the Federal Reserve Bank of Atlanta's Center for Workforce and Economic Opportunity's last Ask Us Anything of 2020. We started Ask Us Anything, came up with a name at April when the world had changed, and we were no longer getting to learn from people in person and see everything that we could learn. And there was so much stuff happening, not only kind of in the world, but kind of really great ideas that we just said hey, let's try to pull together some of these ideas and get them out of the world to think about what's happened and what we can do to help come up with strategies for recovery.
So we are really lucky to have been able to get a lot of the people that we can and have the people on today that are going to talk about what they've seen happen and where we can go still in 2021 as we think about recovery. I'm going to turn it over to Sarah Miller, who has been kind of both the brains and muscle behind a lot of all of this that's happened [and] let her just talk a little bit more about some of the things that we've done and get us kicked off for today. So Sarah, it's all to you.
Sarah Miller: Thanks, Stu. You'll get my thank you check in the mail for that. It's been certainly a team effort, and part of that team is everybody that's on the call here with us today, all of the participants that have joined us on this journey. It's been a very strange year, and we really leaned into these virtual tables and virtual convenings, but you have all been a very engaged and active and inquisitive group, and it wouldn't be nearly as interesting of a series had it not been for your participation, so thank you all very much. I'm going to go ahead and share my screen here in just a moment, but I just wanted to take a second to reflect just on the year itself and why we're excited to have this particular conversation at this point.
So on December 1 of 2020, we've learned a lot over the year. It's been a year like no other with very record high unemployment, rapid changes to the labor market, a quick shift in skillsets in job clusters and job needs, and we're very excited to have the brains on the call that we have here with us today. Excuse me one second. I clicked right out of where I was going. So with that, I want to just kind of introduce you all to the presenters that we have on the call here today. First, I want to introduce you to Nikhil Patel. Nikhil is a partner with McKinsey & Company. We've been working very closely with him through what's called the Rework America Alliance that the Atlanta Fed is a core partner on that you may have heard about. So he's going to talk a little bit about some of that work but really taking kind of a whole systems approach to how we utilize labor market data and what that means for making investments and strategies going into recovery and into next year.
I'm very pleased to be joined by Matt Sigelman, CEO with Burning Glass Technologies. Burning Glass has been a great partner in this strange world of workforce but certainly also to the Fed as well. We use a lot of their labor market data with our Opportunity Occupations Monitor, and he's going to be talking a lot about some of the labor insights that we've seen from 2020 and what some of those trends are that we should anticipate for next year looking at the importance of skills and how skills can become kind of a new currency. And then finally, we're joined by Karin Kimbrough, the chief economist at LinkedIn. We've been incredibly impressed with everything that LinkedIn has been doing, a lot of the workforce reports that they've been releasing throughout the year, and Karin is going to have some very interesting observations about what's happening on the supply side and how we can look at kind of the equity imperative of what this means of getting folks back to work and elevating the skills that they have.
Just a quick kind of administrative note. As always, we are recording the session. You will get a full recording of this along with a transcript of the session and a digest of all of the questions and answers. Our objective today is to get through the presentations quickly so that we can have an open kind of conversation with our experts and with you. So please, at any point during this conversation, utilize the Q&A button at the bottom of your screen, send us the questions in, and we'll be able to kind of triage those toward the end of the call. If we don't get to your question, rest assured we've received it, and we'll track down some responses to send to you in a post-session material. But with that, I'm going to transition over to Nikhil to kind of kick us off. Nikhil?
Nikhil Patel: Thanks, Sarah. Hi, everybody. Maybe a little too ironically, I'm going to start a discussion about labor market data with a single page that has no data, but I think it helps before diving into the specifics of analysis and data just to think about the bigger picture of what does it mean to actually get impact from data. What is it we're trying to do with data? What does it take to translate the amazing and really creatively acquired data that's out there into something that actually helps job seekers to find jobs, helps people to achieve economic mobility, helps employers to be able to draw on the full talent pool that's really available and the best talent and create opportunities? I think as a start, though, I'll verbalize one piece of data that I think is important for us to keep in mind, which is to dimensionalize the need that we see in the country today.
So if we go back a couple of months ago, because the numbers are changing given what we're going through in the pandemic, but from a couple months ago, it's something like 12.5 million Americans unemployed, which is an extraordinary number, and it's just been a year of extraordinary numbers. When you look at the data more closely and you try to understand in the real world who are the people that make up that 12 and a half million, about eight to nine million of those Americans are people that are coming from lower or lower-middle wage occupations. So this is $36,000 median income and below coming without a four-year degree. So this is a large portion of that 12 and a half million are folks that are going to need to find re-employment based on their experience and their skills. And our labor market is actually not well equipped to be able to fully recognize that experience, those skills, and what it can bring to the table.
So there's a real need to understand in detail in a data-driven way, what are those skills, how do they connect to other occupations that may have more demand growth in the near term or long term that may draw on the great experiences of that eight to nine million Americans and give them trajectories to even higher income, more resilient, higher quality jobs. So if you look at it from that perspective, use of data is not just a theoretical thing. It becomes hugely important to be able to support that population and understand what kinds of actions are really going to help. Now with that context, one of the things I want to spend a little bit of time on as we get into this discussion is to think about what does it take to actually get impact from data.
And from my work and the work of my colleagues and many others, including folks in the alliance that Sarah mentioned before actually deploying efforts on the ground in cities or in different states and working with job seekers, working with community organizations and others, there are a few lessons that really jump out. And this is a nonexhaustive list but something to get us started. One is the idea of to really succeed in helping job seekers as well as others in the market, a key piece of the puzzle is linking the system together. The labor market is a system, and the challenges we face today are system and systemic issues. And there's a big piece to how do we actually get better connections between job seekers, community organizations and coaches, workers serving organizations, training providers, employers who are doing hiring, government and other policy makers that are setting the parameters and shaping the environment.
Right now, there is a big disconnect across all of those different participants in the system, and part of supporting job seekers is about bringing those pieces together effectively, so that's one major element. A second element is making the data actionable and this is an obvious point, which is to say we have data, we don't just want it to sit there. We want to make use of it in ways that are actually helpful. But there is an art to making data actionable, and it's remarkably difficult on the ground to take all the insights that you can get with the amazing data that's out there, and Matt and Karin will talk about some of the really creative things and important things that you can do with this unprecedented amount of data that we have at our disposal; but translating that into impact on the ground, there are a set of things that we have to think about very practically, and a lot of times we overlook it.
And then a third element of this is about navigating economic uncertainty. That was always true because none of us had a crystal ball about the ebbs and flows of the economy, and it's important as we think about data not to get into false precision, false prediction, but be reasonable in how we use it. That thought has become exacerbated now that we're in a hundred-year pandemic where we really don't know what the recovery is going to look like. And there are different scenarios that we can think through, but we have to somehow figure out how to make sense of the future opportunities using data as a touchstone to get the thinking started but without being too reductive about it. So I'll very quickly get into some of these points, what are some of the lessons. Again, not exhaustive in each of these different categories and as we have a discussion, we can dig into some of them.
On linking the system together, a key point here is if you build it, they will come—which is a typical attitude—doesn't work. You need a proactive, and maybe this is too aggressive, but a push approach to help different participants in the labor market whether it's job seekers, coaches and community organizations, employers, actually understand how to take the data and make use of it in a way that helps their own interest and also helps the community and brings the system together. One thing that we've seen to this end again and again is that people have created really great tools, they've put together initiatives, they made trainings free, and what we've found is it's actually quite hard to get uptake among job seekers. So for example, maybe an institution will go into a community and say, "hey, we've got this great program for you. We've got some great coaching advice." And the response will be, "look, I don't know the coaches. I don't trust the coaches. I don't know what this is. This isn't for me. It's not a part of my life."
Similarly, working with employers, you create a program and say, "hey, this is really going to help you access talent," but it just sits there without actually helping these different stakeholders understand how to use it. It will lie fallow because this is not the way the world is set up today. Second is the idea it's going to be hard to get a single version of the truth, but at least creating consistency in what the version of the truth that everybody is using on data, bringing different data sets together, translating it in a way that actually tells a coherent narrative across job seekers, employers, training providers, others, is important to link the pieces of the system together, make sure everyone is speaking the same language, and looking at the opportunities in the same way. If we think about making data actionable, one of the critical things is that it's very hard to pick up equity issues through the data and then understand from the data at face value what types of actions are really going to help.
There are barriers to people accessing jobs that don't come through the data that have to do with race, that have to do with gender, have to do with other factors. And there have been creative attempts to try to use data to tease out equity issues and opportunities, but the data alone is not going to be enough to really be able to figure out what needs to be done, so it's important to overinvest in that question. The data is a starting point. What else do you investigate on the ground to understand the interventions that are most going to help job seekers. A deeper understanding of the user. A lot of times what we put together when it comes to community programs, jobs, initiatives, tools is done in a laboratory, either an academic one or a commercial one or a policy one, but it's really not reflective of the realities that people face on the ground.
So for example, when we've done user journey interviews with coaches, they say, "look, I have to talk to dozens of people every day, the conversations I have are maybe an hour at most, and in many cases, there's not going to be a second conversation. Here is what I need very simply in order to have the most effective conversation with a potential job seeker that helps get them going in the right actions and knowing that there may not be a second chance knowing I don't have a ton of time." I think similarly, if we think about the user as being an employer, the way that hiring decisions are made can be quite complex. You can have HR actually setting a set of policies about what kinds of resumes actually get looked at; where does a human eyeball actually look at a resume versus getting pushed out by a computer, but then you also have people that are running the line business that are deciding this is what I want in a particular candidate. So actually thinking through who the user is makes a huge difference, and we need to be more sophisticated about it.
I'll go a little bit faster here but on the remaining pieces, what we hear over and over again [is that] awareness and trust are big needs, meaning job seekers often don't have awareness of the opportunities that are relevant to them that draw on their skills and experience. And then even if they hear ideas of what might be relevant, there's a big trust hurdle to get over: Why should I believe this? Why should I upend my life, go through the big expenditure of time for training for the sake of some recommendation coming from somebody I may or may not actually know and trust? So to get there, simplicity is key so that the suggestions coming from the data are understandable, and transparency is key so that the suggestions coming from the data are not in a black box and you can actually get your head around them and feel like this is real.
Geography matters a lot. Hiring tends to be local. People's lives are local. Mobility is not easy. It's easy to talk about hey, you can move here, there, everywhere. People are rooted in their communities. And so the opportunities that are available and what the data tells you—you need to break it down by geography so you get the nuances. And then finally, on navigating uncertainty, at the end of the day when it comes to a job market, demand is the big driver of whether there's impact or not. Is there actually going to be hiring in a set of occupations? Everything else ends up being a footnote to that very important basic point. The data can tell you some indications of demand, but if you really want to make the system work, putting extra effort into understanding local demand, getting as many sources of real intel as possible—especially from employers—is crucial, and then you pair that with big data on demand.
And finally, is the idea of optionality, which is it's too easy to say worker X is coming from an occupation of being a cashier or a customer service rep, and here are adjacent occupations that could be very relevant to them based on their skills. But if you just give a single point solution, you say computer user support specialist or frontline supervisor on manufacturing—whatever it may be—you may actually be doing a disservice to the job seeker because we don't know exactly what the demand will look like, we don't know exactly what the recovery will look like, and it's even harder to get a sense of that geography by geography.
So the most responsible thing is to say how can we create the most options for a given job seeker that will be resilient through different scenarios in the future. There are a lot of other lessons, but I think the overarching message here is at the end of the day if we want to have impact, there are real nuances about working with people on the ground, understanding the local nuances that make all the difference in the world. Thanks for letting me go on a bit. Matt, I'll hand it over to you.
Matt Sigelman: Well, Nikhil, I think really a fabulous laying out of principles for how we make data both accessible and actionable. So what I'd like to do is to build on that by speaking to that question of actionability. One of the things that... it's interesting. For those of you who aren’t familiar with Burning Glass, our work is in leveraging massive data sets in order to understand what's going on in the market with a great deal of local specificity, occupational specificity, ground level specificity. People often refer to this as real-time labor market data, but we actually think the more salient feature of this is about the granularity of analysis, the ability to actually get down to a skill level and therefore to get to the level of actionability. And here is what we are learning from what we're seeing across the billions of job postings that we analyze, the hundreds of millions of people's career histories that we analyze around the world each year.
First off, even before the pandemic, one of the things that we saw was just incredibly important here to this conversation, is the rate with which skills are changing. And that's important because often a lot of the discussion around labor market dynamism has focused around new sets of jobs. And in fact, what we're seeing is much more than new jobs emerging, is we're seeing that the skills that define jobs changing very fast, and that has very important implications for how workers remain relevant, for how workers get back into work, and how we make sure that the workforce that we need or the workforce we have is the workforce that we're going to need in the future. As you can see here from some work that David Deming at Harvard did, looking at our data, there are some occupations [inaudible] over the last decade, 40 percent of the requirements have changed. And then you sort of think about the pace of dynamism in training systems, obviously a lot slower typically than 40 percent, so it raises some significant questions and challenges about how workers can keep up.
If we go on, one of the things that we'll see is just how this bears out. Again, on the last [inaudible] these new sets of jobs that theoretically 70 percent of us are going to be working in by 2030, what you see on the right is how this really bears out. So for example, on the top left you see [inaudible] over the last decade from 150 data scientists to about 40,000 last year. [inaudible] you look at the universe of jobs looking for data analytics skills in an intense way, you went from about 20 occupations representing the 300,000 job openings to 68 occupations requiring 1.1 million job openings. So it’s sets of skills which are increasingly hybridizing a much broader array of jobs and again, that has significant implications for the shape of the economy. If we go on, what you'll see is how this is accelerating in the pandemic.
Just pick four jobs more or less at random here, top left and bottom right are professional jobs, bottom left and top right are more [inaudible] jobs. And what you see here is this is comparing the job requirements of these jobs over the last 30 days versus the same requirements last year, requirements of the same jobs the last year, and you're seeing some really profound rates of change. And again, even in jobs that have been slower to change like production techs or insurance claims clerks, we're... just among those production techs, we're seeing the rate of demand for these quality management skills and process management skills going up as much as 60 percent just over the course of the space of months at the same time that more traditional definitions of those jobs around manual dexterity, soldering, equipment operation actually significantly going down. And so when you see those kinds of rates of change, it raises questions about how workers can keep up.
If we look here, so as is, really hard for workers to stay abreast of this, really hard for them. As people get displaced, increasingly the kinds of skills that employers are going to be looking for as they hire back are not going to necessarily be the same sets of skills that people needed before they came in. And this is already having a disproportionate impact on communities of color and immigrant communities as well. And so this is some work that we did with policy writing for the JPMorgan Chase Foundation looking at the kind of impact that we're seeing on communities of color over the pandemic. As you see, those communities who were in a lot of cases furthest behind are the ones who are bearing the biggest brunt, who are already facing significant inequity challenges in the market. So I frame all of this because I think it also speaks to on one hand, the challenge that we have, but also to how this pace of skill change provides opportunities for us to overcome the equity challenges that we're facing today and also to overcome the kind of displacement that we're seeing.
And so what I wanted to do here is to look at what this speaks to how we can help people get... not only catch up, but get ahead. In every kind of job, there are sets of skills, which it turns out command premiums. The premiums in terms of demand, you can measure in terms of how long jobs take the skills, you ask for certain sets of skills, but there is also evidence in terms of salary premiums that people command, which give us insight into what are the sets of skills where there seem to be supply demand imbalances where you can help workers get into the [inaudible] and you see around every kind of job. And so then, this whole notion of a skill being a way for somebody to get ahead also becomes a way for people to be able to not only get ahead in their current job but to be able to move between jobs. This is from a wonderful report from the Cleveland and Philadelphia banks leveraging a data set to look at how people can move into what I'll call opportunity occupations.
So those [inaudible] jobs which pay above the national median wage for people who don't have a college degree. So you can see there are sets of skills which consistently can help people make that jump from being in jobs that are low-wage jobs to opportunity occupation jobs. So when we get to that layer and that lever of skills, we can help people open doors and move up. Just to sort of illustrate this, we've spent a lot of time looking at the question of how you can help people bridge the gap between jobs. We found, Nikhil mentioned this before, the notion of skills essentially measures the distance between jobs. So what you can do for any given job is look at what are the adjacencies, what are the different options that somebody has for being able to make transition. These are represented here singularly, but in some ways you can almost sort of think of it more as a web of opportunity with skills defining proximity between where people are and where opportunity is.
You can see, for example, machinists suffered significant decline in the early parts of the pandemic could transition into CNC programmers, which suffered much less of a decline. There are certain sets of skills that can allow people to make that jump. Same thing with data entry clerks, same thing with pretty much every occupation. So as we think about people being displaced, skills become a way of reframing the retraining imperatives. As we all know, retraining programs have historically been challenged to deliver precisely because they are trying to train people from the ground up with new sets of skills instead of capitalizing on the skills people already have. That same model is also key to how we can help people both step down and step up if they need too. We've been thinking a lot about the idea of what we call lifeboat jobs.
If you're in the North Atlantic and on the Titanic in the middle of the winter or, less metaphorically, if you're in July and your benefits are expiring, you may need to take a step down. How can we help people find the right lifeboat, get into the right lifeboat, and then move beyond it depends on skills. So it reframes the work that we have to do in the workforce community, not just to think about this once and done—OK great, we got somebody placed—but how do we teach people to not only get into jobs, but how do we help them move past the sidelines. You don't want to have to cross the ocean in a lifeboat. Somebody who is taking a job as a contact tracer, which may not be a great job, has a lot of but not all of the skills they need to make a move into a range of social service jobs where client interviews are key, to a range of customer service and sales jobs. What are the additional skills that will help them continue to move along?
It also becomes a really important way of helping drive equity, which is so much of the imperative of today because one of the things that we know is that the difference between... or the distance between communities who traditionally have been held back from opportunity and the kinds of jobs that employers are struggling to fill, it's often just a few skills. A shipping clerk making on $27,000 a year and 41 percent of them are people of color who learns Microsoft Excel and a couple other skills can become an inventory specialist making $6,000 a year more, which normally 30 percent are people of color. Now, once they get project management and scheduling skills can [inaudible] an operations coordinator. The talent pool is only about 20 percent people of color. Each step is a life-changing step-up from a compensation perspective, and it's a community changing step-up in terms of the level of opportunity and access that skills can help people achieve.
And I think two quick last thoughts here and [inaudible] going on. One, as we think about how we help people make transitions, it's really important to think about what are the jobs that not only are the right next steps but afford mobility. Here's two pieces of work on that. One again from [the] Philadelphia and Cleveland Fed on this idea of opportunity occupations—what are those sets of jobs that are still open to those without degrees, that still give good levels of pay to those without degrees. And the other, some work that we did with GSS essentially creating a typology of middle-skill occupations and trying to understand of them, and [inaudible] hear industry advocates talking about all different types of jobs and the wonderful opportunities they represent. Which ones truly provide upward mobility for workers? Which ones maybe don't provide mobility but provide stability and good pay? And which ones actually turn out to provide neither despite the claims often of very loud industry advocates?
As you see, it varies a lot based on sector where opportunity is, and so if we can align on what are the sets of jobs that provide not just good opportunities for the time being but future mobility, that's important. And finally, I'll finish here on the idea of if we go on the next page, of how we help people, how we orient to the question of what are the [inaudible] jobs going to drive the recovery both next year and into the next decade. And we're tracking a range of meta-trends that we think are going to be key to where the economy goes beyond the pandemic in terms of the readiness economy, the remote economy, the logistics of the diversified economy, and the automated economy. The readiness economy, just to illustrate, Thomas Friedman from the New York Times likes to say that COVID-19 wasn't a black swan event, it was a black elephant event. It was not an event you can't predict but the elephant in the room that nobody was paying attention to.
Well guess what? We're all paying attention, and you can imagine that industry and government will be investing considerably in readiness not only for the next pandemic but for physical infrastructure failures, for cybersecurity advance, for environmental catastrophes. There's sets of jobs and skills that are going to drive that economy. And when you can map those out and so we've got a report coming on that next week and if anybody wants copies, let me know. If you map those out, you start to think about how you escape to where the puck is going as opposed to building talent for where the economy is today. Sorry for going on so long. Let me, with that, turn over to Karin.
Karin Kimbrough: Great. Thanks. That was fantastic. Thanks, Matt. I'll try to be quick as well building on what they said. I'm going to talk a little bit about LinkedIn, just a real quick recap. So our economic graph is a set of sort of live data that as Matt and Nikhil mentioned, not just high frequency in real time but also pretty granular, spanning 722 million people globally and 55 million companies and 14 million jobs. So there's quite a lot of data in there including skill data and education data, and you can go to the next slide. What we do with that is we try to create reports that can be quite targeted, so this is sort of a screenshot of what we call the Nowcasting report, which is a report we can produce for policy makers that kind of isolates, like this one is for Detroit. So here it will say, "hey, Detroit, this is how many millions of LinkedIn members you have, or we have in your space and here are the skills that they have. This is the hiring rate."
That's that graph kind of at the top that happened during COVID, where people basically stopped getting hired around the time of the pandemic hitting. But we can also say: What are the top industries hiring in your metro area? How many of those people being hired are women? How many are in COVID-related jobs? What are the companies that are employing? What are the fastest-growing jobs, and what are the career transitions people make? And I think both to Nikhil and Matt's point, it's not so easy just to kind of talk about career transitions. There's a lot of devil in the details in terms of whether or not you're just transitioning someone to another, so to speak, dead end job or something that's really a steppingstone, whether you're giving it a localized feel and taste of if it's realistic for them. And if you have the next slide please.
We've tried to create some tools, humbly knowing that no tool is perfect, that let people kind of click through and say, OK, if I'm in the U.S., for example, and I pick a particular region and I pick a time frame, I can understand exactly who is hiring. So if you go to the next slide, for example, I pulled it for Atlanta randomly and I said, OK, what are... it spit out what it thought were the top trending skills, and you see it really ranges. Everything from teaching to time management, from real estate to account reconciliation. So it's all over the map. It definitely reflects who our member audience and customer base are. So we are typically known for pulling in the white-collar jobs, but what I will say is that now, over half the jobs on our platform are what we call first-line jobs—so these are jobs that may not require a college degree or may only require, say, an associate's degree, and a lot of those jobs are some of the fastest-growing jobs.
I think there is still a question around skill quality there in particular. Let's go on to the next slide. So just to continue this idea of the Nowcasting reports that we try to create that are very sort of targeted, we can also look at who the employers are. So for example, based on what we have on our platform, this is what we see as Atlanta's top employers as of September, and you'll see what it's reflecting—probably no surprise, and this is also the case in other places—is a lot of jobs in health and public health, a lot of jobs in real estate, and a lot of jobs in what I would call that transport and logistics sector. So, the Amazon jobs may not be software engineer jobs. They may be warehouse jobs. Home Depot is obviously reflecting a lot of that home improvement focus that's been going on in our economy here as well as the fact that retail is starting to creep back up in terms of the number of jobs that are available.
If we can go on to the next slide. So this is a different tool I wanted to just highlight. So the first tool you saw was really for policy makers to kind of have a sense of what's going on in their region, and this Career Explorer is more of the idea of a tool where you can go on and say what, if I have a certain set of skills in my location; I'm in Denver and I'm a data analyst, or I'm in Denver and I am a food server, what are the potential job transitions that I could make. And these are actually based on skill overlap but also on actual transitions that we have observed in our data of over 100 million people, say, in the U.S. or closer to 200 million people. So what are the realistic transitions people made not just based on what we think are conceptual skill adjacencies, but we've actually physically seen this transition happen and you can kind of sort over... And you should get at least four or five if not many more.
Obviously for data analysts, which is showing up here, you get 72 possible options, but there are fewer options maybe for a lower-skilled worker. But we do have them for many of the cities. Go to the next slide, please. The other thing we do is we try to think a lot about how to use our data to kind of make sense of where we are in the economy, and when it comes to our data, I think the only thing I want to just indicate is that we're really able to kind of track it by industry or sector. We're able to kind of see what skills seem to be more in demand and what skills people are adding to their profile. Go to the next slide, please. And one of the things that we're observing in terms of job-seeking behavior is how people who started off in one industry, when they move, where are they moving to. So here, for example, we're seeing... You'll see [in] recreation and travel and retail, harder-hit sectors, people are more likely post- pandemic to apply for a job in a different industry.
So people are more likely to kind of switch industries when they are coming from one that was hit hardest; and obviously if you were in sectors that were more resilient—and by the way, the ones at the bottom, the bottom two, were definitely more resilient—people are more likely to stay put because there probably is enough opportunity there. I think it's interesting to see entertainment, third from the bottom, because that is a sector that has also been hit quite hard but depending on the kind of job, you may still be able to stay in that sector if you're working remotely, or if you are not working or not working remotely, then you're going to have to switch. That's an interesting one to me. But by and large, what we're looking at is that job-seeking behavior has definitely changed post-pandemic and people are willing to move around.
I think it speaks a little bit to what Matt was getting at earlier, or not getting at, explicitly noted, which is that a lot of skills can kind of populate many roles and occupations going forward, and we're also seeing that in data I'll get to in a second. Can you go to the next slide, please? One of the things I wanted to highlight as well is just that we are seeing a lot of folks focused on remote jobs. Massive increases since March in the number of postings that feature remote work and the views by members looking for work, actually looking for remote jobs or seeking it and explicitly putting it in search functions and the increase in the share of applications to remote jobs. All of this makes perfect sense, but really large increases in that. And I think that's an interesting potential for what we think about for two reasons.
One is it might mean that there's going to, at some point, be a bigger mix of skills in different geographies because people can live in different places if this truly holds to some degree, so you might see that not all tech workers, of course, live in the Bay area is the easy one; but you may also see more human capital moving around to different locations. And we actually do see some of that migration data in our own data of people leaving some of the large and more expensive cities for second cities. The other thing that I think it does is there's a democratization of opportunity. You get a chance to maybe pull in talent that maybe doesn't want to move but is still adequately skilled to take the role, and where remote jobs are possible, you may be able to pull more people in who have transportation challenges or who have other barriers.
Next slide, please. I wanted to highlight, as well, something about the LinkedIn hiring rate and gender. We measure how people are getting hired, and one of the things that we saw was that women in particular were experiencing an unusual dip in their hiring rates relative to men, and this was pretty broad-based around the time of the pandemic. I'm interested to see if we get the same impact from more recent shutdowns that we had in November in Europe and here in some rolling locations here in the U.S. But what it tells me is that the burden of shutdowns was not borne perhaps equally. We definitely saw women not only being hired less but also applying for jobs less and even being approached by recruiters less during the pandemic. So I don't know how to explain that one, but there's definitely differential treatment that's happening or differential experience that's happening for women.
So back to that point about meeting people where they are, I do think that it's women in this case but also black and brown people, underrepresented groups, under-skilled groups are definitely all experiencing this labor market upheaval very, very differently, and I think any successful approach is going to need to be quite targeted. Next slide, please. So one of the things that we did, and I'm coming down toward the final set of slides here, was look at how do we think about what emerging jobs are. In the LinkedIn land, a lot of our jobs are very IT oriented. I understand that's a lot higher skilled than what the audience here is really thinking about, but I wanted to share that with you because you can see where they are and I want to call out number three, which is data, a little bit to the left you'll see data, cloud engineering data. That was one that is really a new area where it seems as if, as I think Matt was also saying, is populating a lot of different jobs suddenly is any agility or familiarity with data.
Similarly, HR and marketing and sales, these are all... I want to just kind of call these out. I know these feel like they're higher skilled, but these are all really broad streams of jobs that can ultimately be accessible. If you go to the next slide, I'll tell you why. So this is a bit of an [inaudible] chart, but the point of this is to show that in certain of these emerging clusters, what we see is a very large proportion of people that take roles in those new clusters are coming from other industries, other sectors, other jobs. So some of the newest areas are attracting people who didn't previously come from there. And the way you think about it is if I take... let's use a good example here... data in AI in the bottom chart, there's about 50 percent of the people moving into data in AI jobs came from somewhere else. So there is really a chance for people to come from other sectors and make a pivot into emerging jobs.
But the trick is really around the skills and what skills you bring, and I think Matt said it perfectly about the skills measure the distance between jobs. So if you go to the next slide, please. We took a look actually at just our own, what job postings we have, what are people doing, and we said if we could pick 10 roles that are on our platform that have been in demand over the past four years that we think are going to remain in demand, that require skills that can be learned online, do not require a BA necessarily, so we have seen people take roles in these jobs without a bachelor's degree, we have a significant number of openings here and these jobs typically pay living wage—what would those be? And filtering through all of that, we came up with these three roles, excuse me, 10 roles, and then linked to it online learning courses.
So I think the idea here is how to let people kind of upskill and train where they are; and obviously you can't learn all skills online, but I think you can kind of start to explore. So we really tried to do that to focus on skills. And if I can, I'll take one more minute and just mention something I don't have on the slides, but I wanted to share to talk about what we're seeing in skills. So we see both a huge demand and increase in people adding to their profile soft skills, which are things like people management, transparency, communication, collaboration, compassion, having hard conversations, talent development. All of these are really highly sought-after skills that I would call the soft-skill side, and I think they are essential. They're definitely growing.
We're also seeing a huge amount of digital skills. And I'll leave you with this. The digital skills seem to be permeating all roles at all levels. So if you think about it, you will have a requirement or a demand for more digital literacy even for someone who is doing inventory management or inventory monitoring in a warehouse. They will need some kind of handheld device to kind of maybe remotely monitor inventories in large warehouses. So you need some minimum level of understanding wireless technologies, handheld devices, that basic literacy. Same in healthcare. You need certain amounts of familiarity with software and data management to work in a lot of healthcare roles. So I think I'm just sort of maybe adding on to what Matt said about how this element of taking a somewhat simple, what might seem lower- skilled job, but adding some digital literacy really allows the person to upskill and access more options.
And I will say the fastest-growing roles that we see are in healthcare and food service preparation and in kind of delivery, back to that transport and warehousing jobs. Not all of these are highly sought after and the best quality jobs, but where they become better quality jobs is where they are married or matched with digital literacy skills. So I'm going to stop here and turn it back over to you, Sarah. Thank you.
Miller: Thank you so much Karin, Matt, and Nikhil. We could have taken a whole half day to have this conversation with you guys, but I really appreciate you laying out some of the insights that you've gathered from looking at the data and giving some guidance to folks in the field that use it to actually inform workers of their options and how to get them from one place to the next. We have a ton of questions, not all of which we'll be able to get to, but please submit them and we'll make sure to track them down from Karin, Nikhil, and Matt after the fact. I do want to pose a question to you, to any of you but maybe I'll start with Matt about the utility of this data when it comes to the workforce system or education training providers that have a little bit more structure around how they set up programs, what their mandates are and what their metrics are. If you could offer just a little bit of insight, I know that workforce boards and community colleges are a huge user base of your data, on how they can use this to stay within their regulatory structure but to be really innovative in how they use it.
Sigelman: So a great question and just a couple of quick thoughts. I want to go back to Nikhil's imperative before about actionability, about granularity, about locality or geographic specificity all as being, I think, key to how this [inaudible] out. I think there's a couple of ways that we think about this in particular; the principles are being played out in the world of workforce and the world of community college workforce programs and lists. One of those is... but I think the overarching theme here I think really is about the notion of thinking skill to skill rather than job to job. How do you build on somebody's existing skills and therefore come up with models of skill adjacency that brings new efficiency to how we help people train, to how we help people keep up?
So I want to start from that frame. And then I would think about here as well how we align, how we use those principles in turn to be able to really understand the work, to be able to work backward, if you will, from the way we normally tend to work in the workforce system. We tend to, and I again I think to your point of the question, a lot of our incentive systems are structured around being supply driven, supply side. We've got a set of people, they are a problem to be solved, and that's a supply, and we need to hopefully help them find their way. I think if we reframe the work that we do in the workforce system to be demand driven, to say OK, where is the landscape of opportunity, how do we work backward to be able to figure out how we can help people skill to skill master those opportunities. I think that's a transformative idea that allows us to be a lot more efficient and to make sure that we actually get to good outcomes at the end.
I think one of the last things I would say here, and this goes to the idea of being demand driven, is we need to make sure that the training that we are offering, that we have a really good framework for evaluating the ETPL, what are we investing in, and are the ETPL programs that we're investing in ones that yield an industry recognized credential for an in-demand occupation. And I think something as simple as that, which sounds like no duh, right? But the reality is I think most of us recognize that the eligible training provider lists tend to be a morass. They're seldom revisited and often there's a lot of politics that goes into them, so the extent to which we can create a set of yardsticks and the extent for policy makers in this audience, that you can create those sorts of mandates, you're doing a great service to the workforce community in allowing them to be more focused in how they invest training dollars.
Miller: Any other thoughts, Nikhil Patel, we can pick another question here.
Patel: I think Matt covered it well.
Miller: There we go. He always does. There's a great question here on what you've seen from employers. Maybe Karin, I'll kick this to you. Are we seeing employers kind of change the way that they're posting their jobs, emphasizing skills that are in demand? And then can you tell if there's been any change to that behavior around skills-based hiring?
Kimbrough: Yeah, sure. So a couple of things that we're observing. One is we definitely—I mentioned the remote work, but we're also seeing certain employers being much more flexible about how they think about the requirements for jobs, and that's something that I personally think is a great thing to show more flexibility. So not so much relying on signals around education, but also actually relying on actual skills assessments, so can you do the job. And a little shameless plug for LinkedIn I'm really proud of is we have a pilot program just inside LinkedIn—very, very small—to say why don't we just hire underrepresented groups of folks. If they show that they have the right skills, we're not going to expect them to have the degree; and then if they pass a certain test, then we're willing to have them come in for a face to face interview or Zoom to Zoom interview.
So I think there's a push toward not being so rigid around schools and degrees and being more focused on skills is one aspect. And I think the other aspect is recognizing that you really need, if you want to try to build up a workforce that you want to have, you're going to have to bring something to the table. You need to offer some training, be willing to accommodate people who don't have the full portfolio of ideal skillsets but are coming in with a gap or two and being willing to kind of play a role. And I'm going to go back to what Matt was talking about, which is I really think private sector has got to play a role here. If they are demand driven, I'm 100 percent on board with that, but they need to show up and say what it is they need and partner to help be really specific about the technology they want to use, the skills they're looking for. They can't just expect workforce groups to try to guess at how to prepare people. That's a waste of people's time and a waste of a lot of energy.
Miller: There was another question that came in too: how do we kind of ground test this and get that on the ground thing, so we definitely need to validate this data with our local employers, make sure this speaks to their reality, but I fully agree with you, Karin. If we can turn our employer partners into investors of talent as opposed to consumers of it then we have a much better exchange.
Patel: Can I add a quick footnote to that point? Karin, I think you captured the answer to that very well. Just a quick footnote, which is I think from the employer perspective. Sometimes we assume that they actually understand what they need and what they want when it comes to coming to the table, and I think we have an increasing... there is an increase in intent that skills-based hiring matters, that we want to create opportunities for a broader set of people, but I think it's still a mystery for most private sector companies of how. So it's incumbent on those of us that are closer in doing work on the ground or work with data, government, other, to help employers to actually figure out how. I think the intent is there.
Sigelman: And I would just say that part of helping with the how is a place where I think labor market data can actually be very effective because one of the things we know is that when you bring employers to the donut breakfast and you ask them what they're looking for, very often you get a bunch of platitudes. "I just want people who have a work ethic. I just want people..." whatever. We hear from people out in the workforce community and the education community when they instead sort of see the conversation or data and say, "look, these seem to be sets of skills that you're struggling, specific skills that you're struggling to get. Is this right?"
Then that really changes the conversation and makes it a lot more tactical. This is a great moment for skills-based hiring because ironically, even with so many people sidelined, we're seeing some increasingly severe skill gaps. And the silver lining of skills gaps is that they force employers to think differently about their supply lines and talent. So I think we can, in this market and the workforce education community, really push to drive for the kind of change Nikhil has described.
Miller: Absolutely. And I think that that's such an important point that you make, Matt. It's not just to show up and say, tell us what you need, but to direct the conversation and use the data to inform—have that be the flashlight, that focus on what the needs are, where your investments should be, what programs you should design. I just want to thank all of you so much for spending some time with us this afternoon. I know that everyone is thrilled to hear from your perspectives. We have more questions than we could have possibly gotten to in two hours, so we'll make sure to follow up and get those off to you. But thank you again for joining us on this journey. We wish you all a fantastic remainder of 2020 and the holiday season, and we look forward to continued work with you in 2021. So with that, I'll give you back your afternoon, and thank you so much Nikhil, Karin, and Matt for joining us today.
Patel: Thank you.
Kimbrough: Thank you.
Sigelman: Thank you so much.
Miller: Thanks, guys.
Andreason: Thank you all, and thanks to everyone for joining.
Miller: Take care.