Podcast
Podcast
- 07 Dec 2022
- Managing the Future of Work
The digital "help wanted" sign. Can AI improve hourly staffing?
Welcome to the Managing of the Future of Work podcast from Harvard Business School. I’m your host, Harvard Business School Professor and Visiting Fellow at the American Enterprise Institute, Joe Fuller. My guest today is Sumir Meghani, CEO and Co-Founder of Instawork, an online platform for shift workers. A graduate of HBS, Sumir was convinced of the need for such a platform by his experiences with a family business and working with small businesses while at Groupon. His firsthand understanding of the staffing problems facing employers and the needs of shift workers inform Instawork’s business plan. We’ll talk about how Instawork uses AI to reduce the cost and time of connecting workers and employers, and how temporary positions provide the potential for longer-term engagements and career development. We’ll also talk about the state of the flexible or contingent segment of the hourly workforce. Welcome to the Managing of the Future of Work podcast, Sumir.
Sumir Meghani: Hi, Joe. Thanks for having me.
Fuller: Sumir, you’re a Founder and CEO of Instawork. Could you describe its business model for us? Because many of our listeners may be familiar with the type of worker you serve, the type of businesses that they work in, but may not know that it’s Instawork behind the scenes making the connections.
Meghani: Yeah. We’re a flexible work platform that connects local businesses that need staffing with about 3 million hourly workers across the U.S. and Canada. And we have a really special business model, where the businesses that use Instawork, who we call “partners,” they only pay when they successfully match with a worker, and we call our workers “professionals.” And so partners get charged when a qualified professional shows up and does a great job. And so our interests are completely aligned with the businesses that use us. We don’t require them to pay up front to post a job or need for help, like a job board. And then, for our professionals, the service is completely free.
Fuller: And how did you get interested in becoming a market maker, and specifically, working with the types of workers that get placed through Instawork?
Meghani: Well, the founding journey started a long time ago. I grew up in southeast Michigan, in Detroit, Joe. And my parents immigrated to America. They were small business owners. And whenever a surge of orders came in to their retail business or someone didn’t show up, some of their hourly workers didn’t show up, I’d get the call. So perhaps my earliest memories from the start of my working career was managing through labor shocks. And obviously, that’s a really timely topic right now. I joined Groupon right before starting Instawork. And I vividly remember when I was helping run the sales team for Groupon, I was sitting in the dining room of a restaurant here in San Francisco, and we were meeting the owner to pitch her on a Groupon to fill a bunch of empty seats that we could all see in front of us. The restaurant was maybe one-third full. And we had all the ROI analysis prepared and how Groupon could be a great customer acquisition tool for this small restaurant. And the owner stopped us 30, 40 seconds into the pitch and just pointed to a sign on the window that basically said, “We need dishwashers.” And here was a restaurant, a small business, that couldn’t handle more customers because they didn’t have staff. And believe it or not, this owner didn’t need more customers, she needed a dishwasher. And it struck me that the labor market for hourly workers was broken, was inefficient. In business school, we talk a lot about search costs. The search costs, the matching costs, seemed very high. The way you find hourly workers traditionally has been a sign in the window, word of mouth, maybe a job board, perhaps a staffing agency. And we felt that a mobile-first, liquid, efficient marketplace to connect these two groups, the businesses and the workers, made a lot of sense, right?
Fuller: Contingent workers of the type that you serve are reasonably recent phenomena because it hinges on things like having smartphones and also businesses getting comfortable with the idea that they’re looking for something other than a full-time or part-time worker to fill their labor needs. What can you tell us about who the 3 million workers you serve are? Who are they? What does your research show about why they’re interested in contingent work? Or is this just a stop-gap between full-time jobs?
Meghani: Well, Joe, we just published a Pinnacle Report a few weeks ago, September 2022, that we’ve been working on for many months. It’s called “State of the Flexible Workforce”. And it was published by an industry veteran, a Harvard-trained economist, Joe, Dr. Daniel Altman, who is building our economic research division. And what we did was, we took a bunch of publicly available data from the government and the Bureau of Labor Statistics and combined it with the rich qualitative and quantitative insights that we’re gathering from our 3 million pros and excited to share that with you and your audience and talk about some of the highlights from that report. We find that flexible workers are more diverse. Populations of color are much more represented with flexible work than they are with the traditional hourly workforce. They tend to skew younger. So 60 percent of flexible workers are in the 25 to 44 age group, which is much younger than the traditional hourly worker pool in the BLS, the Bureau of Labor Statistics. We find that flexible workers are highly educated. Ninety-seven percent of them completed high school, 69 percent completed some college courses or degrees. That 69 percent is nearly 15 full percentage points higher than hourly workers, overall. They get higher pay. The mean average Bureau of Labor Statistics data suggests that hourly professionals are making about 15 bucks an hour in the categories that we said Instawork serves. Our pros are making for those same roles about $20 an hour. So it’s about 40 percent higher. And then lastly, as you would expect, flexibility is a big thing that contingent or flexible workers are looking for. About two-thirds of flexible workers say that’s actually the primary motivation. They use services like Instawork because they’ve got some life event or situation that prevents them from maybe having a permanent job, or they’re just looking to make supplemental income on their own terms that is unique to them.
Fuller: That’s interesting, and I’m not sure many people would imagine that—that most contingent workers, according to your data, actually prefer these types of arrangements because it gives them flexibility. When people describe to you—and when you look at the people, individual case studies, of who you’re serving—what are the common things they’re doing? Are these ways to augment income, or is it a more diverse set of other calls on people’s time that cause them to view Instawork as a good solution for income generation?
Meghani: Well, as I mentioned, two-thirds of our surveyed pros in this flexible workforce report said flexibility is a primary motivation for using a service like Instawork. And we found that that share is even higher in situations where workers are in school, say part time or full time, or they’re caring for someone at home, as you suggested. And it’s interesting. It’s counterintuitive, but actually having more non-work demands on your time may actually create more availability for flexible work from the sheer number of labor hours that you’re available. Those hours, however, might only be available on an episodic or a nonstandard basis. The other thing that impacts availability of time and worker availability is pay. We have the ability to recommend that businesses pay more, for example, if situational worker availability is limited. So for example, say there’s a need for pros during a big holiday, and we think worker availability is traditionally going to be low. Well, having a dynamic labor marketplace actually lets us use price as a way to ensure higher liquidity and higher matching when needed. And so worker availability can actually go up, right? If pay goes up, and because we’re connecting both sides, we’re able to help nudge to those types of outcomes.
Fuller: Do you see other patterns in the data, Sumir? Like seasonality, or days of the week, or holiday periods really both affecting the availability of work and, obviously, labor rates if you’ve got peak staffing phenomena?
Meghani: We do see that. One area we see this is in seasonality, as you mentioned. Probably 70, 80 percent of baseball stadiums and football stadiums use Instawork, right? Because you need people—every Red Sox home game, right?—where then you don’t need them again until the Red Sox are back in town. And similarly on the worker side, you have people coming in and out of the workforce. Their availability may change, particularly the last few years with the pandemic and stay-at-home orders and school openings. And so, because we see the workforce changing in real time and hours available and hours needed changing in real time, we think the concept of flexible work is a great thing for the economy overall, because it really allows us to dynamically match both sides when there’s the biggest need.
Fuller: Why don’t you walk us through essentially a classic Instawork transaction. An employer engages Instawork, and then … first of all, how does that happen? And then, how do we go from that engagement—they become a client—all the way to your professionals showing up at their door, doing the work, getting paid, and all the value that’s created through that string of transactions?
Meghani: So let’s go back to maybe a real example that we can pick on. So HBS, maybe they’re holding a big alumni event on campus, and they need tents set up, and they’ve got to provide food to all the alumni who are coming in, or there’s a warehouse maybe that needs to scale up with a large order, particularly as we prepare for a Q4 and holiday shopping season. The businesses that use Instawork would post a shift through our website or mobile app and one of the first benefits is they don’t have to talk to a single human unless they really want to. It’s a big departure from the days of picking up the phone to call an agency or to put up signs right around campus or at the warehouse to recruit workers. Meanwhile, our pros have signed up on our mobile app, they’ve gotten onboarded based on the work they do, their skills, and once they’re onboarded, they get notified on their phone for shifts that are relevant to them, that meet their availability, that are convenient to them in terms of location. And so, after that match is made, the pro decides if they want to pick up that shift, and they have 100 percent freedom to decide if a particular opening is what they want, they then show up, and then they work the shift. Businesses pay us, and then we pay the workers as quickly as instantly after the shift is over. So we manage all the tracking, and the hours, and the flow of money for both sides. And pros can come together for a single instance, say six to eight hours in the example I gave earlier of maybe an HBS reunion, where you just need people for a day. Or maybe you’re a warehouse that needs people for peak e-commerce season, and you might need people for a few weeks, and our marketplace facilitates different types of matching based on what the two sides are looking for.
Fuller: So if I’m an employer—let’s say, you mentioned Harvard Business School as an example—do I get a menu of, “Here are five or six candidates.” Let’s say I need two people, do I get 12 selections I pick? Or do I leave that to the algorithm, and I’ve delegated that to Instawork for you to just provide me somebody that, more or less, you’re vouching for?
Meghani: It’s the latter. You are trusting Instawork to deliver you high-quality, reliable pros who can show up and do a good job. It’s a really critical part of the value that we add right to the marketplace, which is the vetting and the ratings on both sides.
Fuller: Let’s talk about the ratings, because obviously all of us as consumers are now suddenly expected to see customer ratings of products or books, and of course, I even get rated as a professor from my teaching here at Harvard Business School. But how is that rating data gathered? And what are you looking for in terms of how you put that feedback of the pro reviewing the employment situation and the employer, your client reviewing the performance of your professional? What is that data used for?
Meghani: So when a pro creates a profile on Instawork, Joe, they start by inputting all of their skills and experience into the app. We learn a lot about them—almost the way you and I might create a LinkedIn profile. We collect dozens of data points around, what are your certifications? Where have you worked? What do your references have to say about you? And every Instawork pro builds his profile today using data signals that are predominantly offline today and scattered. And the great thing about their profile is it gets stronger as they work more jobs, as they pick up new skills. After a shift ends, the workers and the businesses get to rate each other. And so there’s actually feedback in two directions. Some of it’s structured, some of it’s unstructured. And the structured data could be things like, did Joe show up in the right attire? Did Joe work reliably? Did he work quickly? How did he work with others? The pros get to rate the business as well. Was it a safe working environment? Were there Covid precautions taken as well? Did you feel the directions to get to the location were correct in the app?
Fuller: So, Sumir, obviously, you work as best you can to ensure there’s really a good match between one of your pros and one of your customers. What happens when things go wrong? What happens if either the process—that’s not the job I was expecting or I found the working conditions objectionable or maybe I didn’t like the way I was treated by an employer or when the employer is saying that person didn’t have the skills we expected or was late or wasn’t customer friendly, whatever else—is that just something you let play out through the system? Or do you intervene in those situations in some way?
Meghani: One of the benefits of Instawork is that structured and unstructured ratings feedback that we get from both sides—it helps in a few ways. First of all, pros and partners can raise their hand and say, “I don’t think that this is a good match for me in the future.” And so we’re able to customize what they see going forward based on the data they’re giving us. In situations where we’re seeing extreme behavior—good or generally bad by either side—that data also helps us take action. A very innocent example of that could be a lot of pros raising their hand at the end of a shift and saying, “It was really hard to find this room at the Harvard Business School campus where the event was held.” And it’s very innocent, but we can use that feedback loop to go back to the business and say, “Hey, can we update the instructions on future shifts so that the on-time arrival percent from pros is actually going to be higher, which is good for both parties?” And so we’re able to use data and technology to really watch how the matches are going and the fidelity of the matches to get them better and better over time. Of course, we’ve got a large number of people on both the partner and the pro sides who are at their computers 24/7 in case of really difficult situations where we need to intervene with humans very quickly. And so I always tell our team, we’re high tech and high touch to make sure we’re driving quality, and the ratings we collect is one input into all of that.
Fuller: So I think it would be fair to describe the service you’re offering to an employer is, you’re really allowing them both to fill unexpected needs but also to variable-ize a bit of their labor costs so that they can sculpt the matching of hours to task and not have a layer of underemployment or non-productivity because I’m staffing to the peak versus the trough. Is that a fair way to characterize it? And is that the way you describe it to your perspective clients?
Meghani: That’s right. Most businesses have a high degree of variability in terms of the tasks that have to be done based on seasonality. I think there’s another benefit, which is allowing both sides to come together—back to the theme of flexibility—like the try-before-you-buy concept, where even if there’s a need for a permanent match—and we think 60 percent of our pros are open or interested in full-time jobs, and many of our businesses may eventually seek to find someone permanent.
Fuller: That rent-to-own model—I’ve heard it described also, as well—has some prima facie appeal because, of course, now I can see what type of job that Sumir, Joe, or Katherine or Jorge did. Do you see that as a pretty regular occurrence in the system—that you have your pros dropping out because they’ve ended up employed by one of your clients? Or clients, they’re at liberty to pursue that, and there’s nothing in your understanding with either side of the transaction that you’d prefer they not hire your best pros, or you prefer that your pros not be soliciting full-time jobs through your clients?
Meghani: Look, at the end of the day, our value is in creating good matches between two sides. So if there’s a good match between a worker and a business, and they want to come together, what does it mean for us? It means our data is really working well, we’ve created a high degree of loyalty in our communities with our users, and it means that we’re going to increasingly be a valuable part of how those businesses will staff in the future. One thing we have noticed, Joe, is there’s a lot of shifting in the status and the income ladder. And so if someone’s looking for a permanent position today, that may not be the case in six to 12 months. Or maybe they’re looking for multiple part-time options. Or maybe someone has a great full-time job today, but there’s some event in their life that forces them to be looking for something more shorter term. Needs change, schedules change, desires change, and our dynamic marketplace really helps people to come together in different ways all the time. But we do see a lot of that shifting a lot.
Fuller: Sumir, what are the most common roles that your professionals are filling in? And how often are they working at Instawork, and for what type of hours? Are people using this really as a substitute for the 40-hour-week type job? Or is this primarily a vehicle for supplemental earnings or flexible earnings based on your availability in other life or employment circumstances?
Meghani: So we find, on average, flexible workers are working about 45 hours a month—so let’s call it 10-ish hours a week, so maybe a couple of shifts a week. So this is really supplemental income to deal with inflation, perhaps pick up new skills, maybe to find other industries or categories that may be appealing to pros. So it’s really a critical, but supplemental, vehicle for economic stability. In terms of your question about in-demand jobs or categories, particularly with the holidays approaching, we’re seeing a lot of demand around supply-chain logistics, as well as warehouse staffing. Both are particularly high demand. And then, in addition, when Covid hit, we saw certain segments of the economy, Joe, that were pretty disrupted due to Covid and stay-at-home orders. And those have been bouncing back very strongly. So, for example, I mentioned earlier, stadiums—probably 60 percent, 70 percent of professional NFL stadiums across the country are using Instawork because fans are back in full force, and there’s a high demand for staff to service fans who are coming back. And, of course, given the seasonality we described earlier, you’ll need a bunch of cooks and concession workers every other Sunday at Gillette Stadium. And so it’s certainly an appealing category for us. But those are some of the places we’re starting to see a really large increase in demand.
Fuller: A number of states have been increasingly concerned that somehow contingent work is going to start undermining the economics, a full-time job, or can lead to abusive relationships between that contingent worker and their ultimate employer. Certainly, our data suggests that the vast majority of people who take advantage of platform like yours are very happy to do it. A majority are not relying on it as a principal source of income, and it allows them to enter or reenter the workforce based on family considerations, other considerations, seasonal considerations, and that the anxiety that a lot of legislators and some regulators seem to feel about this may be just one of those classic “afraid of the new and dynamic and different” phenomena, as opposed to one rooted in concerns that have to be factored into some type of regulatory framework. Basically, the logic of employment has been around full- and part-time employees, and that’s reflected in the way that labor laws, and even the way states and the federal government, finance themselves—by payroll deduction of employees. So it’s a new model, and let’s hope that people don’t legislate it out of existence or constrain it so much that the types of economic benefits that occur both to your employers and to your pros get inhibited.
Meghani: One of the interesting things that we found during the pandemic was there was a bunch of economic research that suggested folks 55 years and older retired early during Covid. And although Instawork—our flexible labor pool—tends to skew younger, the biggest change we saw the last couple years was the percentage of our pros who were 55 years and older, like on our app, going way up. And so, for example, there may be someone who left the workforce, they retired early, but for the purposes of flexible work, they actually have labor hours they’re giving to our marketplace. And maybe they’re working at a Red Sox game at Fenway Park as a hot dog cook or as a bartender. And I think the traditional ways we look at employment as, are you in the workforce or are you not? The truth is it’s complicated. And the overall framework around labor law probably hasn’t changed much in 100 years. And so we’re obviously advocates of a more modern approach to how we think about employment and how we think about labor. And there’s a lot of interesting policy work that’s being done around the country in different states and at the federal level, perhaps around introducing new levels of classification that are not entirely 1099 or W-2.
Fuller: Well, we certainly saw that the so-called Great Resignation was actually several different phenomena, and one of the biggest was the “Great Retirement” of a very significant increase in the retirement rate during Covid of almost 2 percent. And, of course, all the data about people who are in retirement or edging toward retirement suggests that remaining active, remaining employed, is consistent with better health outcomes, obviously a lower likelihood of taking their social security benefit early, which has got a very heavy economic penalty to it if you elect to take your social security early and prevent drawing down on savings, and just generally both social and economic good. So you mentioned W-2, which is, of course, the employment status for someone who works enough hours to be on the payroll of a company, where that company is deducting taxes, FICA, social security, things like that, versus a 1099, which are contingent workers, where the worker gets paid, and that worker’s expected to pay their taxes on their own without the intervention of the employer. Does Instawork work include jobs that are W-2 jobs, or is it strictly 1099 jobs?
Meghani: Yeah. Well, yeah, we offer both. And we really allow our partners and pros to decide the right arrangement that makes sense for both of them. And then we look at either different states...but yeah, we offer both. I think at the end of the day, the most powerful part of our marketplace is its flexibility, which is sort of agnostic from the type of matching our employment, the ability for workers to choose when and where and how they work, for businesses to be able to flexibly staff their business. And so we think that that’s the most powerful part of the marketplace that we build.
Fuller: Now obviously, you’re relying on AI and tools, algorithms, to drive matching. You’re constantly trying to prove that. How do you see all this bow wave of technology and intelligent automation affecting your business and the ability to make the labor market more fluid and efficient going forward in the future?
Meghani: There are three areas today with our technology team where we think AI and machine learning can play a really important role. The first is back to quality. We talked about it as a big concern, right, for businesses in general and why they lean on Instawork and is core to our value proposition. So 3 million pros, we built a pretty rich database of the hourly professional universe—so historical employment, certifications, education training, other types of data points. And so using predictive algorithms, we can identify who the best pros are and onboard them seamlessly and quickly. And this can create a non-biased and fair experience for pros to get access to the shifts that can make them successful. We think we can create a fair playing field for anyone, right, by excluding things like race and gender, ethnicity. The data suggests there’s virtually no gender or racial pay gap on Instawork. So we want to create a true meritocracy, and if people have the will and the skill to work, they should have a fair chance. I think this is different than traditional labor market matching, where there may be unconscious biases that are introduced in many parts of the application process. The next area I’m excited about, Joe, is forecasting. For example, with artificial intelligence, we can dynamically forecast labor needs based on data we get from our partners around order volume, seasonality, wage trends. And then lastly, performance management is a third area we’re excited about. I’ve talked a lot about ratings today, ratings and feedback data for both sides to help create better matches in the future. So, for example, with workers, can we create personalized training curriculum? Think of an AI-driven Instawork university where we can say, “Joe, you’re a great server, but we think you can learn to make a better martini. And so take this class.” Or learn from your peers. Or for a business, you offer a high wage rate, but a lot of your professionals wish they had better onsite guidance or support. And so here are three things we suggest. And so improving quality of matches, improving forecasting, improving performance are three places where we’re spending time from an artificial intelligence or an ML [machine learning] perspective.
Fuller: Sumir, so talk to me about your ambitions for the company. How do you see it growing? Do you see yourself getting in the business of helping people either upskill by directly providing services or directing them to credential-granting institutions or businesses that’ll help them earn more? Where do you see this all evolving to?
Meghani: I’m glad you asked that, because we talk a lot at Instawork about our vision, which is to create economic opportunity for local businesses and professionals globally. One of the things that we’ve already seen start to happen organically as pros are more successful on Instawork, Joe, is they get top pro status. We can give them opportunities for faster career growth. So, for example, we found it takes on average about 60 days for pros to hone their skills enough to move from, say, an entry-level position to an intermediate position. Some examples of that, you can be an entry-level warehouse worker and become intermediate level and make an extra two bucks an hour. You can be a prep cook and then become a line cook that makes an extra $2.50 an hour. You can be a bar back and then become a bartender. So as you get rated and as you build your profile of record, you can unlock new opportunities on Instawork, which we think is really powerful. And many of our pros are able to now do multiple types of jobs and positions. And so, yeah, we’re going to look for more ways to encourage them to climb that ladder, to pick up new skills, maybe to learn from other peers when they’re on the job, or partner maybe with third parties who can offer credentialing. It’s all in that spirit of economic opportunity.
Fuller: Sumir, you’ve become a major force in the contingent workforce space. You grew through Covid hugely. What can we expect to see in terms of business developments? And what types of services that you might be offering either side of your two-sided marketplace in the future?
Meghani: We’re actually hiring. Instawork is hiring. We’ve got about 10 hubs around the world, including Boston. Specifically, with our partners and pros, we’re expecting a really busy holiday season this year, hopefully, as Covid stays low. So we just want to make sure that businesses have the staffing that they need to meet all the demand that they’re expecting. And with inflation really high, over 8 percent, pros need extra income more than ever. In terms of what’s next for us, with our pros, for example, we’re thinking about things like giving them access to training to help themselves, something we talked about earlier. We’re thinking about financial services for them, given that many of them are underbanked or living off of just $400 or $500 worth of savings. And for our businesses, some of the things that they’re looking for is, how could I use Instawork to maybe go find a great permanent hire of my dreams? How can you build software services for me that let me manage my entire workforce more efficiently, not just my flexible workforce? So those are some of the areas that we’re going to explore.
Fuller: Well, Sumir Meghani, Founder and CEO of Instawork, it’s really been a pleasure hearing about the growth of the business and the services you’re providing, and we’ll look forward to seeing some of those future developments unfold.
Meghani: Thank you, Joe, for having me.
Fuller: We hope you enjoy the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you get your podcasts. You can find out more about the Managing the Future of Work Project at our website hbs.edu/managingthefutureofwork. While you’re there, sign up for our newsletter.