$1.4bn Startup Tackling Loneliness

There's like a million old people, and they need help. So it's actually a non-sexy, untapped market, especially when I started eight years ago. That was a pretty important insight, and kind of a ballsy move in a way, even though it seems simple — because I essentially didn't ask "how do we fit in the rules." I just said, well, there are no rules. We're over here.

Yes. I missed the whole dating app thing, so I'm like, I need to experience it somehow.

Papa's a companion care platform. We connect older adults to what we call Papa Pals. They provide companionship, assistance, transportation. We used to call it family on demand, but we've really become a strategic in-home care network for healthcare organizations and employers.

I personally understood how bad it was because of my grandpa. We tried to get traditional services — and he bit a lady. They all wanted 40-hour-a-week arrangements, and he jumped out of a window one day. He had dementia, but he was okay — it was the first floor, he kind of scrappily jumped out of there. Still, it was just a difficult experience for us, trying to find services for him. And when I started looking at it: 10,000 people turn 65 every day — I think it's 11,000 now. It's a growing demographic, and there just aren't enough young people; people aren't having kids like they used to. You see what happens in Japan, where the aging population is becoming the largest demographic. So I saw a lot of trends, honestly. And then loneliness happened — that was the first big break we had on trends: loneliness is important.

Yes, it's pretty cool. I remember one of our investors saying, "We like investing in category-creating companies." I didn't really think of it like that before — I was just building one day at a time. But yeah, we did create a category. Before us, this service didn't exist. It was traditional home care — bathing and toileting. We're focused on companionship, though we're actually expanding into other care areas now. Our main focus is: how do we provide this new kind of provider — non-clinical in nature, but still delivering a lot of impact and solving the needs of members?

We came up with the idea to help my grandpa — my cousin and I together. And since then I've scaled it to support millions of people across the country.

Yeah, there are different kinds of buddy programs. A lot of people visit a senior facility in high school as a school trip and help out older adults — most people have been exposed to this in micro, volunteer-ish ways. But we felt it was important to have an economic component. It's free to our members for the most part — almost no one pays today — but we pay people to do it. They want to give back, they're mission-oriented, all of them, but they also need to make some money themselves. And the need is very obvious; almost everyone is going to experience it. My investors, my clients — health plans and employers — they get it because they've lived it. It's so tangible: no matter how wealthy or low income you are, you likely have a family member with a need, or you're a person who has no one at all — which is 25% of seniors. Elder orphans, as they call them.

The idea was to leverage technology to enhance what already exists, but in a very nuanced way. What we found was important was to decouple home care from companion care — which in health plan world is ADLs and IADLs. We said, wow, we could do IADLs. People like my mom, who want to give back, want to be part of a community, want to help — they could be great caregivers as pals, but they're never going to be home care professionals. So we separated out this new supply that was untapped, where the traditional supply is incredibly constrained. We added a whole new group of people to the supply chain and upskilled them through technology, process, scoring and ratings — kind of Uberified it. And now we can go back and add in personal care, because 20 or 30% of people need that, not 100%. We can properly manage our in-home network and support people across the whole spectrum of home care. From there, the world becomes our oyster — we become the true eyes and ears in the home, through both people and software.

Exactly. We brought in regular people who go through a very thorough process to become a pal. And we've really learned who the best pals are, how to manage them, how to notify them, remind them, nudge them, get them to answer questions — and how to make that happen accurately across millions of interactions. It's taken eight years so far. We're just scratching the surface, but a lot of the promise of what I dreamed up on my couch in 2016 is coming to fruition. A lot of our competitors in the early days got it wrong because they tried to tech-enable something that already existed; we tech-enabled something new. It's almost like the medallion thing Uber realized — if you pair individuals, you can build a network that's more of a peer-to-peer marketplace. And we realized that if you remove bathing and toileting, you don't just solve the supply problem — you can also package it up for health plans in a way that makes sense.

Very critical.

In America, if you do bathing and toileting, you have to have a home care license, or be an agency, or be regulated. And we're not any of those things — we do this new thing called companion care.

Exactly. I created a new little wedge of people who don't have any rules but can still provide value.

My Papa needed support and assistance. My cousin was at my house, and we said, what if we just get a pal to help Papa? We put it up on Facebook: who wants to be a pal to Papa? It wasn't viral or anything, but — think about this, eight years ago was a completely different world in so many ways. A pal applied. She was a med student from Venezuela and spoke Spanish. My grandpa spoke Spanish, and he absolutely loved it — and this guy didn't love anything, let alone a stranger coming into the house. So it was enough to convince us this was worth pursuing. We did it off the side of our desks while doing our jobs, playing around with it, built some technology. About a year later I said, you know what, I'm going to do this full time — which was kind of nutty, considering where I was at the time. But I did it. I bit the bullet and quit my job. My boss was supportive — he always thought I'd be an entrepreneur, I guess.

I had a very successful job. I was part of the founding team at a startup that was growing like crazy — MDLIVE, an early telehealth company. My father was the founder. This was before anyone was really pitching into the digital health space at all, incredibly nascent, and it was growing like crazy. But I'd been there six years, I had my experience, and I decided — financially speaking — that I was willing to go to zero. Now I realize it went negative, a significant amount. But I was able to convince myself it was easy. Honestly, that's what I think I've been doing unintentionally since the beginning. I really didn't have much question about whether it would work — I probably should have. I just jumped in and figured if I quit, it could get going.

Part of the reason was that in that first year we'd had three, four, five users — which is like nothing. But they loved it to such a crazy degree, it was so impactful to them. Maybe ten by the time I quit. Every Tuesday, Olga got her hair done. Every Wednesday, Gloria went for pastelitos at the Spanish bakery down the street. Every two weeks, Sam needed somebody to pick up his mail and help him get to the doctor. Those were literally the early members. I'd seen this Y Combinator thing that said if you can get a hundred people who really love something, it's more impactful than a million downloads. I didn't get a hundred, I got four — but that was good enough for me. At the same time, my boss promoted one of my colleagues. We were both salespeople, more or less — it was a startup, and I now realize I did way more than sales, but at the time it was sales. He made Brendan my boss. That pushed me over. I thought: I've truly maxed out here, it's time to try something new, go all in. We had four people who loved it — I could definitely get five.

It matters, because you have sales DNA. An insight I have about health plans is that you want a configurable solution that solves a lot of different problems, because there's so much nuance. There are bulk categories that matter, but there may be something very specific too. The idea started with my grandpa — but what I also saw was this: if you can do a lot of the different things health plans need using human beings, you'll get more yeses. If you can do a small program that shows value and could grow into a big program, you'll get more yeses.

So the flexibility of what we built is really a deep insight, which was: I'd sold to all of these clients — I sold telehealth software to them — and over the course of ten years, everyone wanted some slice of this functionality. So at Papa, from the beginning, we asked: how do we use in-home engagement — people and their relationships — in a positive way, plus software to guide these configurations and report at all these levels? It's taken way longer than I expected, but over time it pays off, because the client says, oh, I really have this important need. It doesn't mean you can solve everything, but you can say yes to more things — and in an ever-changing health plan world, where priorities shift with CMS direction, a new administration, different bills and rules, that flexibility is critical.

At the same time, a deep insight was that it takes a village to get these partnerships done, if it's a meaningful component of their business. Ten health plans have 80% of the lives in America, but they're really a lot of micro-businesses inside. There are lots of ways into partnering with them, but it's critical to prove your value. Another insight: it's okay to have a pilot, as long as you're not building all these different things you didn't contemplate at the beginning — do it through configuration. So, a long-winded way of saying: understand your customers very broadly, interact with them very broadly. And then the insight on the product side was — wow, this could actually solve a lot of problems, but it's truly one product. It's a friendly person showing up and doing something.

We had an app we'd started to build on the Tinder concept, and we couldn't get anyone to use it — it only worked on a perfectly clean, happy path. Unless I got on the phone and walked you through step by step, it just wasn't happening. So we thought, oh damn, we've got to rebuild the app, we're screwed. I said, we're not rebuilding the app — I'd put all my money into it, I had nothing left. So we kept the app, but when somebody requested a pal, I'd be the one pal available, and it would just be a Papa logo. That's how it started — I would individually text pals who were available. Then we realized members didn't even want the damn app, they just wanted to call me. So I changed it to calling me, and that's kind of when Papa started to take off.

It's true. Text is easy — off the cuff, a quick chat. We want to do more and more with text. We actually enabled texting between members and pals, and had to create a unique masked arrangement — they're not that good at getting into the website or app, but they're good at texting. You also want to protect the safety of the pal and the member, so we went about building that masking system in some interesting ways.

We recruited pals on Indeed, and we used Craigslist in the early days too. We didn't know anything about Facebook marketing, so we just posted and hoped somebody would reach out — on my own personal Facebook, not even an ad. That was the best I had. Getting people to be a pal was relatively straightforward; it came pretty quickly. Getting members was near impossible — I called every person I'd ever met. We ultimately put a little marketing money online, but "sign up online" did not work at all. It was: call us. Call the 1-800 number. They called me, then they called my cousin, then we hired our first person, Emily, and they called her. We set up a little off-the-shelf, low-cost phone system, started integrating with Zapier — which at the time was relatively nascent, at least in my world — and we just kept going from there.

We ultimately realized the most important thing for us to build was a call center tool. Everyone said, oh, you should use Salesforce, you should use this. I said, I want to build a real tech company. I wanted to own the user experience so it could be high quality and meet all our goals, and I knew that to scale this business with quality, at a fair cost for us and our pals, we had to do that. So we spent most of our effort building the administrative platform that now runs all of Papa — the dashboards, the visit success tracking, it all kind of spawned from there. Now, of course, we have a pal app, a member web app, a family care app, and we're about to have client apps. We built those later, but we still very much use our phone service to coordinate visits with seniors. We want to support our users where they want to use our service, not where I want them to use it.

We were at a WeWork in Brickell. When I quit my job, one thing I thought was important for my persona — my ability to wake up every day and think "this is as real as it gets" — was to get a four-person office. And I got a four-person office even though at the time it was just me. I'm like, well, we're going to hire people. I kept saying "we" — it was just me by myself. I had all these grandiose ideas with no real credibility. So we got a WeWork.

I started going on LinkedIn every day — which wasn't used this way as much as it is now; keep in mind, this is pretty much nine years ago — and messaging everyone. The world. Anyone in healthcare, digital health, health plans, employers, business consulting. And I was also thinking about investment — how do I raise money? I had no idea, no plan other than: I'm going to wake up every day and try to get in touch with as many humans as possible. That ended up being a really good use of time. I met a lot of people that way — my co-founder Alfredo, who joined later as COO and helped me build the zero-to-one Papa in a very meaningful way; investors; clients. And I ended up building a relatively big LinkedIn profile. I have 30,000-odd connections now, which matters in health plans — because no one wants to talk through LinkedIn anymore, but I already got it. And there are only 200 health plans.

I was going after anyone in healthcare, in any way that could be helpful, because I knew I'd ultimately have to solve a lot of problems at the same time. I needed customers, funding, partnerships, employees, advice — I didn't know anything. So I kept almost blindly searching, with some relative direction, until I started getting feedback: hey, we're starting a company, we provide companionship, we're going to scale it nationally, we're going to work with employers and health plans — is this interesting? What do you think? And the person is really smart, or works at a huge consulting firm, or runs a utilization team at a health plan, or runs case management, or runs benefits at HR. It's the most rapid method of getting user feedback you could possibly think of. I probably had 20 calls a week, because I was pretty good at the cold outreach. It became the first collating of information to make decisions. And then, ultimately, I was looking for clients, and I started focusing there — that's how I met my first ten clients. Probably thirty. Probably all my clients, actually.

Yes.

It was kind of an omni-channel approach, with myself doing it all. I was simultaneously reaching out to health plans and to local provider groups and practices. I downloaded Twitter at the time and reached out to the Herald and said, I have a local story, I want to talk about it. One of my four members was willing to do the interview, and it made it look like we had this amazing service — which we did, but for like 20 people maximum.

So we had an article. We talked to the local provider that worked with that health plan. The health plan put out a PR piece saying they were trying to solve loneliness, and I said, well, we're the loneliness company. I called the person quoted in that article, messaged them, and ultimately that turned into a meeting.

It's become more and more critical today, as health plans have fewer dollars, new rules, and people using healthcare services more than ever before, for lots of reasons. How do you maintain the solution you want to provide but also show ROI, when you've barely even done the service? That's the interesting part of a health tech business — it tends to be chicken-before-the-egg. We had the supply-and-demand chicken and egg. We had the health-plan-and-evidence-of-impact chicken and egg. Among other chickens and eggs — a lot of chickens in the coop.

I tried to think of things that were super believable because they're obvious. If you're alone, you're by yourself, as an example. Could I use existing research on loneliness that shows the impact? And if someone's there to drive you, and the problem was that you weren't getting to doctor's appointments, but now you did because they were there — and you have GPS showing it — that's an interesting thing. I thought about each element like that: how would I get proof that a pal did it? And even where I didn't have proof yet, because we hadn't done it yet, I needed them to give me the partnership to do it. If I'm in their shoes, how could I believe this could even be half right? So you use different evidence, and you use common sense. We had an inherent value — they got it, it just makes logical sense. If this woman is alone and suddenly has a car service, a helper, an assistant, somebody at least somewhat tech-savvy telling her what to do, probably there'll be some help there.

We were able to bring those two things together and say: listen, we haven't studied ROI extensively, but imagine if we keep them out of the hospital X percent of the time — that saves Y money. And we'll use your data, because I saw an article where your health plan says an ER visit costs $1,000, or $5,000, whatever. Walking through those steps to show how it could be.

What's interesting is that now benefits and programs actually have to articulate the ROI to actuaries — you have to tag it in this bid system and say, here's what we think. So thankfully, what we did — which was smart — was that when I was getting pilot programs eight years ago, I was saying: we believe we can save money, here are some of the measures we think we can measure, we'd like to measure them with you, plus patient-reported outcomes. Eight years later, we've been studying this with third-party validation. It just takes a constant body of evidence: we think we're going to do X, let's study X, okay, the result is Y — how do I make it 2Y, how do I make the improvement on ROI bigger? Maybe I should enrol more higher-risk patients. Maybe people with XYZ gaps in care. Maybe I shouldn't enrol my dad, who's 66 years old and doesn't need a pal.

So refining who we target, how we target, when we target, and what we do with them — that came over time, through feedback loops where there's a long time between the feedback. We've now done something like ten different studies, been published many times, claims-based studies. Every time we learn a little more and refine the model. What's nice is it aligns pretty closely with what we had originally — we use that as our goal, and we've just been chipping away at it ever since.

That's insane, right? And if they used it more than three times, it went up to 19%.

We're working on it.

It felt incredibly terrible. Really bad — pretty shocked as well. A lot of people say, well, this happens, you've done millions of visits. Yeah, but this is happening on a service that we created, to users who trust us. I think it was more of an eye-opening thing — not necessarily that it happened, more that it could happen, no matter what. Things can happen — there's just nothing you can do to guarantee they won't. But you have to have things in place to make sure you're solving for as many edge cases as humanly possible. 99.9% of our visits have no safety incident. We want it to be 100%. And we have protocols in place if, God forbid, something does happen.

We built our own background-checking anomaly system, advanced AI-enabled identification systems, interviews, training. Though that was a terrible moment in our company's history, it really opened our eyes: we have to take what we're doing incredibly seriously. Not that we didn't before — we always had trust and safety, and we face even more scrutiny because we work with Medicare; we're held to a very high standard compared to a consumer tech company. I saw Uber just release that they'll pair women with women — we've been doing that for two years, as a feature members can leverage.

It was one of those important moments for us as a company to come together and take it very seriously. We effectively stopped doing everything at Papa except focus on trust and safety for about 18 months. We were very transparent about the process — here's everything we're going to do, here's our progress, and 12 months later, here's what we completed. That resulted in a safer, more tracked solution that really solves for the edge cases — because on the happy path, or even the medium path, it goes well, but the amount of nuance that occurs when real human beings meet in real life is enormous. We've essentially built a human logistics platform that's pretty robust. From that, we've become ironclad, and we're proud to say we're the safest in-home company by a significant margin.

Not really, but I can see why you'd ask. I want to build a public tech company that's respected as a brand, and that comes with elements that aren't always great to be part of. I like being held to a very high standard. Our complaint rate right now — how many visits get a complaint — is 0.8%. Below 1% of the time somebody complains. Think about how incredible that is. It was 3% before, which also doesn't sound that bad — but how do you get it below 1%? How do you get it below 0.5%, which we're getting close to? I'm proud to be held to a high standard, both on service level and on scrutiny — a home care company can make some mistake, even a dramatic one, and the chances of them being in a Bloomberg article are very low. It's just not interesting for news.

That said, I think it ended up being a positive. I don't think about fairness. Is it fair that I came up with an idea and now I get to do what I'm passionate about every single day? I don't know — sounds great, I'm glad I am. I wanted to build a company that has meaningful impact, and this is part of what comes with it. I'm okay with that.

Well, I think robots. Human labor is the number one cost driver here, and for one-to-one companionship delivered this way, the only thing that can meaningfully improve the unit economics is robots. That said, I think we're pretty far out from a pal actually becoming a robot — though I could imagine a next step being virtual companions that complement, as a co-pilot to a pal. When the pal leaves the house, the member could then speak with the virtual companion. But I do eventually think we could have some version of robots helping with the assistance, and maybe some elements of empathy. For us, I'm more focused on building the system that delivers this value to our users and our clients. From there, if we end up replacing a pal in some cases with a robot, they'll still need our system, our logistics, our tracking — to understand what are you actually doing, why are you doing it, and how are you doing it.

I'm open to a lot of different things, at least exploring them. I'm very curious whether there's a way to leverage voice AI to build a companion. I didn't necessarily think to clone your grandkids — that could be fun. My grandparents would love it, I'll tell you that much. If you have more than one, that would make them very happy.

But I do think we're getting to a point where these virtual agents are pretty engaging. Whether it happens now with early adopters — I think we're still five years out from it truly feeling normal. Frankly, I talk to ChatGPT and all these other things myself all the time. Nothing too personal — usually strategic questions, or what I should do for dinner, or how to say something in another language. But it'll get more and more comfortable for people. My dad's 66 — in ten years, maybe he'll be using Papa. Hopefully he doesn't need to, but if he does, we'll be here for him, and we can be virtual, or in person, or perhaps robotic.

Whoa! Real Andrew's feelings!

Resilience is probably the key skill, or a key characteristic, that I feel I've earned — but also tend to feel quite comfortable with. Though I think it's less that I'm comfortable with it, and more that I'm willing to sit in the pain. The willingness to feel pain. I recently found out, for example, that Tom Cruise says he's not scared of heights — and I'm like, no, he's obviously scared of heights, but he accepts the fear and is willing to have it happen to him. By the way, I'm not willing to accept the fear of heights. I hate it. But when it comes to this, I'm willing to go to the dark place, as they say in extreme exercise. It's not that I don't feel anything — it's that I'm willing to feel it. I joke that I'm a masochist, but no, I think it's just an important element of being a founder.

The other part is surrounding yourself with a very high bar of talented people — both in raw, tactical skills and culturally. It's very easy for anyone — and I've done it a hundred times — to say, oh, they're so good, it doesn't matter that everyone hates them, or they're causing a ruckus, or people are talking behind their backs, because they're doing X and X is so good. But X would be so much better if they were able to fit into your world. That doesn't necessarily mean they're an evil person — it's just, what are you trying to establish culturally? Though culture can also be a weapon, so you've got to be very careful with that.

And just being myself. It's really important to be yourself and not try to be someone else. You can get insights like, oh, Jeff Bezos does this thing, it's so smart — but I'm just not Jeff Bezos, I'm always going to do stuff my way. Or, oh, I saw this founder launch this thing, I can't believe I never thought of that — a lot of this stuff is shiny objects. Just keep your head down, build something meaningful, hopefully have some fun doing it, and build a good team to help you do that.