Ben Hackett, Senior Vice President - Product Management at Accolade, Inc.
===

Emily: [00:00:00] Welcome to launch pod, a product management podcast from log rocket. Today. Our guest is Ben Hackett, senior VP of product management at Accolade, a personalized healthcare company for employers, health plans, government agencies, and consumers.

Ben began his career as a product manager at BlackTea Systems, an integrated technology services firm, before transitioning to LabKey software. Accolade eight years ago, Ben co founded Jido, an IoT platform that helped facilitate managers manage, monitor, and control thousands of conference room devices.

On today's episode, LogRocket CEO Matt Arbisveld talks to Ben as he shares his insights on decision making in M& A deals, cultural alignment in acquisitions, and the role of AI in medicine. So here it is, our conversation with Ben Hackett. 

Matt: Super excited to have you on the pod today, Ben, who we learned not only are you a great product leader, but also a plant expert. so, Yeah, thanks for coming on.

Ben: Hey, thanks for having [00:01:00] me. I think that's very generous. We'll talk to the plants in a couple of months and see how they're doing.

Matt: Yes, continuous status updates are important. Cool. I'd love to dive into a topic. I've heard you talk about in the past around. When to go from debating, working on a product initiative versus just diving in and trying and. The tension between talking about things versus actually. Biasing to action.

So just would love to learn how you think about when to talk about stuff versus actually dive in and work on a problem.

Ben: Yeah, definitely. I think there's sort of two dimensions to the problem. There's the I have a problem in a space that I need to identify a solution for, and then there's the, I have a problem , that if I don't get started, I'm just going to delay and continue to. Not do the work to find the solution and obviously for like user problems Or depending on who you're building for you need to you know Do your user interviews identify?

What the pain is and try and [00:02:00] address and solve and figure out what your leading indicators are to resolve that issue But then there's a secondary piece which is Sometimes I feel like just humans as people don't like to solve the problem and would like to just debate the surrounding issues of the problem versus actually being like, okay, how are we actually going to solve it?

Maybe like we take a problem should I buy a new car or not? Really? The question is like, what are the, of our goals? What are we trying to achieve? What's the most important factor? Is it for community? Nor is it for work? Okay. Once those things are identified, it's pretty easy for us to narrow down the car, but oftentimes I feel like a lot of the conversations are about like what is a car or like, why do cars have four wheels or well, the licensing and registration process and see how it's like complicated.

So if we did buy a car. Are we going to, have to go down to the DMV? Oh, but I don't really like that DMV. And then you just start [00:03:00] going, at least I just start going insane. And really the question is just go back to the core piece, which is like, why do you need the car? Okay. I have to commute to work.

You're not a construction worker, whatever your budget is X, Y, and Z. Like it narrows down the choices pretty quickly. And it feels like in product management where I think there's just a ton of opportunity where I push myself and my teams is just get down to, it really comes down to first principles thinking, but then using that to quickly identify what you're trying to solve.

Matt: Yeah, that, that makes sense. It's easy to get analysis paralysis. Like I remember when we started the company, we weren't thinking about all the things that could go wrong, but now. Now that I know more, it's, it almost seems impossible to go start another company when you know all the things that go wrong.

But if you just get started you eventually will find how to get through every sort of roadblock

Ben: Yeah, for sure. And you would never start if you knew all the potential downsides, but there's also the opposing piece, which [00:04:00] Also is crazy making, which is not doing any of the analysis, which is, like the logical fallacy of identifying a problem and then identifying a solution.

Like logically, yeah, I can see how that makes sense, but you're not measuring the potential outcome. You're not looking at. The conversion rates or your overall funnel of the existing data to say, Hey, does my hypothesis line up to the outcome that I think I'm trying to get to, and , there's just so many surprising outcomes that you didn't think would be the problem or the solution, but that once you get started, you start measuring.

You start getting feedback. You're like, Oh, okay. I wouldn't have thought about that upfront unless I had talked to people. Like recent example for me is. On our team, we provide reporting for our customers on the status of our offering. So we work with an advocacy in the healthcare world.

So we help employees navigate and their families navigate healthcare. [00:05:00] And one of our product areas is. If you're an employer and you have really specific clinical conditions within your population, you're trying to control. So something like, or impact like diabetes or musculoskeletal, high spend, high life impact for people that if you improve their life improves and your cost as an employer to serve those employees improves as well. When you talk to employers, What they say is, and you say, Hey, you purchase, let's say Virta, they're a diabetes reversal company. Super cool. You bought Virta. Tell me about what success would look like so that when I build your reporting on what we've done I can know what you want to see. And what they would say is I oftentimes I bought this clinical solution I would like to see an increase in the reversal of diabetes, and I would expect Accolade to increase utilization of that partner.

When people call in, I want you to send them to Virta if they have [00:06:00] diabetes that is reversible. Cool. How would you know that we're successful? What like when you say increase in utilization, what does that actually mean? It's I just want to see it go up and you're like by how much eventually like you have a million of these conversations, what you find out is there's a lack of sophisticated ability to set targets because our customers just don't have the data to know okay, what percentage of my population has diabetes?

What percentage of people who are referred to that program enroll will percent graduate? And so through that, what we learned is that really what we need to be able to do is set targets and be able to say, Hey, one half standard deviation above and below the mean within our book of business, you want this level of referral engagement.

And that will get you to the outcomes you're looking for. If you just talk to people internally, or even myself before these interviews, what we would have built [00:07:00] is more reports or what we were hearing internally team that were a bit further away from. This product was, we need more reports or we need additional transparency in our operational metrics, like how fast we pick up the phone.

But when you talk to customers, it's really none of that. They're just like, Hey, help me understand what good looks like through context of the data.

Matt: That makes sense. And are you able to set those targets for them or they give the kind of existing Yeah, I guess how do you set that target,

Ben: Yeah. For those targets, there's two areas. There's identifying our own internal customers where what is the highest level of achievement that's good and it's going to vary but that sets the ceiling and then looking at, in this case, like a hand, like half a standard deviation above and below the mean gets us to a reasonable.

Sort of. Area and then the other is external benchmarking, which is in health care, there's a lot of things like Millman index, [00:08:00] which will say, hey, here's what the average increase of your overall cost of claims is expected. So bringing that in and then saying, Hey, relative to this external index, this is what good looks like.

And here's where you're at within that context,

Matt: that makes sense? And you were able to get. Set that strategy based off. Did you show customers the reporting you had and they'd asked for, or I guess, what did you bring to them to get to that kind of results or finding?

Ben: it was a lot of just starting with the initial conversation of, Hey, in our partnership together, what outcome are you trying to achieve? That would make it so you would say, Hey, this has been really successful. And I think people's jobs are complicated enough. In all worlds, but in this case, the people we work with that sometimes I think for all of us, it's hard to step back to know what the answer is of what is my fundamental goal I'm trying to achieve in this metric represents that, [00:09:00] especially if.

You aren't, , an actuary or someone who's job requires to think in this way. And so , it just takes a lot of prying out, I think of just like classic good product management interviewing to get what it is that they're trying to achieve. Cause if you ask them, they don't know initially.

Matt: Yeah. If you ask what do you want, but everyone should have one sort of two KPIs that they're really trying to drive and they're talking to you cause your product should help with that. So I think that's a great

Ben: and it's not that, there's still a lot of data driven components, like especially around engagement and spend utilization that our customers are required and are, I think, really sophisticated in. But sometimes the, I think it's our job as the deliverer of product to be able to say, what, here's the problem we're trying to solve together.

And here's how we're doing and achieving. [00:10:00] And here's what good, it's like, it's our responsibility to make sure they know what that is versus expecting them to know always what they should have to do. Like we, we should offload that cognitive work for them as well as I think a good product partner.

Matt: the best solutions, usually no one will have asked for it, but it will when someone sees it, they're like, wow, that's amazing is what I found. And getting to that point is a really good feeling when you find that product that does that for them. Yeah,

Ben: such a good proxy for that, which is what I think, especially in the world of startups and what is so painful that someone would be willing. Accolades, not a startup anymore. But what is so painful that you would be willing to take a large risk in your career and purchase from a vendor who, Doesn't have a significant reputation.

Maybe then there's some VC backing that gives you some reputational signals, but it's [00:11:00] so painful that you're willing to make that bet. That's such a good problem to solve. And so indicative of. You're solving something for someone that people will pay for. And we've all had that, like for me, one thing that I would have paid so much money to solve in the healthcare is there's this issue of provider directories.

So if you're doing a search for a doctor, so someone calls in they have a specific health plan and they need to know whether a doctor is a network or not. The problem is that all of these directories of who's in network or not are constantly changing and they're not updated frequently, or they're just completely inaccurate at a level and a degree that is baffling, but there's no one, there's a couple of companies, but there aren't many people who can cleanse and provide updated in real time directories for providers who are [00:12:00] absolutely in network and creates a terrible member experience because or patient experience because you go whatever you, especially if it's a specialist, you wait two weeks, three weeks, months to have this appointment, you go, that's it.

It's now out of network. It's so painful and it's such a hard problem because you're dependent on that accuracy, not from yourself, 

Matt: I think you've just inspired 20 new startups in that space from this podcast. Hopefully in a year, you'll be getting a lot of cold emails about that. Yeah. That's it. You've done work in M. N. A. And I think that's a really interesting area now, where there are a lot of these companies that were series a series B, they can't get more funding, they have good products, and they're looking at sort of strategic options and would love to just learn your experience running an M.

A. D. L. How you thought through that? Any learnings about that? That process for you and maybe starting you could share what the deal was. 

Ben: My end. I was less on the due diligence piece of [00:13:00] some of that and much more of the folding into the loop and merging as one company, but I can talk to to both pieces. So we did two large acquisitions. Plushcare, which was a virtual primary care and still is, virtual primary care company that serves at the time consumers so direct to consumer virtual primary care.

As well as second MD, which was, and is expert medical opinion. If you get a diagnosis and like my mom's about to get a hip surgery if you want to talk to someone about, Hey, I'm considering this hip surgery. Here's the recommendation I had. Here's the type of surgery you can talk to one of the best, highest quality specialists in the country and get their opinion on your records and your information and confirm whether it's the right option. So that was close to 900 million in acquisitions, Which was wild and I think the most important [00:14:00] pieces for us were three things is this align with our strategic or overall strategy, i. e. for us, that's delivering care and there's an overall thesis. Let's take virtual primary care that there's a huge primary care shortage in this country.

Getting people to primary care is. One of the largest signals to improving their downstream health outcomes, and it's also the source for many of their referrals into specialty. really important types of care that if we can support for our membership, then we just have huge impacts on the outcomes in their lives. The second piece is, does this, from a technical perspective, are we going to be able to merge our platforms in a way which isn't going to be a significant investment, both in time and resources to complete. And then probably the one that I underappreciated the most [00:15:00] and which ended up being so important was the cultural aspect.

Are these people and huge respect to Raj Singh, our CEO. He was a former CEO of Concur and I, and, has a pretty good book and experiences doing this. And it was something that he told me early on that I didn't fully appreciate till we got into it. But are these people that you want to work with day to day, are they waking up thinking about improving patient outcomes?

Are they cooperative in, in their work? And do they want to build something together versus the opposite, which you can imagine is pretty, pretty difficult is there a lot of political gamesmanship is there only looking for your own incentives to be achieved? Are you completely checked out post acquisition?

None of that happened with plush care and emo. And that is its own due diligence which so important and so thankful for those teams to be kind.

Matt: And probably think so much experience having done a lot of these [00:16:00] and getting to know people and how to get a sense for what there'd be like post acquisition. So yeah, that must be a challenging, did you ever, I think a lot of times one thing about M and a, it's a build versus buy versus partner for each of these.

Did you consider a build solution or even a partner solution or was it was that not was it just so difficult? It wasn't really an option.

Ben: Yeah, in these cases, they're like, let's take virtual primary care, for example. The really big challenges are things like the licensing for physicians in all 50 states. It's the recruiting, retaining. And the workforce management component of staffing physicians, it's the sort of brand component of, this is a place where people trust that, takes years to build.

And so from a speed to market perspective and something that aligned with the people we wanted to work with, that alignment was a huge [00:17:00] part of that acquisition. So, I think if you were going to build a virtual primary care from scratch, all those components. Just are going to take you so long that you're probably going to miss your opportunity to make the impact you want.

And what we ended up doing is building virtual primary care into our advocacy offering for employers. So taking a direct to consumer offering and then making that available Within a B2B enterprise space so that our employers that we work with, their employees could have their own primary care physicians directly.

And what's cool is you can get an appointment through plush care and, an hour, 30 minutes, like pretty immediate and, that's just a significant investment in work and also just product management and a kudos to a team that was able to build something. With such high respect and really strong NPS on the consumer side.

Matt: It makes sense that products that have a physical regulation component just takes so much longer that there's a premium on, more [00:18:00] likely than say, a pure software company. Maybe a little easier to build instead. One last big topic I'd like to cover is so my fiance's in medicine.

I just love to learn your take about where AI and medicine will be. I'm guessing you're recording tons of patient interactions with these platforms. So just how you see that evolving and where I will fit in or where it's, you think it's still a ways off from really impact

Ben: definitely. There's the, like, two areas. There's the really boring dystopic piece, which is not dystopic, but I don't know, depends on where you land in the spectrum of. Glass half full, half empty. It's probably multiple things, but there's like the cost reduction side, which is how do you more efficiently do things like patient record summarization such that when you see that patient, which all of that's done by MAs now,, it's a pretty nice place for LLMs to be able to summarize a patient record.

There's a lot more models now coming out that are , much more healthcare oriented where those summarizations are pretty [00:19:00] good and it's relatively safe. And you can reduce your cost of labor in that regard. And I don't think people really like to talk about that because it's.

People's jobs and that makes sense, but if we're being realistic, I think that's one area. And then the second piece, which I think is really interesting is on, so I've been interviewing a lot of doctors recently on their use of their overall patient journey and the challenges they face. And one big area is around one, the prep for visit, and then two.

The amount of time that you have with the patient to meet two demands. One is the demand to be present in your visit and make the patient feel comfortable, listen, be prepared. And then there's the, how you translate that into something that eventually you have to code and bill and be for, which oftentimes are [00:20:00] in direct conflict with each other.

I eat one common thing you'll hear when you talk to doctors is. If I write this note and I give it to another doctor and let's say hypothetically I'm a high performing doctor. That doctor is going to look at the note and be like, Oh, this is a good doctor. They followed the right protocol and they're clearly advocating for the patient and the right treatment of care.

But if you take that same thing and you bill against it, it's not going to necessarily pass reimbursement. And a lot of these conversations where burnout occurs is the thing that's good for patients isn't good for reimbursement. I didn't get into medicine to care about reimbursement.

I feel a lot of conflict and this becomes uninteresting to me. And so you see a lot of like primary care physicians moving to direct primary care where they no longer take insurance. It's cash only and they can treat the patients exactly how they want. So what my sort of thesis is, I think there's gonna be something really interesting, which is, can you give in the direct primary [00:21:00] care space or with doctors, An opportunity to have almost like a way to manage and work the conversation.

Ben: Have a conversation with the patient, not work the conversation, have a conversation with the patient and then translate those pieces to whatever needs to be from a billable perspective. And then oftentimes there's a lot of patient follow up that needs to occur for also. The billable component and so just making it so you can separate out business side from allowing doctors to provide the care that they want a lot of that, I think, occurs through summarization and points for all of the medical records information and then interacting with patients via.

I think greater acceptance of, chat UIs with patients that are tied to pretty sophisticated medical models that can then translate that back to the doctor where they need to do follow up or treat the patient or in some cases bill.

Matt: And so right now is our primary care doctor. If I see. [00:22:00] Say a skin issue. And I, am I better off referring that to a dermatologist that I am treating it directly? Is that the situation or,

Ben: Yeah, where I think it gets interesting is, I don't think we want to rely on , the models, even the new clinical ones for care decisions, but I think they're sufficient in summarization in a high level, not recommendation, like when I talk to doctors, they don't say I don't want, an agent to tell me what to do, but I would like it to review all the information, tell me what's relevant, and then ask me if , have you considered X, Y, or Z.

Almost oftentimes what I hear is , I really liked it when I've had a med student in residency that I'll explain to the med student. Here's how I'm treating the patient. Here's the medications I'm ordering. Here's the exact amounts. And then to then do a review with that medical student for what did you see?

Does this protocol make sense? [00:23:00] And can you look at any additional literature that's occurring on this condition that you can bring back to me? So that I can make sure that I haven't missed anything. So it's less of this AI agent is replacing the doctor and much more. I would like the additional support for someone who has a significant more amount of time and resources to Do all the background research to support the patient and the record. 

Matt: from what I've seen, the AI solutions that seem most likely to work, it's already being done to some degree by someone or somehow, and it's streamlining that or making it faster, more efficient. But if the thing =, was not getting done already, it's very unlikely that I think AI will be able to get there.

In the next few years. So

Ben: And I don't think that's the end state, obviously, but I'm where I'm much more interested in is what the midterm state is because I feel like so many startups are, I don't know, maybe this is a trope now, but essentially build wrappers around [00:24:00] LLMs and then sell that. And then I think there's two problems with that.

One is that the models are going to continue to become more sophisticated and whatever use case you initially solved will continue to change. And so it should probably be somewhat model agnostic and then just get the workflow. And then two what is the actual problem to be solved? That's interesting.

There's a market for that improves patient outcomes. That's not some like hypothetical long term thing. And isn't, some short term. Really small problem that is going to be hard to scale. It's like that movie hurt. Do you ever see that movie? What's what I like about the movie is it's set in the near future and it's like close enough that it's interesting.

Everything still looks familiar, but it's far enough, at least then, it's crazy to think now at least then it felt still like sci fi.

Yeah. It's like the things that you can see it actually working in the next couple of years versus the fully virtual doctor is still a pipe dream and super unlikely. 

Ben: Yeah, and with liability and all that, [00:25:00] it's probably be faster than we think, but there's a lot of steps between here and there that need to be solved.

Matt: Awesome. A really interesting conversation. We covered a lot healthcare, M and a. How to make big product decisions. So yeah, I really enjoyed the conversation. Thanks so much for the time, Ben. And yeah, good luck with the plants.

Ben: Cool. Thanks, Matt. Appreciate it. Thanks for having me.