Transcription:
Jennifer Overhulse: (
Hi everyone. As Tom said, I'm Jennifer Overhulse with St. Nick Media Services. I'm the moderator for today. We're gonna be talking about in the customer experience track, we're gonna be talking about personalization and how much is how far is too far. So, that said there's a lot of insurance companies after a lot of data these days, and it's a lot more accessible from a lot more places than it used to be. Customers willing to give up some of it, not willing to give up others of it and I think that's probably where we come to the rub at the moment. I'm gonna introduce our panelists and then have them talk to you a little bit about their companies and their perspective on the topic of data gathering and how much is too far in terms of personalization. We have Itay Bengad from Sensa Insurance, Ray Lynch from Coterie Insurance and Rumana Hafesjee from Guardian Life Insurance. Itay do you wanna start introduce yourself and tell us a little bit about your perspective and where you sit in the industry?
Itay Bengad: (
Sure, hi everyone. Thank you, first of all, for having me here. So we founded the Sensa Insurance and we launch our product about four months ago with kind of a different perspective. So on my personal background, I'm a medical doctor and we're trying to bring the, what we called bedside treatment in medicine to the insurance world with a very basic agenda of people should never be left alone in a car accident. Our first product is auto insurance. We've launched itto Illinois, Ohio now, Indiana and Texas. Our entire perspective of this industry is kind of what insurance is about, in terms of what a consumer expect from their insurance is. Insurance is like an ATM where you ask for your money back, if something happens or insurance is that thing that should handle things right there. Then when the sorry hits the film and we try to shift the idea of the moment of truth from the I'm gonna get my money back now to now someone's gonna handle that from the second it happens. So that's pretty much what we are about.
Jennifer Overhulse: (
Ray!
Ray Lynch: (
Ray Lynch. I'm VP of external products at Coterie Insurance. We're a digital MGA focused on the small commercial space in particular the really smaller end of smaller micro as some people might call it, offering a business owner policy, general liability, professional liability, and soon to be workers compensation. I think what's great about the panel and as you'll see, as we have pretty diverse backgrounds in terms of segments in the space and some of the personalization strategies can be a little bit different in commercial insurance, their businesses, the privacy is little bit different, but when you're down in kind of that really small end space, like we are a lot of times there's a lot of overlap between individual preferences and kind of themselves as a business owner. And we're really focused on simplifying that underwriting process for the small business owner for the agent that's serving them just to kind of help them with these transactions quickly and be able to find the coverages in a fair price in a simple way.
Jennifer Overhulse: (
Rumana!
Rumana Hafesjee: (
Yeah, Rumana Hafesjee from Guardian Life Insurance. I am the head of product management there. So I look after a ton of customer platforms. I think our perspective is generally that people don't like to deal with insurance. And so we have to try and make it as easy and as fortuitous for our customers as possible. I think one of the things that's great about this panel is where is that line that we really draw between what data is too much data. And what I found at least in insurance is people are very happy to give you very personal information because they feel like that's gonna give them a better experience. So that expectation us as a brand, we really have to live up in reality to that expectation. So super excited about this panel. Thank you for having me.
Jennifer Overhulse: (
Absolutely. And in that carrot and stick kind of.
Rumana Hafesjee: (
Exactly!
Jennifer Overhulse: (
Exactly. And she led right into my first question. Ray, what is the line when you're collecting data for the purpose of maybe bettering the customer experience? Sure. Where is the line between what's really gonna make the customer experience better and what's gonna invade someone's privacy?
Ray Lynch: (
I think there's important thing to kind of make sure you have is what I like to call relevance, right? So, the information that somebody's willing to share with you if it's in not an area where it makes sense they're gonna be much less likely to share it with you. So in small commercial, if we're asking them to share information through their accounting or their payroll provider, things like that in the effort for insurance, they're much more willing to do that as opposed to some other non-relevant touchpoint and like, through a paper good store or something like that that maybe they use to get their banners printed, things like that. So, I do think it's important that, it first has to be relevant where you're asking for that data. And then as far as when it goes too far, I think the customers are pretty clear and it's, as long as it's gonna either, you can demonstrate to them, I'm gonna get a better price or demonstrate to them that this is gonna make this process very quick and painless. Then they're willing to kind of open up a little bit more it's a little bit different with an individual than a company in terms of like, what is actually privileged information or not.
Jennifer Overhulse: (
So again, we're still kind of carrot and stick, right? Yeah. So they, you ask for information, they'll give it to you if they're gonna get something in return. Does that change, does that shift a little bit, Itay, maybe you can, if we're talking about information about my health or an individual's, own medical history, for example, or as opposed to their car, are they willing to share more information if it comes to, we'll give you the telematics data about how I drive, but I don't wanna tell you how much I weigh?
Itay Bengad: (
So I really think it's a matter of how intuitive it is for the consumer, that the data that is being gathered will result in something that again, will benefit with the consumer because we might find like a great piece of data that will improve our models by an order of magnitude. But for the consumer, it won't be intuitive enough that this piece of data will actually serve him in the right way. So I think that we always try to find, and to draw the line between the point where the consumer feels that intuitively without getting any inclination, because every eventually and there's a whole UBI industry based on that, you can solve many things with money, right? Consumers not always understand something but they understand money. So you pay them back, you reduce the cost, whatever, but what happens when you don't wanna offer money because, you know, you wanna make profit and there are many other things.
Itay Bengad: (
So I believe, at least in our company, we always try to find the very fine line between accumulating data that on one hand will serve us only when we think it's crucial enough and stay away from data points that it's benefit might be incremental and not exponential. So that's where we try to draw the line. So back to your question, we rather provide 95 or 90% of accuracy and staying away from some data that people might feel uncomfortable sharing rather than being aware of a hundred percent of things, but maybe eliminating 90% of the population that we wanted to have in our book anyway, by that. So this is what we try to do.
Jennifer Overhulse: (
Ray's point about relevance is really interesting to me too. Because for example, I think it was max Drucker with carpi data that told me once that the, how long someone has actually held their single email address is a better indicator of stability than their credit score. And personally, I would be more willing to give someone access to the time length of to how long I've had my email address. Then my, because I don't necessarily see the relevance of that does that may or may not make sense, but consumers probably have a different perception of relevance. I think when you say than then the insurer.
Ray Lynch: (
Always. Yeah. I mean, it's why we all get mailers in the mail because we went to a certain alumni group or because we've found that maybe this type of customer segment is attractive to the insurer. So we want to find those data that says, how do I find or get access to customers like this? But the customers actually don't like that or don't necessarily like that kind of targeting. I think it's more relevant when we're asking them to share versus finding them because of what magazine they subscribe to.
Jennifer Overhulse: (
Good point, Rumana on the life side, you probably have a pretty good perspective on this. There's all kinds of sources of data now that are available to you. Wearables, I mean, social data, everything from the medical information as well. So, what data would you say is really necessary in your opinion to provide a better customer experience? How do you see that?
Rumana Hafesjee: (
Yeah, I mean, I think we've done a decent job of actually asking our customers of what they find relevant for life policies and even outside of life and non-medical insurance you have to give every bit of information, your contact details, your spouse's details, your kids' details, your social security numbers. And most of the time we find that people are actually quite comfortable doing that because the trade off is there. They know that when they come into their authenticated sites, then their policy information will be correct. If they need to file a claim, things will be much easier if they've already provided us with that information. I think where people start to draw the line is if I haven't given you the data myself then, and you've bought it from a third party, that's when it starts to become, not as relevant, right.
Rumana Hafesjee: (
To your point, Ray, I think that where the expectation has sort of moved is if I've given you this information, I expect you to use it well, and keep it safe. I think that's just a social deal that you have. Similarly we asked our customers too, like where do you draw the line? And it was funny. They, drew the line at social profiles. So if I asked you for your Twitter handle, it was like, absolutely not, but if I asked you for your social security number, you were like, absolutely. And so, you know, I think that's the other realization we have is what do we consider to be personal versus what does our customer consider to be personal? So we do have an array of personal based information on people and their families and their makeups and their social security numbers. But, we don't resell that data obviously and so we use it when people are submitting claims, we use it for when they wanna exercise their policy, when they wanna change their beneficiaries to make that experience much more personalized.
Jennifer Overhulse: (
Yeah. So kind of a two fold follow up, right? So you get John Hancock, that's been fairly public about the vitality program, right. With the Fitbits and that kind of thing. Do you find that? Because I didn't, I was not necessarily aware of this, so your smartwatch or whatever it is, the Fitbit or whatever you're wearing, if I wanna share my step data with my insurer in order to do that, I have to share all of my data. There's no way, as far as I know that you can parse that. And I think I would be more protective over certain things than others, because there's some of these watches now that'll take heart data and ECG, for example, so maybe I've got AFib, right? So is my insurer able then to use that data that they didn't necessarily get? I mean, I authorized the data, but maybe I authorized the step data, not necessarily that can they adversely select against me, would they adversely select against me or would they call me up and say, hey, you may have a problem. So, number one, do you find customers are generally aware of all of the data that you are getting or that you're able to get? And well, let's start there.
Rumana Hafesjee: (
Yeah. I mean, I don't know if everybody has the same level of awareness. I think generationally, it matters too. I think there are certain generations that are much more savvy about, I know I'm giving you my smart watch data for my steps, but I know you're getting a whole bunch of other things. Even if you have it in a disclaimer, right? Like at the end of the day, I think there's the generational differences, but I think it's also about the trade off. So if, if you're giving us all that data and you get a reduction on your premium as a result of providing that data, I think a lot more people are open to saying, okay, well, you can have that data. The other side is, are you gonna tell my employer? Right. I think that's one of the bigger questions is who are you going to share this data with? And if people feel a sense of, it's really just gonna go in there for aggregate means anonymous, not shared with my employer. And by the way, I get a discounted premium, you start to get a lot more people wanting to participate in those types of programs, and you just have to be super transparent about it.
Jennifer Overhulse: (
What about context in terms of, and I know this wasn't on my little list of questions back there? So, again, kind of thinking this through right. Social data in particular, if I'm sharing information on my social feed that says that I've contributed to walks, runs, or rides on for breast cancer and autism and heart disease, my insurer looking at that information could assume many things. What about context? How important is context in terms of, of the data that you get? Because I think it's important to the insurer or to the insured. Is it important to the insurer? I know it's a bad question. It's a terrible question.
Ray Lynch: (
Right. Jen, you had to do it to me again. Should have known being on a panel again. But, so it's hard, right? Coz, my head automatically to a small commercial and it's just, it is different. So I definitely wanna caveat with that but I think context is important because there's like the famous website underwriting, right? So I was an underwriting manager, New York city a lot of issues with contractors. There's lovely labor law in New York that causes people a lot of issues. And it was always around like where did they actually operate? And their website would say Hamptons and the claims would be Manhattan because they had the same customer that had a place in the Hamptons, and then they came into the city to obviously do work with them.
Ray Lynch: (
So that's where I go is that like, you can actually be led astray or kind of, because it's not always, it's not going to be perfect, particularly if it's things that you're scraping from a social or a website or things like that. How often is it updated? How often is it relevant? So I think it's kind of one of those things. You can use it as a tool. But if you're gonna actually make all your decisions based off of what somebody's website or Facebook feed is, I think you're gonna get yourself into some issues.
Jennifer Overhulse: (
Well, playing into that in the context kind of scenario a little bit. Let's look at UBI just really quickly. So if you look at UBI, let's assume that you've got a customer and insured who has a telematics device in their car, and they drive generally in the state of Kentucky. I know these are not geographically next together, so don't worry. And I happen to drive to Texas, my speed limits in Kentucky, where they may be looking at how fast you're going. Are you speeding? Are you this? Are you that the speed limits in, in Kentucky tap out at 70, in Texas there's 75. So if I'm doing 75 in Texas, my insurance company, I mean, that's the kind of context I'm, it seems important. But I'm not sure how you get that context. Is there, can you think of any?
Ray Lynch: (
You hire a very smart data person, like we have, he's probably in the audience over there and he'll figure out how to get that context. But I actually have no idea.
Jennifer Overhulse: (
That's good. Okay. You Itay any perspective on context how that would maybe if you provide the context in which you're going to use the data or would that make your customer more comfortable?
Itay Bengad: (
So based on our experience, the answer is definitely very much. Yes. You mentioned before that your watch has ECG and number of steps, etc. And the current usage of data due to many reasons, there's like a huge wall between the data that I provide. Then it goes behind the wall. There are few decisions that are being made. Then I get something I don't know now. I told you that it's all about doing things that are intuitive, right? If you would have like a price list, say if you provide us with the ECG, this is a hundred dollars discount. If you provide us with a number of steps, that's 150, right? Like a price list that would be very intuitive, then people would understand what does it mean to actually provide the entire set of data or just one simple piece.
Itay Bengad: (
That's very helpful. In our case, we decided to gather data only upon accidents. That's a very intuitive things 99.9% of people do not want to be left alone in a car accident. It's very intuitive. So when you tell them, so we gather the data only upon accident, we know everything. So you won't have to do a thing and we take care of stuff. So I think that it always, we must remember that it must be very intuitive for the consumer. Why do we ask for that data and what will be used when we are taking this data for they have to actually understand either the number or the benefit itself or what is actually gonna materialize by providing another piece of data. But I think that's what consumers are.
Jennifer Overhulse: (
Looking for. So thus far we've focused on the data, collecting the data, what you're gonna do with the data, how you're gonna use the data, the context and what you're gonna use the data. That's a lot of fun. Right? So how are we gonna use it to personalize someone's experience with say Guardian Life. So they definitely, they want you to open the webpage and they wanna see hi Jennifer, right? They are accustomed to that from Amazon and so many other places. So they definitely wanna know it's their experience, but how fine tuned can you make that and how comfortable are people with it?
Rumana Hafesjee: (
Yeah. I mean, people are getting more and more comfortable. We ran a campaign where we asked people to kind of update their profiles. We never have enough contact details. Right. And one of the reasons we asked them to update, it was to say, if you ever are in a state where you need a claim, we can start sending you text messages about the status of your claim, which by the way, for us is like a number one pain point. And number one reason, people call us in our call center. And we had like a 12% conversion rate, which is pretty high for us. And so people are much more willing to give that information. If they know the context to your point as to what, so now you can actually get the status of your claims. They might say, Hey, Jennifer, by the way, your claim has been processed and it's going to be paid on X date.
Rumana Hafesjee: (
I bet you almost a hundred percent of people would want that type of text message. Right? Yes. And so I think that's where it really counts when you can personalize to that level. And it doesn't matter which channel you're in because that's the other thing that I think particularly with COVID happened overnight, this idea that you needed to personalize your experiences now, across every channel people wanted to interact in whether it was in an app, a website, whether they call the call center, whether they were on a virtual assistant, the expectation was almost immediate that, I want a personalized experience. And I want each of your applications to know what I may have said to the contact center today. And now I'm in your website and there should be some information for me too, that connects the dots. And so I think that's where we are heading when we think about personalization and it's really gotta be multichannel omnichannel across the board.
Jennifer Overhulse: (
So part of it for your company is also making sure that there's consistency across the channels that people are using, right? Yeah. So if I have an exchange with a chat bot, for example, and I don't know if you have one and then later on call in and talk to someone, my expectation would be or at least I would hope that I wouldn't have to repeat the same information. And that's where you're, that's across platforms. That's where you're going. That's using some of this data to,
Rumana Hafesjee: (
There's that element, which is you might be in the chat bot and then call the next day. But there's also the element is when you're in the chat bot, it's not gonna be able to answer maybe everything, like, for example, why was my claim denied? That's not a conversation you wanna have with a bot, right? At that point, what we're looking to do is get people straight over to a contact center rep. And that rep has the full history of the conversation they had with the bot. So immediately they might have to do a quick authentication, but then they know the exact conversation that's happened. And the reason they're calling is they wanna understand why their claim was denied. It's looking at it like, yes, I want a seamless experience across all these channels, but I also want within these channels to be super intelligent, if I move from one to the othe in the same conversation.
Jennifer Overhulse: (
Right. It's gotta be different in the small commercial space, right. Because you're using that data and you're interacting with a corporate customer. And maybe you eventually get down to a driver level if you're talking about things like that. So how are you using that data to personalize a customer experience or an instance for a customer and how far down do you take it?
Ray Lynch: (
Well, I think we also have to talk about like enabling agents and brokers to serve their customers and things like that too. So, I do think it's very important, what Ramona was saying around kind of that omnichannel experience and kind of being able to keep the left hand, knowing what the right hand's doing or not having to over repeat point cause that's gonna be it's time. You don't wanna waste your time to have to rehash or resay or redo something that you've already done. And I think that goes for agents that are trying to help their customers or for the direct policy holder. If we interact with them, big frustration point, it's less about like I'm a landscaper. So I wanna see a picture of a lawnmower on my webpage when I log into like a customer portal. It's more about just like that. I can, that the record is easily transferable across different channels, right? Just like, other experiences that you're used to, whether it's in your banking or where you might start with a chat bot end up at a contact center and then but they already know who you are. They know your account, they know what you called in for. And it just speeds up that whole interaction.
Jennifer Overhulse: (
Itay, from a regulatory perspective, there've got to be things that you're not that you're not allowed to use or that you're are kind of skirting the edges, maybe of things kind of pushing regulators. I don't know. Where are you in terms of a regulatory experience? What's the feeling on personalization? How far do you go?
Itay Bengad: (
Yeah, so to be honest on our end, I think we're pretty much on the very conservative side of the scale, because we generate data only upon accident. And this is where vast majority of people would be more than happy to share what happened, because what they're looking for the moment is help. Whether, if it's a, the worst case, it's a life saving matter. Or if the most vast majority of cases, it's a mild accident. But they want to be helped. And when they want to be helped, they feel very comfortable sharing everything that will start to move things forward. So we interact with those people about 30, 40 seconds after the impact itself. And they are more than happy to share every single detail. Obviously, in some cases it works a little bit against them. This is part of the system, right? We see people that are not covered in many other things but in general, we don't use any data that is pushing it, would say other than I would say at the moment, at least.
Jennifer Overhulse: (
So, would we all agree that more personalization is where the entire industry is going, as opposed to less, it's gonna continue to get more personalized. We're moving away from things like risk pools and more to customize coverages. Is that fair? Anybody wanna?
Ray Lynch: (
To a point, what point I think at the end of the day, I'm thinking of it as a standard PNC type claim, right? It's like it's property or it's a liability. I don't know if you need to change that, but you might bundle some additional type of coverages. And I think that's where, I mean, it's already happened, right? There's segment endorsements and things like that. I think what it is is giving more relevant recommendations of some of the additional coverages versus like a truly unique package of customization. I just think the expense to maintain file deal with regulators for like anything that unique is where I would draw the line. I think it's just more of like, let's help you agent either provide a better recommendation to your customer or business owner. Here's a recommendation of what you should buy based on these characteristics that you have.
Jennifer Overhulse: (
I like that. I mean, from the perspective of insurance, as a insurance, as a concept. I mean, what we aggregate and we homogenize risk in order to mitigate it. We spread it across groups. So when you make it so bespoke, it fundamentally kind of breaks insurance, doesn't It?
Rumana Hafesjee: (
And isn't that a great thing!
Jennifer Overhulse: (
I don't know, is it?
Rumana Hafesjee: (
I think so, look, I think we're an industry that needs to change faster than we have been traditionally. And our customers is changing drastically. If you look at the generation Z that's coming in and coming through, they don't have a lot of patience for insurers to treat them as a homogenous group. And so I think if we don't start thinking about things differently and disrupting, I think we talked about that this morning at the keynote speaking rate that we have to think about how we disrupt intelligently for that, that group, that generation that's coming in, because we don't particularly an insurer like ours, that's been around for 165 years. We don't wanna lose them to the new entrance, like lemonade and some of the other ones that are doing this really well, by the way. And that means we have to change the way we do things. So, yeah, I do think it's a great.
Jennifer Overhulse: (
Yeah, the younger generation, they don't know what a risk pool is for the most part and they don't care. The idea of it is even a little bit offensive. I think, yeah, it's like I'm not those people I'm me. I want my own coverage. I think it's, yeah, still the carrot, the stick they'll give you information they wanna know how you use it. They wanna reward for it. So, any thoughts out from there, from the audience, anybody have any questions that maybe for the panelists that, we could answer for you in terms of personalization? Yes, sir.
Audience Member 1: (
Yeah. Experience versus the factor, right?
Rumana Hafesjee: (
If I got, yeah. So, what we know is that 70% of customers feel really comfortable with personalization. If they've given you the data themselves, what becomes creepy is when we sit and we say something and our phone's listening and all of a sudden you start to see advertising against the thing that you were talking about at dinner with your family. So I think it's clear that when it comes to advertising, right, there's like a clear delineation of how far you can go in many cases, brands go too far. But I do think that when it's information that people have given you the ability to use it is sort of exponential.
Ray Lynch: (
Yeah. The creepiness is definitely in more of the distribution and the marketing or targeting of customers is where I see it, a bigger risk than in like the underwriting or pricing or kind of helping that customer with their need.
Jennifer Overhulse: (
Prefills, I mean, prefills a perfect example, right. I'm perfectly comfortable with prefill because, I see my address as public. Right. Those are the the things that are but you get yeah. To Ramona's point the advertising sometimes. Yeah. You don't, you assume your phone listens to you, but yeah. Anybody else it does though?
Ray Lynch: (
It does. It's actually scary that the data that you can get on someone,
Jennifer Overhulse: (
Yes.
Ray Lynch: (
Like some of the providers, exactly. It is you see the list of fields and you're like,
Jennifer Overhulse: (
How much do they really know? Yeah. Anybody else? Any other questions? Nope. Okay. Well, please join me in thanking our panelists. Hopefully we've given you a little insight into how far is too far in terms of personalization and you guys enjoyed the rest of your day.