Imagery that only your customers will know, with Matt Salisbury

I'm bad at passing KYC checks. I know the classics, mother's maiden name, first pet, colour of my first car, all that sort of stuff. But, chances are, so do the fraudsters. Where I struggle is the financial data that banks now use, I simply don't pay that close attention to what they name their products, I don't know if my wife having a secondary card means I have two credit cards or one, I don't know which email address they have as their one on record, and I don't remember when I opened my last credit card.

The problem with these events they want me to recall, is that they are once off tasks that have quickly faded from memory. But every day I travel through my neighbourhood. Actually, that's not the best example as I work from home and the kids school is a thirty second walk away, but I've just got back into golf, so I do drive to the range a few times a week and I'm not a complete hermit yet 🤣 Anyway, as each day passes and I remember less about the details of opening that credit card, I remember more about my neighbourhood. And that's what Honey Badger's geo authentication product relies on.

In this episode, I'm chatting to Matt Salisbury, co-founder and CEO of Honey Badger, who are creating user-friendly authentication challenges to stop fraudsters in their tracks. Matt loves to chat to lending people on LinkedIn, so feel free to hit him up, and tell him the podcast sent you.

You can book a demo by, or first learn more about, Honey Badger at https://honeybadgerhq.com/ - including direct jumps to their geo authentication and their mobile data for credit risk and fraud solutions

Honey Badger were also at the Future Identity Finance event in London a couple of weeks ago, and technically this post is speaking about that, but it also gives a wonderful timeline of that relationship with Amplifi Capital that is discussed in our chat: https://honeybadgerhq.com/future-identity-finance/

You can learn more about me on my LinkedIn page, my action-adventure novels are on Amazon, some versions even for free, and my work with ConfirmU and our gamified psychometric scores is at https://confirmu.com/ and on episode 24 of this very show https://www.howtolendmoneytostrangers.show/episodes/episode-24

Oh, and I'll announce it properly, but I'll be recording an episode of this show at Money2020 Europe in June so if you're going to be there, too, let's meet up. And if you're not, but you would still like to participate in the show, reach out to me at https://www.howtolendmoneytostrangers.show/contact-us

Regards, Brendan

The full written transcript, with timestamps, is below:

Matt Salisbury 0:00

Many organisations use approaches like 'what's the name of your first pet?' Or 'what school did you go to?' The challenge, of course, is that people forget the answers to those questions, and also a motivated fraudster can go ahead and find the answer in many cases through different methods.

So the approach we took was to say, what is it that we can present to you that isn't just memorable, but it's also frictionless, and in some cases even enjoyable for you to do?

Brendan Le Grange 0:25

So taai soos 'n ratel.

It's an Afrikaans expression that literally translated means as tough has a honey badger. That link may have needed a little more explanation, or it not for the classic YouTube meme, 'The Crazy Nasty Ass Honey Badger, narrated by Randall' that showcased the powers of this nearly indestructible beast to the world.

Welcome to How to Lend Money to Strangers with Brendan le Grange.

Well, Matt Salisbury, CEO and co-founder of Honey Badger, welcome to the show.

Matt Salisbury 1:13

Thank you, Brendan. It's wonderful to be here.

Brendan Le Grange 1:16

I introduced you there as the co founder of Honey Badger, which is true, but it's also an incomplete summary of your entrepreneurship track record. So, before we talk about how you are now ferociously fighting fraud - tricky one to have put there in the script - would you indulge me a bit on your background what you were doing before Honey Badger?

Matt Salisbury 1:36

Yes, sure. So I've had quite an active career, let's say.

I was born in the UK, in Manchester, and started working in tech soon after I graduated from university. I started in London, but then, after a few years, moved to Sydney and spent five years working and travelling around the Asia Pacific region. And throughout that time, I was working very much with technology, data analytics - smart cities being the focus area. But then I moved to San Francisco with my work and spent another five years there. And that's really where I got into the mindset of thinking more like an entrepreneur and tackling some of these challenges myself, rather than doing it for other people.

As part of my journey and time in San Francisco, I started a company called Ntropy, which was around customer experience around consultation and collecting survey data from people. And eventually, that was acquired by Civica, a large UK company.

So Honey Badger really is a second time round for myself and my team.

Brendan Le Grange 2:28

What is Honey Badger? And what was the gap in the market that you saw and want to do full with it?

Matt Salisbury 2:36

One of the first questions that people ask is why we called Honey Badger. And if you're familiar with the animal, it's very kind of persistent, it's intelligent, and in some ways, it's quite an aggressive animal that can really hold its own. And we had the mindset and the mentality that, when we founded Honey Badger, we wanted to fight on behalf of our customers to help protect them against fraud, and to give them the tools and support needed to fight back.

That was our mentality. And it became our philosophy. And hence, it became our name.

In terms of the opportunity that we saw: going back to when we founded Ntropy, one of the things that we did was create software that was used for surveys or consultation by governments with the community. And what we saw very much during that that work was overseas, bad actors would try and sway results. This isn't a new concept. We've seen it at the political level before, but it actually happens at the community level with overseas bad actors trying to sway the outcomes of surveys or just disrupt them.

It's very difficult in many cases, to prove that somebody is who they say they claim to be, especially when you don't have to factor in multifactor, or any previous kind of knowledge on that person. And so we saw an opportunity for technology that could play a role in helping organisations really prove who somebody is who they're claiming to be, but do that in a frictionless way.

Because there are technologies to help do that, but they're so disruptive to the customer experience that often they're not actually helpful. And so we saw an opportunity to be creative and innovative, and bring something to the table that could really play into that aspect of it.

Brendan Le Grange 4:08

As we said there yet, this is your second go at entrepreneurship. You were a successful entrepreneur before, you exited, you had some time being an advisor, consultant, and now it's the second bite. Now, I've not been a founder myself, but I've got two kids and I guess the first time you have a child, you almost do so without quite knowing the challenges that are there and that allows you to take on this new stage of your life. But then when it comes to the decision to have a second child, you're going in with your eyes open, in some ways better because you've learned a bit and in some ways worse because you know what's to come!

So for you, for Honey Badger, now about a year old, but what's that experience been like founding a second business going into the startup game again, with your eyes wide open?

Matt Salisbury 4:54

The benefit comes when talking to investors, raising money getting in support that you need, because you've already got a track record. You've got previous investors who can support you and are willing to support you. And that makes that part of it much easier. And just the fact that you're experienced how to navigate those initial complex situations. So that's the good part of it.

But then the other side to that is to grow a business, any business, you have to have a really solid product, you have to be able to convince people convinced the market to buy that and you have to address the problems that the market has. And honestly, that doesn't become any easier. You have to be on your game every day, listening to people having conversations, adapting, pivoting, in some cases, if need be, and continuously be willing to work and to adjust and improve how effectively you can do that determines your success.

Of course, I've got some of the same team members. So I know I've got people I can trust - that takes a little bit of the stress away - and certainly you're a bit wiser, but the challenges are still there. We had a saying with our previous organisation when we would go into kind of opportunities, and we'd have a sales pitch. And we'd do a proposal, we've had this saying where we'd say no news is good news. So if you haven't heard back from the customer, we haven't heard it know that it's probably good news, we've not heard back. But actually, what I can tell you now is that when you have no news, it's really just a delay before the bad news. And because of that I can take action and find ways to work around those situations. So we're not just waiting for a no, but it's things like that, where you learn that experience.

But then you ultimately have to adapt and go if you don't get any news back from a customer, what are you going to do to try and fix or correct that that opportunity?

Brendan Le Grange 6:26

Let's talk about those products that you're building this time around, as we said in the intro, they're around 'geo authentication'. But what does 'geo-authentication' mean? For those of us that might be on the user side, or those of us who have a online KYC problem we want to solve are looking for ways that don't involve your mother's maiden name and the the old favourites?

Matt Salisbury 6:48

Yeah, geo authentication is a really unique product for pretty much any financial institution, whether it's in the consumer finance or through to the traditional banking sense. At some point in time, you usually are in a situation where you need to prove somebody is who they claim to be. And you often rely on asking people questions that they should know the answer to: what's your mother's maiden name, as we said, or what's the name of your first pet? Or what school did you go to? what it was a whole variety of questions there, which are commonly used by the industry.

The challenge, of course, is that people forget the answers to those questions, and also a motivated fraudster can go ahead and find the answer in many cases through different methods. And you do actually see that in some cases, you get organisations like Equifax, that can go a little bit deeper, they can do what's called dynamic knowledge based authentication and maybe they can look at your bank statements and ask you have you done a transaction for this amount on a certain day? Or how much do you pay against your mortgage each month, etc. And that's a little bit better.

But ultimately, people often don't get those questions right, especially when it's on the spot.

So the approach we took was to say, what is something that you should know, that you should always know you'll never forget, that we can present to you that isn't just memorable, but it's also frictionless or very easy, and in some cases, enjoyable for you to do. And we took this concept of, okay, if you live at a certain place, or you know, a certain area, then why don't we go ahead and grab different images from that area. And we're talking Google Street View images, in many cases, these are images that aren't easily searchable, you'd have to walk around and try and get to know a place if you're going to do it yourself.

But we'll go ahead and we'll use our intelligence AI to go and find images from a location that people who live nearby should recognise. And then we put those images in a lineup with other images from different parts of the country that look similar.

And for a user, they simply click the images that they recognise. And we can then compare the results and see if they've passed that challenge. It makes that whole experience of knowledge based authentication, really straightforward, really easy, and really powerful in many ways, because we can also generate totally different images or totally different challenges each time. It's not the same questions over and over again.

In a nutshell, that's what geo authentication does.

Brendan Le Grange 9:02

I've ran afoul of the verification process before. Once, I think I've spoken about on the show before, but we were moving house, I paid for the movers, and the movers arrived and they got there and said we still haven't got the payment, we're not going to load the truck until that's sorted. And I had made the transfer from internet banking, but the bank had put a freeze on it and I hadn't noticed, and I thought I paid, I hadn't, the movers are there, it's a big rush. So I phoned into get the payment cleared. And it went through these things. As you said, the bank has more data, so it wasn't my mother's maiden name but it's things like 'what was the last transaction on your credit card account?' Or 'what day was it?' Or they ask you 'what is the email address we have on record?' And then you would I can't remember if I've updated them with or did I use my work address and you say, it could be this or it could be that, and they say we have to enter one in and then we don't know if you're right or wrong. Anyway, long story short, I failed the verification. There was a branch not too far from my house, I had to run down to the branch and show my driver's licence, passport and whatever else there to get it done.

And it was really a painful process, the questions I've been asked, were valid, but certainly not part of my day to day life, like, just the natural environment around me.

And I think especially true, if we think historically asked me what my last postcode was, or my last address, and I'm going to be a bit hazy, certainly my last phone numbers, no chance, remembering that, but you could show the streets or their neighbourhood that I grew up in as a child, I'm going to remember that. So I think as an idea, it's got a lot of potential.

I guess, one worry would be, obviously, as per there I've moved around a lot, and we've moved into the house we're in now, just a couple of months ago, s would that be a barrier would would I need to be well established in my neighbourhood to be able to pass these tests?

Matt Salisbury 10:51

Yeah, there are ways around that a challenge doesn't have to be generated based on your current address. For one, it could be based upon an address you've previously lived in, it could be an office place, somewhere, you've spent time during your childhood, whatever, different ways and different places you can generate challenges for that's one aspect of it.

Of course, if you are, let's say, creating an application for a loan, and we want to prove as part of that KYC process that you live at the address that you've specified, then we want to generate that challenge based upon the address you provide. Then there's a few things we can do, we can make challenges harder or easier.

We can tweak it so that we grab more distinct landmarks, for example, in many cases, we don't want to do that, because we don't want to make it too easy. And in that case, maybe you'll fail that challenge. And then we'll revert to some other technology to validate you. And that's okay, because that really is a kind of an outlier in this relative to the general types of challenges we do. But certainly there are things we can change on our side different ways of making it easier or harder.

And then really just another note to bear in mind is we have a patent pending on this and it it actually extends well beyond just using imagery. We can take over information such as maybe a local coffee shop nearby that you visited or a supermarket, we could ask you, how long does it take you to walk that hour drive there? Is that uphill or downhill? There's all these different types of derived information that we could, in theory generate that you only you should know. And so we can look at other ways of asking questions that don't necessarily rely on you having known the area for a long time.

Brendan Le Grange 12:22

Yeah. And what I like about that is and honestly, you're always living somewhere. When you think of history in the traditional sense, like 'when did you last open a credit card account' is a question I've been asked. And I happen to know that because I opened my credit cards when I moved to the UK - now I've given away that, so now fraudsters know my answers to that question - but in normal things, like 'when did you open an account' is a one-off thing in history, that the further along it goes, the harder it is for us to place.

And even if you're saying the last six months, how many accounts that you open, when something was open seven, eight months ago, nine months ago, or three months ago, that can be hard, because it's like a one off event that drifts further and further from your mind, but the neighbourhood you lived in, it's a continuous process where you're building these memories, maybe what we could even do is try that and do a bit of a demo.

I'm terrible at remembering these questions for verifications, so definitely I am somebody who would love a solution that is based more on real life events. That said, I'm also terrible with directions. So this is a good test, I think, if I can pass it, then you don't need to be a local genius.

Matt Salisbury 13:28

Yes. So let me go ahead then. And we'll do a live demo and see if you actually pass.

So what I'll do is I'll explain as I go for people that are listening in and can't see the screen, but essentially, I'm going to open up our demo interface here and type in a postcode that you provided to me. And what you'll notice is that the technology now goes away, and it's finding images from nearby to that address. And it's finding images that hopefully you'll recognise, and then grabbing other images that look similar from different parts of the country, and then putting them together into this kind of night free by free grid with nine images total.

Brendan Le Grange 14:01

Yeah, just for those listening, it's very similar in concept to the 'are your robot' checks: I've got nine images here, immediately, I can see middle left hand side grocery store, I used to go to a very frequently so I know that I'd say middle right as well looks like a couple of coffee shops on the high street, that would pass fairly often. And I'd say bottom middle is down the bottom of the high street where I used to finish a lot of my bike rides, so I can recognise him, but the pictures themselves all look very similar.

They're all from similar UK towns.

Matt Salisbury 14:35

I'll click submit that and let's see if you pass that challenge because we do have a time limit.

So hit go and you've passed that the good news is that in this situation, you did recognise the images from that postcode and you passed. And you can see from that, that someone like myself who has never lived there, those images would have been a wild guess

Brendan Le Grange 14:52

As you're setting it up, I could see there there's a toggle for how hard you want to be at so any loan situation you can make it stricter. If you're doing something far less risky for the business, you can make it easier but you've certainly got to stop and look, I said at the start, it's a bit like the are you a robot test - obviosuly, it's like a fire hydrant, it's not a challenging thing to do, you just click the things and you go, this does force you to look because it is similar, they could all fit together, you could easily tell me that those nine were from the same place that he would believe you.

So you do have to have that local knowledge. And you also want to feel slightly challenged. I mean, it's always a pain to go through the process. But you do want to feel secure, right? So you do also appreciate when somebody asks you a question that feels challenging, and I think that really does strike that nice balance.

Matt Salisbury 15:41

Now, of course, if you see an image that isn't particularly useful, for example, this top left one, you may that say that's not a great image to make a decision based off of, so you could actually go ahead and swap that out for a different image, and it'll grab a new one for us, it does happen from time to time, it's usually let's say, a large truck is parked in front of the Google Streetview camera or something like that. And we do have image classification that we run across these images. And we can filter out and see that there's not much detail in the image. So therefore, it's probably not very useful. That's one mechanism.

The other is we can actually add the option to swap images in and out and you would have maybe seen there's a button next to each image there. And so users themselves could say this isn't useful, swap it out for a different one, and we'd grab a totally different image. So we have mechanisms in place to counter that situation.

Brendan Le Grange 16:25

Yeah, a much better experience than sitting on there trying to remember, we ended last I opened a credit card account. Now, my first ever role when I started working, was in credit card fraud prevention, far enough ago that literally when the internet opened up as a channel, we just made it as painful as possible, so that consumers would go via our branch, like the old days, because we didn't have any idea how to restrict fraud. But the volumes were so tiny, that it was perfectly acceptable. Obviously, that isn't the case anymore. People have to have processes that work online and that work like everything else that we do.

And we obviously understand that in the banking world, we need to verify a little bit more than just can identify a fire hydrant, but we used to very fast processes to get things done not 'okay, now phone this number and tell me these things that you struggle to remember'.

When you are implementing the Honey Badger solution, in terms of a sort of a customer workflow, what does that look like in practice?

Maybe you can talk a little bit about how it's been implemented, how that fits in with some customers. And what that early reception been like. I think one of the hard parts of selling fraud solutions to some extent collection solutions is those tend to be the most conservative people in an organisation, the people that see what goes wrong, when people make decisions to benefit the customer for 20 years, the worst case scenario, so they tend to be a bit conservative. And now you're bringing this very new approach here.

Matt Salisbury 18:02

What we've seen, and I think the industry has seen over the recent years, is that there's a lot of challenger organisations out there now - whether it's in the finance space, consumer finance, or whether it's traditional banking sense, and these organisations are almost like startups, they are basically tech first organisations and they think and act very differently to those traditional banking institutions. And because of that, they are open to new ways of doing things. And especially if those new ways help improve the customer experience, because that's actually what they compete on, people go and open up new bank accounts with these challenger brands, it's the customer experience that really attracts them and keeps them banking with them. Because you don't have those limitations of having to go in branch etc.

And the other really big part of this, and this is how we've evolved. Amplifi Capital is customer of ours, they operate the brands My Community Finance and Reevo, which are loan providers. Now, we started working with them to stop fraud and using geo authentication to do that. But what they immediately latched on to and recognise was that one of the biggest challenges for loan providers is that people abandon or drop out of the application process, because the experience or the friction is too high. They realise that actually, they could use geo authentication as a way to perform KYC. And if people pass that they can actually bypass some of the more traditional checks, maybe document uploads and that kind of process, which is very high friction. And because of that, they can directly move the needle on revenue.

They're thinking about this not just as a cost to the business in the sense of investing to stop fraud. But actually, how do we invest to stop fraud, but also increase revenue? Because we're eliminating friction, reducing abandonment and getting more people for it to have loans etc. So I have a totally different mindset.

And I think because they're becoming so competitive that will actually end up forcing those traditional institutions to adapt to change or to shut down.

Brendan Le Grange 20:06

We're so used to such good service everywhere else that needs to catch up. And we give a little bit more particularly the older folk among us will remember the days when it was truly terrible to get a banking product. So we see the improvements that the industry is made, but certainly we also see where some of the other one click solutions are and we want to get somewhere in between at least.

Talking through the demo, there, one, the way that you select the images and screen those, and two, then you've got the image recognition, and three, you've got all sorts of analytics and technology sitting behind your solutions. And indeed, when I went to the Honey Badger HQ website, click the About Us page, you list technologists first among everything. So obviously, it's a tech first company, but what is your technology, your data and your analytics philosophy there?

Matt Salisbury 20:57

Yeah, I think first and foremost, our high level philosophy is that we believe data ultimately is a great way of stopping fraud and solving some of the challenges that we've talked about. But we have another side to that, which is really that data itself isn't enough. What we also need is domain knowledge and context. Otherwise, you end up with really a solution that's been built in a vacuum, which doesn't take into account the nuances of the different industries you're working with.

And the other part of our philosophy is that we partner with our customers to bring in their domain knowledge, and then build solutions which ultimately have data elements, best in class data and machine learning elements, with their domain and industry knowledge.

And that's what we've done with Amplifi Capital, is what we've done with other brands as well.

And I always say to the people I speak with, look, challeng us to solve a problem, let's work together and put our heads together, because we've got some of the best technologists, you've got some of the best data at hand, and also some of the best domain knowledge. And so that is our kind of high level philosophy.

When we drill down into how we then operate or implement that for our customers, as you'd expect, we're 100% cloud based, we use no virtual machines, everything's service based. And so we have a philosophy that is agile in the sense that we can build, prototype and deploy very quickly. But we can also scale very quickly as well. And we can have deployments in different countries and be receptive to our customers.

And then so we're very modern in that approach.

And then the final thing to mention it is, we track everything that we create, we look and we monitor, we look at how people use it, where do people struggle? Are there ways we can improve our user interface? Are there books that we can identify without people having to submit them to us for free forms, and then we feed that back into our product development and into our roadmap, and we fix those and we refine, and so we use data in that aspect for that continuous improvement cycle.

Brendan Le Grange 22:52

But Matt, we've spoken primarily about your authentication, it's not the only fraud defence of arrow that you have in your quiver.

So maybe you can also just talk a little bit about how you take that technology, how you take the analytics and the insights you have, and what other fraud tools are you using?

Matt Salisbury 23:09

A really, really fast growing area of our business is that what we've built around mobile intelligence, which basically means we can take a phone number and we can determine risk associated with that phone number.

It could be fraud, identity risk, it could also be in some cases, credit risk, and how likely is somebody to repay a loan based upon the attributes we see ultimately, what we have is integrations with mobile network operators around the world, for example, where we can go ahead and check as n is a number of being forwarded it overseas has the sim been swapped.

And we can even go to the level of taking some personal information, first name, last name, date of birth, and postcode or zip code, and comparing that against what the mobile network operators have on file. And so we've really invested in that part of our business, because there's so many insights you can get there. It's instant to do, but it's also frictionless. And the mobile network operators have some amazing data because it's often very up to date. So to have somebody let's say, apply for something, and then say instantly check and go does this match what Vodafone and EE or Verizon or whoever have on file, and then confirm that gives you a lot more confidence that you're actually dealing with the right person.

And then you can move them through that onboarding process much quicker. And then we orchestrate all of that together. So you've got that the KYC, the mobile intelligence, and then you've got the geo authentication part of that which could come into play if they fail one of those earlier tests, or you just want more confidence that it's who they claim to be. So that's really where we've invested recently as well.

Brendan Le Grange 24:39

A lot of interesting stuff here for people in that side of the lending game. If somebody wants to speak to you, Matt, or just learn more about the honey badger projects maybe wants to go and try or have a look and see that visual test in a bit more detail - where can somebody go to learn more?

Matt Salisbury 24:57

Welcome to visit our website, which is honeybadgerhq.com that has some material there, you can hit me up on LinkedIn, I love to chat to people on LinkedIn and just catch up and make connections there.

The other thing that I'm happy to do is introduce people to some of our customers that the institutions that we work with Amplifi Capital, for example, have great relationships there. So if you ever want to speak to somebody that's doing this and has done it and is getting the results and the impact and want, you want to learn from what they're doing, I can also make introductions to those types of people. And we do attend events. So there's lots of ways to connect with honey badger.

Brendan Le Grange 25:40

I'll put the details in the show notes as well. Matt, thank you so much for your time today.

Matt Salisbury 25:44

Thanks for having me. And I just like to say I love your podcasts. So please keep up the great work because it's very informative for people like myself, so thank you.

Brendan Le Grange 25:52

Thank you very much for those kind words and anyone else who loves to follow it on the players as well not just listen one on one, but yeah, thank you again, Matt.

And thank you all for listening.

Please do look for and follow the show on your favourite podcast platform and share the updates widely on LinkedIn where lending nerds are found in our largest concentration, plus send me a connection request while you're there. This show is written and recorded by myself Brendan Le Grange in Brighton, England. Show music is by Iam_wake and you can find show notes and written transcripts at www.HowtoLendMoneytoStrangers.show

And I'll see you again next Thursday.