We talk with the CTO of Kiva.org, Sam Mankiewicz about how digital goals are measured and used in the development process. Kiva.org, a leader in the micro finance world, has loaned over $60 million to date and they are still growing. With this growth come many challenges that must be balanced by the tech team to manage an aging platform with increasing demand for innovation and measurement.
Speaker 1: This is Using the Whole Whale, a podcast that brings you stories of data and technology in the nonprofit world. This is George Weiner your host, and the Chief Whaler of wholewhale.com. Thank you for joining us.
Speaker 1: Well, I’m incredibly excited today because I get to talk about metrics. And any day I can talk about metrics, that day is a happy day. Welcome to Episode 20. We are talking with Kiva actually and their CTO, Sam Minkwich (SP), I’ve known him for a while. I’ve loved Kiva for even longer. If you don’t know Kiva, please be sure after listening to this to go check them out. They’re offering an incredible platform that provides microloans to underdeveloped countries and people that are, you know, entrepreneurs there as well as small grants expanding into the U.S. and other regions. Alright, let’s go jump into this interview. And by the way, I had to fly to California for this one. So please enjoy it.
Speaker 1: I’m here today at the Kiva offices, and I’m here with their brilliant CTO, Sam. Actually, can you just tell us what you do here?
Speaker 2: Sure. My name is Sam Minkwich (SP), I’m the CTO of Kiva.org. I’ve been with Kiva for over seven years now. And I’m responsible for the overall technology plan, lead the engineering and product development team here at Kiva. And sort of being one of the original technology member of the team, I’m also the institutional memory of how everything works and why did we exactly decide to build it this way five years ago now that it’s broken.
Speaker 1: So, “I’m here something breaks, let’s go to the internal repository which is, Sam.” But that is incredible to me because I understand it about Kiva. But could you actually tell us what does Kiva do in a nutshell and, you know, how does that relate to technology?
Speaker 2: Sure. So, Kiva is an online marketplace where anyone with as little as $25 can come online, find an entrepreneur or a borrower in the developing world that they wanna help support in their business, contribute online, and help see that business get the funding that it needs and might not otherwise have access to. And also, not only do you get your $25 back, most of the time, but you also get hear from the borrower and hear updates about the progress of their business.
Speaker 1: Pretty incredible stuff. And, you know, you’re coming up on 10 years of being around. You were in the game before it was cool, as it was the most cool, and now as people are, you know, taking more of a critical eye on it–I’m curious, what are the metrics overall that Kiva is looking at as, you know, as indicators of success?
Speaker 2: Yep. So, you know, Kiva first and foremost sees itself as a social enterprise. What we’re trying to do is give access to capital to people that don’t have it normally. And the best way to measure that is just, how many borrowers? How many people are getting access to that capital? And how much money are we able to put in their hands on any given basis. So, we just recently crossed the lifetime threshold of $600,000,000 in loans in our 10 year history. And we’re on pace for about $115,000,000 of loans to be made in 2014. And so that is millions of lenders that have helped fund loans for millions of borrowers all around the world. Right now we work with about over 400 different partners in 72 countries. I might have gotten those numbers wrong so you can check out our About page to see what the up-to-date stats are. But it’s a, it’s a huge network and it’s a ton of people. And it’s something where, I think we take it for granted today at Kiva the scale of which we’re able to operate. But there’s, you know, 10,000, over 10,000 loans get funded on the site every month. And when you take that to some of the other crowdfunding, or international development programs it just a magnitude more than what they’re able to do.
Speaker 1: Yeah. So, you’re just quietly becoming more and more awesome every single year, giving away millions of dollars. And I think we just, we can’t phantom this.
Speaker 2: It’s a loan, George, not a donation.
Speaker 1: Thank you for the correction. It is a loan probably a misconception.
Speaker 2: It’s something that’s a wrinkle in our model that’s a little different from how international development had done in the past. And I think that that’s part of what is appealing to the lenders, is that they, they realize that it’s a hand up, not necessarily a hand out. But then I can also, from a technology perspective, the fact that not only do we have to take your money today, but then hopefully give it back to you in six months, or a year’s time when the loan is repaid. It adds more complications and it makes the whole thing a lot more complex than, than it might otherwise be.
Speaker 1: Yeah. And I’m glad you kind of like jumped on that because donation in nonprofit. Just kind of like accidentally falls out of your mouth even though it’s definitely familiar. And maybe this would help a bit more. Let’s say like, I’m in Kenya, and like in Kibera and I need a bicycle. I’m an entrepreneur, and I’m like, this bike is expensive. But if I could get this bike, I can then run my business and selling stuff to tourists.
Speaker 2: Yeah. So, the vast majority of Kiva loans are, actually come through what we call field partners. So, it’s a local organization that will be based there in Nairobi, based in Kenya. And you would go to that organization and say, “hey, I’ve got this great idea for a business. I wanna get a loan.” And then that organization would be the one that would help to post your information up on the Kiva website. And in turn, when the loan was funded and money disbursed, they’d be the ones that would help get the cash in your hands, and then collect the repayments when they come back as well.
Speaker 1: So, you’re telling me, you want $25 of my money put through an online website that’s just gonna go to someone random on the ground in say, Kibera.
Speaker 2: Mmm. Mmm.
Speaker 1: And like, do you think that person is actually gonna pay me back?
Speaker 2: The, let’s see, I need the caveat that past performance is no guarantee of future performance. But our current repayment rate today is over 98 percent of, you know, all loans that are funded get repaid in full.
Speaker 1: That’s like almost better than like, well that is better than let’s say the U.S. housing market. If I were making loans out there.
Speaker 2: Yeah. The U.S. housing market has a lot more volatility but it’s definitely something where the cost of the loan might be a lot less than that $25. So, we actually have some of our most prolific lenders will, will actually put a whole lot of money into Kiva and leave it sitting there knowing that they might not be making the same financial return, but knowing that they’re making a lot larger kind of social return on those funds. And that, well, it’s not guaranteed that you’ll have all of your capital preserves. The historical performance is, is petty encouraging for that.
Speaker 1: And it’s about two percent that the return usually falls around?
Speaker 2: The default rate is between, it’s between one and two percent right now. So, the amount that you would, you know, lose on average.
Speaker 1: I think it’s easy to overlook the fact that like, oh yeah, all we have to do is just create a simple website where, you know, someone can find a great loan opportunity and just press a little button, you just probably throw it through PayPal. You know, really simple site, right?
Speaker 2: Yeah. I think a lot of the, let’s see, because not only do we have to collect the money from lenders upfront, but then also disburse it out to borrowers and then have money come back the other way around. A lot of that complexity comes from the last mile of I’m sorry, the last mile of disbursing the funds. And the other piece of it is that because of the small size of our loans, and because of the large number of lenders that might contribute to any one loan, we have a very small average transaction size. So while you might see a, there are $600,000,000 in loans from a financial services perspective, that might not seem like that large for volume. But when the individual transaction gets broken into $25 pieces, and your repayments might get broken into, you know, two or three dollar chunks at a time and then that gets spread across, you know 40, 50, 100 lenders have all contributed to the same loans, you end up with a whole lot of really small transactions. And so from a data perspective, we have a lot, a lot of really small transactions.
Speaker 1: It seems like a website you don’t want to, to crash.
Speaker 2: And, and it’s pretty important that we get all the pennies into the right spot. You definitely don’t want to crash and part of that, you know, that trust of giving us your $25 to then, to send to someplace over the world is that you also want to have confidence that we’re not gonna lose your money and that’s getting to where it’s supposed to go. And that we have some processes, and tools and frankly people in place to make sure that’s what’s happening
Speaker 1: This is a pretty complex system and obviously, I mean, I’m joking around with the whole PayPal, obviously. You know, you’ve built a platform over these like past 10 years with Kiva, and I’m curious, you know, what does your team look like, and what is your role as CTO of managing that team?
Speaker 2: So, the, the technology team here at Kiva there is a, there’s a component that sort of broken down by product area and who our different constituents are. So, probably the piece of Kiva that you’re most familiar with is our, our lender facing website. A place where folks can come and contribute $25 to, to a borrower all around the world. But we also have a portal for the field partners. So, those 400 plus organizations that need to share with us the information for each loan, process information month about who’s repaid, how much, which of those monies should go back to lenders or not, as well as to send updates out to the lenders about what’s going on with this borrower, or that borrower. And then we have a whole suite of tools to internally manage all of the operational, and all of those sort of admin functionality. And I mentioned the, little bit of, scale 10,000 plus loans a month. Ten plus million dollars of loans moving around every month. Literally millions of dollars both being sent around the world, as well as, coming back in to be redistributed. It’s important that we, that those internal tools aren’t just sort of an afterthought. But become a first class problem things that you want, that work well, that they’re efficient for the team. Frankly it’s part of our organizational sustainability and scalability that with a smallish number of staff we’re able to handle a lot of loans, a lot of partnerships, and a lot of money. We have about, Kiva has about roughly 100 full-time staff working mostly out of the San Francisco headquarters today.
Speaker 1: And 30 percent of those are just technology. They’re kind of under your wing.
Speaker 2: Yeah. About a third is in the, on the technology side.
Speaker 1: So, you mentioned that sort of like staff efficiency and when we were talking before it seems like 4ish buckets of your metrics. The acquisition metrics, of like new users, transactor, depositors, you’ve got the retention of how we keep people returning as investors and efficiency, like each staff member, if they were twice as efficient, you’d need half the staff and that’s obviously a metric of how well their managing partners and then the impact of “do these loans actually matter.” Thats a lot of metrics and obviously they definitely matter, but im curious when you close the door, and it’s your sprint meeting,and you’re like “here are the metrics that i care about, for the technology team.” How are you quantifying that, and using it to manage your team?
Speaker 2: Yeah.
Speaker 2: At the macro level, these types of metrics help not just the engineering team and the technology team prioritize their work and figure out what to work on next, but it’s really across the board. And that we’ve realized over the last few years that this, Cuba is a complex model. There is a lot of interdependencies here. And it makes sense for the different departments and groups within Kiva to try to focus on the same sorts of problems at the same time. For example, if one of our big goals here is to increase the efficiency, we won’t be able to support more partners, more volume for each employee. That’s something that’s not just for the engineering team to go and then make the product more efficient. But it can be symbiotic with say, the partnerships it’s going in looking at their policies and seeing which things they can simplify. Do we, is this the review process that we put in place a couple years ago. Is it serving the need that it’s intended to do? And so, because of the interdependencies, the metrics help us focus in on what areas we wish to work on and then each group or department comes up with some of the what, or how they want to, and try to improve that metric. And in the past when we were doing it in a little more isolation it lead to some really disjointed things where we might go and rebuild some tool and then three months later say, the user of that tool is like, well, we’re not actually going to use those things anymore. Can you build us a new one instead? And so, we wanna try to catch those things upfront and really be working with each other rather than just, you know, sort of prioritizing things in asylum.
Speaker 1: And so, that misdirected development, let say, because of the futures backed whatever communication issues came up, by having the metric lead the conversation, lead the feature or product, do you find that that helps focus your team and create in future proof development?
Speaker 2: No, I wouldn’t say, certainly not future proof. That’s a, future proofing is a hard thing. But I would say that it helps to, it helps to target the conversation on the why and not just on the what. So, going in, you know, lots of people have ideas about how you can make the product or software better. Let’s build this. Let’s build that. Let’s make the buttons blue. No, let’s make them yellow.
Speaker 1: definitely go with yellow…[inaudible 15:38]
Speaker 2: And so, rather than you fall into the yellow camp around here.
Speaker 1: No. I fall in the camp contrast by the way. Whatever has the highest contrast..
Speaker 2: Yellow and green just don’t necessarily go together quite as well.
Speaker 1: Contrast.
Speaker 2: Why, not what. And so rather than getting hung up on yellow versus green, or yellow versus orange, or, you know, contrast versus esthetically complimentary. Let’s start by really zeroing on why we’re gonna make this change. Is it to increase conversion? Is it to, you know, have less staff time involved in working on this thing? Is it to make the customer happier? Because at the end of the day, sometimes things that you do, it’s great where you would find the win, win, win where the future changes both, you know, easy to build, makes your staff happier and the end customer likes it as well. But sometimes, you know, those are not always in alignment ,sometimes you do something that’s easier for you even though it’s neutral to a little worse for the user. Or maybe you pick the thing that easier to build but won’t be quite as much as much as an efficiency gain. So, the metric helps you zero in and give you a common language to evaluate the worthiness or the fitness of the, you know, the future of the project that you’re considering. Then just asking that simple question of, alright, after in an you’ve been in the room for an hour you’re kicking around lots of different ideas about what is you could build to make this page or this aspect of the site better. To have that staff, we’ve got, you know, three best ideas up there. Let’s refresh ourselves about why we’re actually, where we’re gonna spend, you know, precious staff time on this right now. And sometimes that could be really illuminating. Like, yeah, three would be cool but that’s not gonna drive conversion at all. That’s gonna, you know, do something else along the way. The thing that we agree at a high level we need to fix this conversion this year, then we should be doing one of two.
Speaker 1: We’re talking about some of these new features and I know you did a very impressive thing as like you applied a tactic of a freemium model to buy a kind of thing. Can you talk to us about how that feature rollout worked and what some of the results from it were?
Speaker 2: From the early days, the way that Kiva work is that you come on the site, and in order to make a loan you had to put in at least $25. And while we had plenty of people that were willing to do that, a few years back when we sort of looked at our conversion funnel, having that high of a price point was definitely an impediment. And something that was popular at the time and still popular is the freemium model. Who can say no to free. And so we spent, I think, in 2012, although I might be proved wrong. We spent 2012 building what we call free-trial model. One of our very generous board members Reed Hoffman funded 40,000 free-trials for a total of $1,000,000. And the idea there is that you come on the site, you’d get to pick out, pick out an individual borrower that would receive that $25 but you wouldn’t actually have to put your own money in. And so, but you’d get all the updates from that borrower throughout the lifecycle of the loan. You would still get to have all of the Kiva experience with except having to pull out your credit card right at the beginning. And the hope was that those folks have the Kiva experience would become lenders in the future that were willing and able to spend their own money. And, you know, like a lot of freemium models it’s not 100 percent of those folks that will convert but, you know, some subset of those. So, we had some really, some early success with that particular free trial program and encouraged us to find other folks willing to put in money and do more campaigns in the future. But one of the, I guess, counterintuitive, one of the things that we saw in looking at the data was that those first couple campaigns had a conversion rate that was about twice what the 16 campaigns would end up having. And so, what in the beginning looked like it was a very lucrative, high ROI kind of program where we could continue to attract more people that would put in deposits into the Kiva platform, make more loans of their own based on this freemium model. And in later years, the conversion rates were about half as much and the ROI was, went from being like a slam dunk , yeah we should do it, to almost more like a, oh, this probably isn’t worth it actually. And so I guess both…what’s the take away there. So definitely, you know, if we hadn’t looked at the data after the fact, we wouldn’t actually see what the ROI was. And it’s prompted us to move in slightly different directions. We’re not promoting the freemium model quite a much anymore. And instead we’re trying to look after things that will either get people to become depositors right up front, or also, you know, look at things that are even in between that, somewhere between free and the $25 price point. How can we get people over that, over that hump?
Speaker 1: Yeah. So, just to like come back on this, the great thing here is, you know, of your buckets of metrics that you’re looking at. This comes under the acquisition model. You’re looking at new accounts created, meaning I’ve credited the account I’ve got in the system but I haven’t really done any sort of giving it a transactor, that is in your words, one of those 40,000 people say, have donated the money. I’m sorry, invested the money.
Speaker 2: Yeah. Invested.
Speaker 1: So, you’re trying to get those people into depositors, right? People that have put their own cash out.
Speaker 2: [inaudible 22:13]
Speaker 1: Fifteen percent was the…
Speaker 2: So, a new transactor is somebody who’s made a loan with not their money.
Speaker 1: Yeah.
Speaker 2: And we’re trying to convert them into people, into what we call depositors. Someone’s who’s made a loan with their own money.
Speaker 1: And you’re getting 15 percent there-abouts when you first did this project. But since you kept measuring it, that dropped in half to the point where you’re like, ah, you know, we’re not connecting in terms of getting the transactor into a deposit.
Speaker 2: Those early campaigns were doing about twice as well as the latter campaigns. And I think this was the one thing that actually made this difficult for us is that, the caveat is that we would see that conversion rate sort of eventually over the lifetime of the campaign. The typical loan on Kiva these days is about a year, a year and a half loan. And so, we couldn’t really, you’re most likely to turn from a transactor to a depositor once your first loan is fully repaid. And we kind of have to wait a year or two to see what the, you know, the final conversion rate might be for a campaign. And that’s a really long cycle time. And it really, frankly it sort of prolonged how long it took us to notice that the latter campaigns were not performing as well as the early campaigns.
Speaker 1: Yep. At least you were measuring. You didn’t just one time look at it this works now. It’s always gonna work. Let’s keep throwing Hail Mary’s, and tell the defense not to adjust.
Speaker 2: And it’s prompted us to move subsequent, how do I say this, subsequent manage lenders encouraging them to put their dollars in a different kind of lending experience.
Speaker 1: Yeah. And that’s really smart to say, you know, hey, Reed Hoffman you are brilliant. That was a great first step but if you’re gonna give us another $1,000,000, by the way, you hear this Reed hopefully if you’re listening, another $1,000,000, we can inform you a smarter way to do that. So, another feature you kind of danced around, you talk about this managed.
Speaker 2: I think frankly it’s not just, hey Reed frankly, we can show you this information but something that we’ve seen from funders and lenders at that level is that they’re, they’re sort of expect and or demand that sort of information today. To go in and they want to see in a quantifiable way what’s the total amount of their impact. Because Kiva is a platform with a lot of data on it and a lot of very precise actions it’s, it’s a, frankly it’s a cost of entry for us to be able to come back and give them very robust, aggregate information about all of the activity that their dollars have helped to fund. And so, it’s not really an option that’s nice to have, it’s a must have that the funders are asking for.
Speaker 1: Yeah. And I don’t think, you know, Reed is an outlier in that. I definitely see that across the board. But you talking about impact and so before we go into any other cases here, that’s another one of you metrics. So, let’s say I donated $25, you know, to five different people. You then come back to me and say, nice job George, your impact is $125. How are you getting to that next level?
Speaker 2: Yeah, that’s, I think that the short answer is basically, yes. You know, we present back your impact as having made $125 in loans to five different borrowers and here is a set of their stories that you’ve helped to be a part of. I think one of the, trying to blend those things together and resolve them back down to one number is one of the super quantify sort of way, is a challenge that development and international development overall over the years, is how do you, how much better is someone’s life because of whatever it might be. And so, you know, at its heart Kiva is about connecting people. It’s about stories. It’s about narratives. It’s about community. And so, we’ve sort of left our, our impact summation if you will at that level like, these are the action you’ve taken. These are all of the loans that have happened because of you. These are all the people’s lives that’s been affected in a positive way because of you. And havent really editorialized much beyond that. Now if you’re funding one of these larger campaigns, then what can get interesting there is both seeing the scope and the diversity, you know, across this many countries, these kinds of loans. You start to seeing patterns there in ways that individual five loan person you may not see there. And you can also see some of the network effects. The 40,000 people that joined Kiva because of you, because of your birthday campaign. We just launched this birthday feature a few weeks ago. They’ve come in and, those people have made this many loans, and this many dollars inside added to the system and it’s touched this many borrowers lives. And so, we try to, while you, you know, you wanna use your data and you want to present that in the best way possible. We also don’t want to dehumanize the data. We don’t want to take the people, we don’t want to take the individuality out of that as well. We’re trying to, you know, compare people or individual circumstances across the world in ways that don’t necessarily make sense. It’s a, like I said, in the start we have, we work with partners in 70 plus countries. Huge range of different circumstances in which these loans are being made. Hugh range of difference circumstances in which that quality of life, or sort of the kinds of benefit that they would expect to get. And part of the big umbrella of Kiva is that we try to have lots of different opportunities out there available. So that you as a lender can find the kind of borrower, the kind of loan that speaks most compelling to you. And every Kiva lender can have that sort of their own definition of what that best kind of loan is. But we don’t think it’s Kiva’s place as the platform to say that, you know, this kind of loan is better than that kind of loan.
Speaker 1: Now, hopefully, you will see why I was so excited to speak with Kiva and Sam over there. You know, brilliant use of metrics here because it’s not just freaking numbers on a wall, it drives me nuts when it’s like, you know, over one billion burgers served. One, you know, million loans, actually for them 600,000,000 loans provided. You know, they’re looking at fine tuning their marketing and how they, you know, you saw, should we continue this freemium model in earnest, or are we getting diminishing returns to the conversions we hope to see of getting new lenders out there. So, hopefully you can look at how you’re dealing with your own metrics and making them actionable, holding yourself, your marketing, your communications against those measures. And saying, how do we create more change at a faster rate? Alright, so this was one of two series of talking about Kiva. So, there’s still more to come if you like what you’re hearing. You know you can skip ahead to the next episode. As always, resources will be available at wholewhale.com/podcast. Thanks for joining us.
This has been Using the Whole Whale. For more resources on today’s show, please visit wholewhale.com/podcast and consider following us on Twitter at Whole Whale and thanks for joining us.