the founder sits down with our reporters kim income mcnicholas, this week on press here. good morning, everyone. i'm scott mcgrew. if you use facebook to find old friends and linkedin to find new jobs, what do you use to find a copy of the great gatsby at 10:00 at night the night before the book report is due? you can run from house to house asking the neighbors or do the

very same thing on your computer using a new social network called neck door. it's kind of a craigslist meets facebook, a social network of neighborhoods. membership in nextdoor is restricted to people who can prove they live in your neighborhood. which means the conversations are ultra local. discussions about the new principal at the school or questions about why they're tearing up the streets and often warnings of a break-in that you'd never hear about on television or read about in the newspaper. >> both successful, occasionally controversial. he's the founder of next door, which is less than one year old and is growing quickly joined by tim mcnicholas of fortune, as well. let me start with this idea that i think facebook is vulnerable

here. there is -- you can post on facebook, hey, does anybody have a copy of great gatsby and my cousin in iowa can say, i have one. not helpful. this idea that i'm only posting to my neighbors and i don't think you necessarily need to know them, either. >> that's right. i don't know that facebook is vulnerable. i probably wouldn't agree with that. i think facebook has done an amazing job building a platform to connect to friends and family. but there's another opportunity in terms of a social network and that's connecting with the people around you. they may not be your friends one may not even know them, but they can be useful to you and you can be useful to them. >> if facebook were to some degree vulnerable, you wouldn't be in business. in other words, there is a missing spot here. you could post to craigslist and crazy people could bring you a copy of "the great gatsby." so there is a sweet spot there, if you want to use that instead. >> but i don't know any of my

neighbors. i don't know anybody's name. i havendon't know anybody on my block. >> next door is really not about being social. it's not so much about just knowing your neighbors. it's not about becoming friends with them. it's about being useful in your neighborhood and creating a stronger neighborhood for everyone. so, for example, you may not want to know your neighbors, but if someone broke into your car, you would want to ask the people who live around you if they saw anything. if they lost their dog, you want want to ask the people around you if they've seen your dog. so this is not a social thing that we generally think of. it's a problem solving network and it helps you solve problems in the area in which you live. >> one of the interesting things that i actually saw, i just signed up last night and created a profile and such. and i actually go running in my neighborhood.

what was disconcerting to me is i had no idea that there was someone hanging out that looks like trouble on one of the trails that i go on. i had no idea until i went on last night. that's beneficial. but my question is, so why wouldn't you create an app for that on facebook? why do we need a whole other social network for this? >> we have certainly thought about this. when we start today company two years ago, we thought maybe it would be better to take the distribution of facebook and it turns out that only 2% of your facebook friends are your neighbors. so you could create that app, but you wouldn't connect with anyone. the people that we friend on facebook don't necessarily live around us. we had to build something from scrat scratch. so we had to say people don't know the people that live around them. let's figure out an easy way to collect in neighbors.

and maybe it would be an easy way to communicate with these people, not as friends, per se, but as people who can help you solve problems. >> so how do people find out about it and sign up? if it's not your friends, then how does it spread virally? in it's a great question in fact almost 100% of our growth has been neighbors inviting neighbors. and we make it very easy for you to do that. we actually have a map of the neighborhood, which you probably saw when you signed up, and you can click on homes around you and we would send postcard invitation on your behalf. >> like actual postcards? >> real physical postcards. >> in fact, that's one of the ways that you verify. i can't just say i live at 123 main street. i have to receive a postcard in the mail or you can do it through a visa card because that will go to my home so that you know the person signing up for the social network lives at the spot he says. >> this is one of the most

unique things about right next door. we don't actually show anything publicly. you can only see what's said in your neighborhood if you prove that you live there, right? so there's no search engine optimization. there's nothing you can find on google in terms of what people say in your neighborhood. you have to prove that you live within those boundaries and then you get access. >> but still, you're not completely safe. your address is out there. some of the people that are lingering and looming on some of these trails probably live in that neighborhood and do have access to nextdoor. they may have criminal records. they may be a sex offender. is it possible for you to make sure that those people do not, you know, get into this community? because they can easily stock people. >> there are two things. first of all, the only thing you're required to reveal about yourself is your name and your street name. you don't have to reveal your exact street address. the sex thing is, when it comes to addresses of sex offenders,

we've integrated 350 state databases of those address and if you try to join from one of those addresses, we don't allow you to do that. so we've done a lot of work to make this as safe and intimate a place as we can. >> how do you make money? >> we don't make money yet. >> how do you eventually make money at it? >> we believe there's an opportunity to connect people who live in neighborhoods to local business that are in your neighborhood. >> do you need a sales staff for that, though? >> well, we've been able to launch almost 5,000 neighborhoods with about 30 people, and so we've always looking for ways to use technology to create more scale bble ways to create quickly. i don't think we would take the path of a yellow sales force, if you will, where you have someone in every region trying to call. what we've seen in our neighborhood today is that the businesses are actually joining, if they can verify that they

live in the neighborhood, and they're advertising themselves, self-service that way. >> that doesn't do any good for you. >> it doesn't today. >> how do you prevent them just from advertising through their own profiles? >> we probably won't prevent them. becausin' the people that you're communicationing with, if you're someone who comes in and you're advertising all the time or you come in and you create trouble, people won't respond to you. there's real reputation here. this is different than other social networks. this isn't about communicating with people that live 2,000 miles away or 20,000 miles away. this is about people you may see when you drive into your garage every day or when you're walking your dog. as a result, the behavior tends to be very civil and very collaborate. >> what are some of the models that you've checked into or that you're together around with aside from advertising? >> we have 100% focused on user

acquisition. we have not spent much time, if any, thinking about modernization. it's not because we don't think it will eventually become important, it's because we're trying to take the path on focusing on adoption and creating a good experience. if we're successful on doing that, we'll move on. >> thank you for being with us. >> thank you for having me. >> google won't say, but it's been estimated 1,500 ph.d.s work that company. up next, a company that could haufr 30,000 ph.d.s.

welcome back to press here. you might expect nasa has the smarter astronomers, the smartest actuaries. but those companies routinely ask for outside help from strangers.

when allstate wanted to know which cars hurt people most in car-pedestrian accidents, it took the data it had, the gig ba bytes and big ga bytes and dumped it on a website called caggle, where perfect strangers downloaded the numbers and competed to perfect an algorithm that would supply the answer, all in exchange for a cash prize of $10,000. there are dozens of these contests on caggle. nasa offered $3,000 to someone who would map dark matter. a grocery store offered $10,000 to anyone who could predict when super market shoppers will motorcycles visit and how much they'll spend. anthony goldbloom is ceo of caggle. he's a sailor representing australia in the world championships. and over the commercial break, you told me it's actually 50,000 ph.d.s working for you, as well.

when they come in there, there is a cash reward, but i am amazed at how much of it is just the spirit of competition. yet people will spend hours trying to figure out the data. >> well, that prize to the nas ka competition was actually a trip to the nasa laboratory. so there wasn't actually a cash prize in that case. so there are all sorts of motivations as to why the scientists compete. it starts off as an opportunity to get ahold of something that you might not get ahold of. you might be one data scientist in your company. >> there is a scoreboard. you can tell how good you are to everyone else. >> there is a scoreboard. and the scoreboard is crucial to why we get the results we do. so in every single competition we've ever run, we've outperformed the best that a

company has been able to do. and i argue that if you're a -- you know, you're a data scientist at allstate, you'll try a few things and you'll start. but if you're computing and there's a leaderboard, you get up in front of the person ahead of you and they'll keep working until they leap frog you. they keep leap frogging you until the increments get smaller and smaller. everybody is converging on the same level of predicting accuracy. there's so much signal in a data, there's only so much information in it. when you put people in front of each other, they compete null they've drawn everything out of the data set. >> do these people have full time jobs? >> interesting question. i would say the majority are in industry and there are some in academia. what we've rolled out in the last six months is what we call private competition. and private competitions are worth significantly more than the sorts of competitions that you spoke about earlier. so they're in the six figures.

and what happens is everybody in a private competition will invite ten or so for a private competition and everybody wins some money. so what we're finding is that this is becoming a source of full time income for the very, very elite data scientists. so we funnel them to private competitions both so they can rely on ceggle as a full time income, our very best data scientists, but it's also so that companies can take data sets private. that's an issue. >> that is an issue because i don't necessarily want to put all of my data out there. even if it's raw data where somebody is being given away, i'm giving this data to the smartest people in the room. they can reverse engineering, couldn't they sthp. >> the easiest information is you would give your data set to build a model on. similarly, you can give it to keggle's trusted competition data scientists. so the other issue, of course, is that if i build a model for

wells fargo and it's a credit scoring model, then bank of america might say, so this person finished fourth. he said before that they converge around the same -- to being around the same thing. so bank of america might go and purchase the fourth best solution. whereas in a private competition, because everybody wins some money, all the ip goes to the company that sponsored the competition so there's no possibility of any ip linkage. >> what's to keep a coop from not using you guys to going out directly engaging with the data scientists. netflix has done it and google has done it. why do they need to use you guys? >> well, one thing about keggle, our very best data scientists are starting to earn their full time income on keggle. as soon as they start moving off keggle, they're not getting the ranking points that are

required. so they're kind of blocked into us. it's kind of like, you know, trying to set up an alternative to the atp tennis tour. you know, roger federer gets invited to wimbledon because he keeps winning the tournaments that lead up to wimbledon. if you set up another, i think it would be quite hard to get roger federer play on another tour. >> what's your cut on this be, a percentage of the prize money? >> where we make our money is from private competition. we charge in the order of around -- so it's a six figure fee to host a private competition. half of that is prize money and half of that goes to us. and so the other thing we do is we host many of the algorithms that end up winning competitions. so sometimes it's not so easy for a company that doesn't have a very strong ip infrastructure to get the algorithm into production. we actually host it on our

platform. >> and the way you know somebody won is the algorithm makes a reasonable prediction hopefully every time? >> yeah. we back test our algorithms on reasonable data. the way this works is you want an algorithm on which borrowers will default on a loan. we'll take half of those loans or those loan applicants and we won't give you the answer, we'll test you. >> and see if your algorithms will do it. >> exactly. and we rank the accuracy on those historic flowing application. once we know which ones will perform best, it would be the future line applicants. >> i'm curious, here in silicone valley, we have a shortage of engineers, a shortage of data scientists. so how can folks here actually take advantage of all the data is in scientists that you have

versus burning them and they use a service by sorts. >> i have a somewhat vested interest in saying yes to that question. i think we're a much better option for the difficult problems. and the reason i say this is take a hard problem. there are any number of ways that you can approach it. maybe i'll give you specific examples. but we did a competition to say algorithmically, like high school essays from around america. you're trying to predict a grate grade. you hire a single data scientist, they'll try a bunch of things. it may be the case that they get you a good model, a not so good model. by the end of hosting a competition on keggle for that essay scoring competition, people have competed against each other, you've converged on

this, the best that can be done on that problem and you don't draw from that a sort of -- you know that that problem is solved and you can kind of move on to the next problem. you know that there's no point in continually revisiting that ath ga rhythm. >> anthony goldbloom, we appreciate you being with us this morning. >> thanks a lot. up next, morgan stanley smith has gone wild.

morgan stanley smith barn neefr has taken radical steps. the company allowed brokers and wealth managers to use social networks. as an experiment, 3% of employees were allowed to use

twitter. derek fowler is one of those three percenters, a private wealth adviser. he tweeted three times, retreated once. a recent tweet, quote, investors with high yield credits may return 4%, 6% in 2010. see page 10. derek fowler has 14 followers on twitter. i don't want to make you feel bad about how many followers you have. 14 is a darn good start. are you expecting to get maybe some more? >> i think i expect to get some more, yeah. to put this into perspective, social media and financial services, i've been lucky enough to deal with corporate executives, founders and early employees in silicone valley. >> so a high quality 14. >> a high quality 14. >> maybe if you tweeted a few more times, maybe you might get a few more followers, more than three? >> yeah. i think the idea is so we started morgan stanley in 2011,

a strategy to develop a social media process, right? when we first started with link linkin. but what i think we see with our clients is there are people here in the valley and they always told me, please stay technologically relevant, right? so for us, this is i think something that is more prevalent in the future. >> you called it a process and it is. there are probably companies that have policies that we have not read. but with brokers and advisers and financial, it's fairly specific, this process. i mean, it's not just wild on twitter any more. >> it's very specific. so if you think about trying to manage a workforce of 17,000 financial advisers all tweeting possible customers, right? we have strict regulations around the industry. so i think we're moving slowly but surely but thoughtfully. there's a reporting requirement. morgan stanley started with 600

financial advise remembers. we're rolling out this process more broadly. so i've been slow to move, but i think we started with linkedin. i hope to get more tweets. >> are there specific words that you're not suppose to use, not allowed to use? >> yes. at present, we cannot customize tweets. so i can't say a personal message. i have to have a scripted message. >> so you don't actually write any of those three tweets? >> not yet. >> you mean you choose from a variety of different tweets? >> yes. >> go local football team or whatever football game you appreciate? keep it nice and neutral. >> that's right. so i think that's coming, but once again it's the supervision of how do we do this but do it thoughtfully. >> so because it isn't customized, i feel the last

thing i want to do is necessarily present every single idea so i want to be targeted with us. >> is it your account or their account? let's say you leave, do you take that account with you? >> great question. it's my account that's overseen by morgan stanley. >> how do you prevent this just from being selling? ultimately, selling is the bottom line, but i wasn't come across that way. >> for me, i wouldn't look at it a different way. to me, it's another medium to provide financial and perspective clients. >> but you've got to be careful that's not just what it is. >> i'll give you an example. we're deemed market experts. if i make a comment on a company, either positively or negatively, and someone goes out and purchases shares in that company, is that construed a solicitation for sale? these are issues in the financial industry we have to work around. so we're moving slowly.

>> you guys, there's a reason why you guys are tweeting that. >> i'll give you stats. vit forum said investors actively use tweeting for usage and that directly influences their decision regarding financial assets. so i think it's critical. >> even larry ellison is on twitter, right? >> what if you decided you wanted to set up your own personal account, would you be allowed to set up something under your own name and be able to tweet what you had for breakfast? >> i suppose i could. if i did, though, i would probably keep it away from financial services. so for me, i'd rather use this as my centralized vehicle for twitter. we'll ease him into it. >> derek fowler is with morgan stanley smith barney and is

fairly new to twitter. give us your twitter handle so we can grow you from 14. >> dfowler@mspwm. you have at least three more followers today.

>> and you're looking at kim's vacation pictures. sth a mai tai conference. >> it's just a group of sill von valley investor that's share the passion for kiteboarding. they shared the passion to expand it. we decided this time to go to the george, which soefr in portland, so we could expand beyond kiteboarding, do some cliff jumping. >> you'll post this and mccall,

you've been writing about nokia. windows 8, got a chain? >> this is not as exciting as your job. windows 8, there was a lot of anticipation. i think there was actually a lot of disappointment after nokia announced these new lumia devices. but, you know, they're knot nice looking devices. it's just i think people wanted them to leap frog apple and android and it's not going to happen. >> that's our show for this week. i'm scott mcgrew. thank you for making us part of your sunday morning.