SitePoint Podcast #141: Pygg with Andy White
Episode 141 of The SitePoint Podcast is now available! This week our regular interview host Louis Simoneau (@rssaddict) interviews Andy White (@arcwhite) from startup incubator Pollenizer on their social payment startup, Pygg.
Download this Episode
You can download this episode as a standalone MP3 file. Here’s the link:
SitePoint Podcast #141: Pygg with Andy White(MP3, 27:00, 25.9MB)
Andy and Louis discuss the challenges of getting a social payment system up and running technically, and getting people to take it up and use it.
Browse the full list of links referenced in the show at http://delicious.com/sitepointpodcast/141.
Louis: Hello and welcome to another episode of the SitePoint Podcast. My guest on the show this week is a web developer from right here in Australia who’s the technical lead on a very cool little web startup called Pygg, that’s p-y-g-g, Andy White. Hi and welcome to the show, Andy.
Andy: Hi, Louis, thank you for having me.
Louis: How you doing?
Andy: Yeah, good, busy as ever. I’m currently migrating the Pygg service to a new Cloud Hosting environment with another local group, Ninefold, so that’s keeping me up until all hours.
Louis: Yeah, were you with EC2 previously?
Andy: We have been sort of, the company I work for, Pygg, is part of a startup incubator called Pollenizer, and we’ve used AWS and also Rackspace for a number of projects, Rackspace is Cloud, and we’re trying to rationalize around Ninefold because we target Australian audiences typically with our startups first, and just getting that latency down is a big win.
Louis: Yeah. So let’s maybe just back peddle a little bit and talk about what Pygg is. So do you want to just give a little bit of background on what this product is and what it does.
Andy: Yeah, sure. So I mean the elevator pitch for Pygg is that we’re trying to disrupt social payments, we’re trying to change the way that people pass money around to one another, and the long term goal is to disrupt the way people interact with their money and interact with their banks.
Louis: Cool. Let’s take it down from the elevator pitch to maybe the first floor pitch.
Louis: Which would be how to use it, what do I do with it.
Andy: So the current implementation is that if you have a Twitter or an email account you can signup with Pygg and you can send money to anyone else who has a twitter or an email account. At the moment it’s only in Australia but we do have long term plans to go global.
Louis: Right. And you say it’s only in Australia because the way you get money out is to an Australian bank account, right?
Andy: That’s right. And dealing with international currency conversion turns out to be quite a complex process when you’re talking about if I as a user pay you one U.S. dollar what does that mean when you receive it and how do we actually present that to the user in terms of the interface and the messaging. We tried an early experiment around that and it got very complicated very quickly, so our strategy has been to work on the Australian market first while we’re proving the idea, and once we’ve really got our legs underneath us then we’ll jump off and look at potentially the U.S. market or somewhere else.
Louis: Right. So the idea here is I send a tweet to Pygg or do I send a tweet out into the open or to the person I’m sending the money to in theory?
Andy: So, we run a Pygg account @pygg, p-y-g-g, and there’s a bit of code that sits there an listens to Twitter on that account, you send a tweet to that account saying who you want to pay the money to and the amount you want to pay, and that little robot will then give you a response to either say it was successful or to invite you to signup or whatever needs to be done, and the person who receives the money just as a result of the way Twitter works gets a notification as well because they’ve been mentioned in the tweet.
Louis: Right. And if they don’t have a Pygg account so that the next step for them is do they get another reply from Pygg telling them, hey, you’ve got some money in an account, that you can sign-in, create an account and get that out?
Andy: Absolutely. Which can be a little disconcerting for some users we’ve found, they tend to think that it’s some sort of spam setup if they haven’t been sort of briefed on what Pygg is.
Louis: Yeah, right.
Andy: But, yeah, we hold onto the money for anyone who doesn’t have an account, and once they signup the funds just appear in their account straightaway.
Louis: Right. In order to send the money I create an account at Pygg and I prepay with PayPal or credit card so that I have money in my account and then I can just send that away by using Twitter.
Louis: Right. Well, I can definitely see a use case for that in splitting dinner bills when everyone has only plastic and no cash.
Andy: That’s what we use it for internally quite frequently. Coffee runs are another big one; we’ve developed a new workflow around coffee runs where someone will basically just walk out of the office and say, “Hey, guys, I’m going for a coffee run,” won’t even stop to take orders, by the time they get down to the local coffee shop the orders have appeared in their Twitter timeline, they’ve got the money for them so they just order the coffees on the spot based on what’s in their Twitter timeline and bring them back.
Louis: Right, perfect.
Andy: No more dodgy Post-it notes.
Louis: (Laughs) Yeah. And no more excuses for getting the order wrong.
Louis: So, maybe dive a little bit into the technical details a little bit because I’m curious and I assume a lot of our listeners will be curious about the geeky side of the project, but maybe before that how long have you been working on this and how long has it been available to the public?
Andy: So I’ve been working on it since it would’ve been August I think, so we’re about getting close to being sort of six months in, a bit under six months in; I’m a programmer and my math is obviously not that great, right.
Louis: (Laughs) Just iterate a variable until you get to the right amount and then —
Andy: Yeah, exactly, that’s it, fail fast. It’s been available to the public for probably all but I think seven weeks of that, the first seven weeks we had only a very limited open beta, we were up and running very, very quickly.
Louis: That sounds pretty fast.
Andy: Yeah, the Pollenizer model is that we at the very least have something that we can test within two weeks, and typically we only have 12 weeks to get a business up and running and prove that it’s viable which is an extremely tight timeframe as I’m sure you can imagine. So the early iteration has been built extremely quickly on PHP, we’ve been using Zend which is potentially not my framework of choice.
Louis: That sounds terrifying. I don’t mean to offend any of the Zend guys in the audience but go on.
Andy: Zend definitely has its place in the world, a very rapid turnaround startup in my opinion is probably not the place for something like Zend, but on the flip side if that’s what you know and that’s what you’re comfortable working with then maybe that is the safest choice for you.
Louis: Well, yeah, especially if you’ve got two weeks to get something up and running, if you’re starting with a framework you don’t know it’ll take you two weeks to learn it.
Andy: Exactly. So, yeah, at the moment we’re just going through and we’re starting to pull the application apart, we’re rebuilding it with a service oriented architecture, so one of my big missions for the next couple of weeks is to pull most of the business logic out of the web app and into an API layer that will go sort of in between the database and the web server with the long term goal of eventually enabling people to write their own applications that will interact with Pygg.
Louis: Right. So you did mention sort of briefly there that the idea is you’ve got I think you said 12 weeks to prove that it’s a viable business, so what is the business model here, do you charge a fee per transaction or –?
Andy: Yep. So the actually transactions in between people are completely free and probably our hope is always will be, we charge a nominal fee when you put money into the system which I think at the moment is a flat $2.50.
Andy: And a part of that just goes to cover our costs in processing the transaction, and our goal at the moment isn’t even to make a profit, there’s sort of long term plans around profitability and how the business will work, but step one is just to make sure that we have something that people will use and that they’re happy to use and the business model can follow on from that.
Louis: Yeah. And how have things been going with users, have you seen a fair amount of interest?
Andy: Yeah, actually, a fairly significant amount of interest. Users tend to fall into one or two camps, they’ll take a look at it once and either because we haven’t explained it properly or because it’s not something they want to use or in some cases because we’re using PayPal will walk away very quickly or they get engaged with it and start using it quite extensively, and there are some really interesting groups of users up here in Sydney where one or two people have gotten involved and just fallen in love with it, gotten really passionate about it and then dragged all their friends along as well. So we’re definitely seeing in some cases the network effect which is quite heartening.
Louis: Yeah, that’s great, and it seems like the kind of thing where if someone demonstrates the usefulness of it and they use it to pay their friends then suddenly that could sort of spread out. I mean you see that happening with things like FourSquare which aren’t even particularly useful, so one would hope if something actually serves a purpose that people would be all over it.
Andy: Yeah, right. I guess the difficulty here is that it’s not always useful until you actually have other people using it, and so our mission for the next couple of weeks, or at least the mission of our customer development team, is to try and get the barrier for entry as low as possible. And long term we’re hoping an iPhone application might make that even easier, but we’ll see how we go on that one.
Louis: Right. So you talked a little bit about the technologies that you’ve been working with, so you built it in PHP and Zend, were there any particular challenges given that you’re dealing with money, you’re dealing with potentially things that people view as important from both a security and privacy point of view and that you have to interact with banks and with PayPal, what were the challenges that you guys faced or the most important things that you had to deal with?
Andy: Yeah, so early on you know we had to make some decisions around whether we were going to be processing credit card transactions ourselves, and you know we identified that as just being way too risky; anytime you’ve got to take credit card details or interact even with a merchant gateway you end up doing something that to get it right is quite complicated. So we’ve mitigated a lot of that by just pushing the responsibility off to other people and other organizations where we can to try and keep things nice and lean and agile, so we don’t actually store any information, I mean we store email addresses but we don’t store anything more sensitive than your bank account number which in some ways can be a little — I mean people feel a bit funny about putting their bank account number out there, but at the same time it’s very hard to misuse it.
Louis: Yeah, right. I think the only thing you can do with someone else’s bank account number is give them money.
Andy: Right. And, sure okay, great, give me more money; you can misuse that all you like. So a lot of — I wouldn’t say there’s been a lot of technical challenges around that, but it’s definitely been a process of thinking things through and trying to minimize risk exposure and stay as agile as possible. But, yeah, in terms of tracking the consistency of transactions, making sure that if I pay you $5.00 that’s what’s reflected in the database, has been quite a big part of my job, you know there’s a lot of testing that goes into a product like this because like you say you are dealing with money, and the second that we get a transaction wrong or money disappears out of someone’s account without an explanation to them that’s obviously going to be really bad for our users, and we definitely don’t want that.
Louis: Yeah. And in the case of things going wrong, you’re also in addition to relying on your payment processors, which are pretty nice and stable, you’re also kind of relying on Twitter; how have you found that in the case of duplicate tweets going through, has that caused any issues or things like that?
Andy: It did early on, one of our developers made an assumption that a tweet would always be, a) delivered in order, and b) never duplicated, and we’ve found that both of those things are not true. Twitter actually says in their API documentation that tweets aren’t delivered in the order that they’re sent, they’re delivered in the order that Twitter servers process them, so you have to be careful to write your code not to assume that they’re in any sequential order. Most of the time there’s enough of a gap between any of the tweets that we’re processing that that’s not a problem, but there are a couple of edge cases where things can actually get kind of hairy. Thankfully we’ve sorted out most of those but there were some late nights.
Louis: Yeah, I can imagine. What was the idea behind going with Twitter from the get-go rather than maybe building your own interface or allowing people to just use a mobile app, for example, to transfer money; why piggyback on Twitter?
Andy: Yeah, the advantage of Twitter is that you instantly get visibility to entire networks of people, it has a really well documented API that’s quite simple to use comparatively, it’s a restful API, it uses nothing but JSON, I think you can get XML out of it as well but I’d rather not.
Andy: I’d really rather not. And, you know, just getting that visibility early on was really important to us; if we’d built our own interface then potentially we would have been using email, and one of the assumptions that we went into the business with was that we wanted transactions to be public, we wanted people to be loud about their money, we wanted to challenge that conception that money and payments have to be a private thing, and I think so far that’s worked out pretty well.
Louis: Right. So you launched initially with Twitter and have since added email transactions.
Andy: That’s right. So as of I think about three or four weeks ago we pushed another update to our production environment, and that now lets you pay people via email which is a little bit more private and a little bit more quiet, but I think so far the vast majority of people are still using Twitter to send money around.
Louis: Right. We talked a bit earlier about the sort of obvious use cases of, for example, dividing up a restaurant bill or paying for a coffee run, you know, if you borrow five bucks from someone and you want to pay them back, what are other use cases that you’ve seen that sort of defied your expectations of what people were going to use the system for?
Andy: Melbourne Cup Day was a really interesting one for us.
Louis: Okay, this one we’re going to have to stop a little bit and maybe give some explanation to non-Australian listeners.
Andy: About Melbourne Cup Day, of course.
Louis: Yes, arriving in Australia that was a bit of a culture shock for me. So it’s a national holiday around a horse race.
Andy: Yes, that’s right.
Louis: (Laughs) So everyone has the day off and they all gamble on horse races is pretty much the given.
Andy: Ah, not everyone.
Louis: Not everyone has the day off, right.
Andy: Yeah, only some states in Australia still have that as a public holiday, but it’s very much culturally a bludge date, and that aphorism I’ll explain that one as well, it just means that people have a very lazy afternoon and end up watching the horse races and quite frequently drink copious amounts of alcohol, which is an Australian tradition.
Louis: Alright, now that we’ve got the background so tell us about how Pygg fit into that.
Andy: So we saw a lot of people making bets and challenging each other over Twitter, so there were a couple of groups that were doing — here in the office, for example, we were doing a competition to see who could be the best dressed, and the winner, I think the winner ended up winning some small amount of money, and another amount of money went to charity. So and other groups were getting involved in that as well and actually voting with a dollar.
Andy: We’ve seen a few people doing things like that where they’ll just throw someone a dollar and say, hey; I really appreciated that blog article you wrote, you know, here’s money for a coffee or money for a beer, so people were thanking each other.
Louis: Right, sort of a tip economy for online helpfulness.
Andy: Yep, exactly. And we’re actually hoping to see more of that because I think that’s really cool.
Louis: Right, have you seen anyone sort of plug this kind of thing into their website, sort of say if you like this article here’s a Twitter API integration with pay, whatever, a dollar or fifty cents?
Andy: We haven’t publicized that as well as we possibly could, but there are a few people out there who are using it.
Louis: Awesome. Have you guys encountered other people in doing a similar kind of thing?
Andy: There are a number of companies both here in Australia and especially over in Silicon Valley who are starting to move in sort of Bank 2.0 in the social payments sphere. We’ve seen Square, who I don’t really know how to describe their business model, but they recently released an application that does location based payments, so you pull up the application on your mobile phone and you select the store that you’re in, and the store owner sort of gets a list of people that are currently checked in at their store and they can pick you out via your photo and charge you using Square’s API. So that’s a particularly cool implementation that we’ve seen. And, yeah, there’s quite a few people doing similar things; I haven’t seen anyone else using Twitter in exactly the same way that we’re using it, but it wouldn’t surprise me to see some competition spring up very quickly.
Louis: Yeah. And you also mentioned that you do have eventual plans of moving beyond Australia.
Andy: Yeah. I’m not sure where we’ll move out to first, but there’s a bit more work to do, the application’s not quite at the level where we’d want to expose it to hundreds of millions of users at a time, yet; we’ve got a long way to go, we’ll get there don’t worry.
Andy: And once that’s sorted and a business decision has been made then we’ll definitely be looking at international markets, but that’s sort of above my pay grade, that’s something that the founders need to decide on.
Louis: Right. So you mentioned one of the things was you withdraw the money into a bank account via the bank account number, was there any discussion around potentially using PayPal for that, allowing people to get money out into their PayPal account?
Andy: There was early on, from memory it had to do with PayPal’s API and the way that we would have to integrate with them, they had some very stringent business rules around making payments back into PayPal. At the moment our PayPal implementation’s just a PayPal buy it now button.
Louis: Yeah, right.
Andy: It’s very straightforward, easy to implement, very quick to churn out and get up and running, actually integrating with their API in a more structured way, early on at least, was more than we wanted to do. So that’s something we’re looking at over the next probably four months we’re going to be expanding the options for getting your money in and out of the service.
Louis: Right, cool. So you mentioned sort of earlier on that the company you work for is called Pollenizer, and Pygg is one of a few projects that have on the go at any given time.
Andy: Correct. Pollenizer is at the moment I guess startup incubator is the best way of describing it, simply it’s a little bit more complicated than that, but that’s sort of the essence of it; we build minimum viable businesses.
Louis: Right. Is Pygg one of the first projects or is this, you know, or have there been a few in the past from coming out of Pollenizer?
Andy: Pollenizer’s been around; we’re coming up on our fourth birthday on the 29th of February. Pollenizer has built a number of different startups, obviously not all of them successful, it’s sort of guaranteed in this industry that you try a bunch of things and you run with the things that work and try to fail as quickly as possible with the things that don’t, but some past notable names include Dealized and Spreets; Spreets was sold to Yahoo for a sum of money that I’m probably not allowed to mention.
Andy: You know; it was a fairly significant exit and a big victory for Pollenizer.
Andy: We have been working for a while, fell into the trap early on of doing a lot of client based business, and we’ve recently started developing what we’ve come to call the Pollenizer way which is more focused on sort of the startup industry and building startups in the right way as opposed to as a client contractor, and that’s working out much more successfully for us; Pygg is one of the first businesses to be developed under that model.
Louis: Right. And is this something where they’ll be a team hired to work on a specific new startup sort of under the umbrella of Pollenizer but that team stays with that project or is there a core Pollenizer team that works on a number of different projects and then if those things spinout into successful businesses they have to take on additional staff; how does the organization of staff work in those cases?
Andy: That’s an interesting question especially as someone working internally at Pollenizer. At the moment the model is that there are three or four what we’re calling Pods within Pollenizer which are self-contained teams consisting of a customer development manager, a product manager, an engineer and usually either a front-end or a graphic designer, that one’s still a little bit nebulous, and under the engineer there will be anywhere between one and four junior developers working on a project as well, they may be here in the office or they may be located overseas or it depends on the requirements for a given project. Each of those Pods will work on two businesses at a time, typically two of those four developers or one of those two developers will be working on each of the businesses, and the other four members of the Pod sort of spend half their time between each of the projects. That model is probably going to be refined soon, you know, we’re constantly learning from the things that we’re doing and changing the way the business is structured to make things more efficient and more effective, but I have to get back to you on that one.
Louis: (Laughs) Right.
Andy: And once a business graduates to the point where it needs to leave the Pollenizer fold and go and be its own thing, staff traditionally have had a choice of whether to go out and work on the business or whether to stay within Pollenizer; I don’t know if that’s going to be the case going forward, but I think it’s valuable to Pollenizer to hold onto staff who’ve worked on a variety of projects as long as possible, and it’s valuable to the businesses that they build to hold onto the staff that built them, so there’s an interesting tension there.
Louis: Yeah. You’ve also written a couple of articles for the various SitePoint sites over the past few months, a lot of it’s been focused around mobile app development, so do you sort of split your time between working on backend server side code and, say, IOS and Android dev?
Andy: I am a serial over-committer and workaholic, and my personal time quite often gets used to work on whatever I’m feeling particularly passionate about at any given moment which has often been iPhone development, I’m working in a lot of stuff at home in my own time on that platform. Haven’t done a lot of Android stuff yet, though, I’m starting to get very tempted by some of the new handsets that are coming out, so if anyone wants to donate me an Android handset feel free, Google, ahem.
Louis: (Laughs) I don’t know if we have any listeners at Google.
Andy: Uh, you’d be surprised, maybe.
Louis: Yeah, it’s a big company. Alright, well look, it’s been great talking to you, and thanks for making a bit of time to come on and talk about the stuff you guys are working on, I thought it was just a lot of cool stuff and I was sort of intrigued by the platform and it seems to have come together very quickly. I didn’t have any clear idea of how long it had taken to put together, I just remember sort of vaguely hearing about it and then the next time I went to the website it looked to be fully functional and everything was there, I’m like, oh, what happened there.
Andy: I kind of feel the same way some mornings.
Louis: (Laughs) Yeah, I imagine. Thanks again for coming on the show, it’s been great, wishing you all the best of luck with Pygg and I hope to see more of your work on the sites in the future. If the listeners wanted to find you online or check out Pygg or Pollenizer what are the Twitters and the web addresses that they should be looking up?
Andy: So for Pygg you’ll want to look at Pygg, p-y-g-g.co, that’s pygg.co; the Twitter handle for that one is @pyggau, so p-y-g-g-a-u, for Pygg Australia, and my Twitter handle is @arcwhite, a-r-c-w-h-i-t-e.
Louis: Yep, I think that’s all good.
Andy: Thank you very much for having me on, Louis.
Louis: Thank you, it’s my pleasure.
Andy: Take care.
Louis: Take care. And thanks for listening to this week’s episode of the SitePoint Podcast. I’d love to hear what you thought about today’s show, so if you have any thoughts or suggestions just go to Sitepoint.com/podcast and you can leave a comment on today’s episode, you can also get any of our previous episodes to download or subscribe to get the show automatically. You can follow SitePoint on Twitter @sitepointdotcom, that’s sitepoint d-o-t-c-o-m, and you can follow me on Twitter @rssaddict. The show this week was produced by Karn Broad and I’m Louis Simoneau, thanks for listening and bye for now.
Theme music by Mike Mella.
Thanks for listening! Feel free to let us know how we’re doing, or to continue the discussion, using the comments field below.