Get in touch

Understanding an organization’s older technology systems: this week Paul Ford and Rich Ziade discuss legacy software and the work cultures around them. Topics discussed include how companies put systems in place and how they become unworkable, resistance to change; clashes between engineering departments and broader company culture; and tips for dealing with the social dynamics when dealing with — and trying to change — legacy software and systems.

Transcript

Paul Ford Hi, you’re listening to Track Changes, the podcast of Postlight, a digital products studio at 101 5th Avenue in New York City. My name is Paul Ford. I am the co-founder? Of Postlight and the co-host of Track Changes. 

Rich Ziade And I’m Rich Ziade, the other co-founder and co-host of Track Changes. 

PF Rich! Let’s tell people what Postlight does. Very quickly, and then get to the meat of the subject. 

RZ Postlight designs, builds, and ships software, oftentimes web-driven software: APIs, apps, mobile apps—we kinda run the gamut. We don’t discriminate at Postlight. 

PF All the things that you like to use on your computer [yeah], we like to build those things. 

RZ We like to make ‘em beautiful, usable, and scalable. 

PF I mean who have we done it for? We’ve done it for Vice Media. They’re a big client. 

RZ They are a big client. 

PF We relaunched a whole new platform for The Village Voice. 

RZ Mm hmm! Goldman Sachs. 

PF We’ve done some work for Goldman Sachs. 

RZ Barnes and Noble—

PF Barnes and Noble Education—

RZ [Jinx with Paul] Education. 

PF That’s right. So, you know, if you’re a really large brand lookin’ to get something done [Rich laughs], you can send us an email at hello@postlight.com. 

RZ Yes, I think that’s the end of our promotion. 

PF Alright! So, Rich, you know, we’ve been talking a lot recently about the concept of legacy, right? 

RZ Mm hmm. 

[1:36]

PF And a lot of people think of legacy as old code. Old data. Old sorta like [yup]—stuff that’s been sittin’ on a server for seven years and, you know, Sally knows how to update it but nobody else does. And you can’t really get something new on the website. And all the orders for some reason have a picture of a cat on the top right. 

RZ Yeah, you’re doing a lot of manual stuff. 

PF That’s right. Or there’s a guy whose job it is to put in a tape in at 5:45 pm everyday. 

RZ Yeah. 

PF And nobody else knows what he does. He reads the paper. 

RZ Right. 

PF I’ve been in environments like that. 

RZ Yeah, yeah. 

PF So [sighs] there’s a lot of things that make up legacy but we get called in a lot to deal with legacy situations. That’s one of the things that if you’re a software driven agency slash studio, that’s one of the things that happens. 

RZ Not to patch a thing—

PF Not to—no. 

RZ To build the new thing. 

PF Yeah, yeah, or to really build a pretty big layer that connects to it or all sorts of stuff [yeah]. And it’s probably worth a minute—before we even get to like what do you with all the data, what do you with all the code, which is almost another podcast than this one, for us to talk about—and this is something that’s really important to you—the social dynamics around legacy systems. 

RZ Huge. 

PF Because people react to software in their organizations and change is hard. 

RZ Yeah, absolutely. You know it’s worth noting that we call ourselves a digital products studio, and we’ve had this internal debate . . . a digital products studio makes it sound like we just build shiny little boxes. 

[3:05]

PF Well and I’ll give some context there. We borrowed that name from another digital products studio—somebody wrote an article. It was ustwo, they wrote an article about what is a digital products studio and it really defined us. Like it was like, “Oh ok, well we’re one of those two.” Like we’re very focused on product. We’re not a creative services agency that you come to around advertising. 

RZ Correct. 

PF And the tricky thing is that really what we are I think what you would think of us as an agency if you weren’t in New York City. Like you come to us, you say, “I have a certain amount of money; I’d like to solve a problem; I need to get a product built.” And you’d say, “Oh well they’re a product agency.” The word agency is overloaded in New York City by ad agencies: marketing firms; ad agencies, and so on. And there’s a real set of assumptions about the things you do and when we described ourselves as an agency, I couldn’t get through that, people thought that, “Oh, we’ll come to them to build the cool interactive game for the new gum brand.” 

RZ Right. 

PF And it was like—

RZ Which is not really us. 

PF It’s not us, we don’t do that. 

RZ No, we do deeper stuff: more complex stuff, applications that . . . you know an app, as complex as Uber or Seamless Web. 

PF You know how I would put it is we tend to work on problems where there’s—and this ties into the subject that we’re going to talk about—

RZ Yeah, we’re not continuing to sell now, just stay with us for a second. 

PF But this is the world we live in: to function as a firm like us you’re always thinking about, “How am I gonna portray myself?” The things we do tend to stick around for years, five years, ten years. We’re building for that. 

RZ Not against campaigns. 

PF Not against a campaign where you’re like, “We need this to just be killer for a year. We’re gonna put it on the side of a bus—” 

[4:43]

RZ Or it’s part of the Grammy’s. 

PF Yeah, that’s right. 

RZ We’re landing this right around Grammy’s time. 

PF See and no one would come to us—they could and we would do our best do it cuz it sounds awesome—you know some of the work we miss are things like, “At the Grammy’s we’re gonna do a visualization of all the different songs.” You know? 

RZ Yeah. 

PF “All the hot lyrics that were there, we’re gonna like do a word cloud that flies around behind the presenters.” People don’t come to us for that. 

RZ No. 

PF They come to us because they’re like, “Right now we have 500,000 subscribers and we need—we wanna get to a million. We have a growth plan but our platform isn’t functioning—”

RZ Or isn’t scaling. 

PF “It’s not scaling. You know, we’re seeing some real issues: it doesn’t integrate well with the newsletters.” 

RZ Right. 

PF “And so we’re wondering if you could help us migrate to a new API piece of software so that we can serve those million people and make our money and, you know, pay our mortgages.” That’s kinda where they come to Postlight. 

RZ Yes. And it could be really anything. 

PF Now the thing is is we make that sound less attractive and exciting but for us that is the motivating work. I like the kind of like dental pick like, “How are we gonna make this thing scale? What services do we use? What are we building?” 

[5:52]

RZ Yeah, it’s hard. It’s actually hard to market. We think a lot about how to tell that story. So, a lot of times people walk in and say to us—you know, sometimes they come in and they say, “I need that shiny new box. And it needs to light up a certain way and make certain noises.” 

PF Sometimes they say, “Do you have any PHP programmers?”

RZ Sometimes they say that but we’re not a body shop so we kind of turn that away. 

PF We like projects. 

RZ We like projects but very often people come in and say, “I need to get off this thing.” 

PF Sometimes they need to get off of WordPress onto something else; sometimes they need to get off of something else onto WordPress, right? Like it’s—

RZ Sometimes they gotta get out of Excel! 

PF Right, I mean, I’m talking too much about media but it’s much more broadly—it’s Excel; it’s Oracle [yeah]; it’s these big old databases that are floating around [yes]; it’s sometimes technologies that were really useful back in the day but—XML is a good example. 

RZ Yeah. 

PF Ok, I’ll tell people what XML is. XML is like HTML but a little more abstract and you can use it for any kind of data. So, in the same way you might markup a web page, you might markup a financial transaction [yeah] or . . . something with insurance. And there’s all these standards and rules and systems around XML but they’re gettin’ a little long in the tooth. 

RZ It’s dated. It’s hard to maintain. They systems that were optimized to run them aren’t as good anymore which gets into—the word we haven’t used yet is legacy. It’s—

PF Not in this context. 

RZ Not in this context. 

PF Meaning older systems that have been around for sometimes decades—often it’s as young as five or six years ago. 

RZ They’ve become outdated. So I wanna talk through sort of that backstory of how these things take hold. So, in technology, it’s kind of the shorthand way of saying, “The old system that is outdated.” 

PF That’s right. 

[7:54]

RZ And why does something get outdated. It’s various reasons. Usually when a system goes into place it’s reactive. You didn’t decide if a business is just rising, like is really starting to accelerate, you don’t say, “Ok, business, chill for a second. I’m gonna build the next five years of software for you.” What you often do is react [right]. It’s very reactive. So you start to build tools. And you start to cobble together these tools to just keep business going: to accelerate things; to help accelerate things, and it works for a while because your path is relatively narrow, right? 

PF Well it’s really specific, right? Like you might build something in-house not knowing that there are 200 things that you could buy that would do it better for you. 

RZ Correct. 

PF You might take the first sales call that comes along and decide to spend half a million dollars a year on something that does your logistics and shipping for you and it turns out that that’s actually like a 30 dollar a month problem. 

RZ Right. 

PF You know given this other system over here. And that’s actually—these are the deltas you’re talking about. Like it could be, “I’m spending 1.5 million dollars a year to support a system that I’m, you know, paying to this company, they don’t answer the phone anymore [yup] and I know that there’s a better way to do this.” 

RZ Well, they keep trying. I mean, what they’ll do is they’ll bandaid and create patches and add features because as the business is changing, like if they add a new product line and it has new data requirements, you’ll make changes to that system. So the system starts to get—has all this patchwork starts to go in—

PF Let’s use an example to help people out, right? So like shipping is a great example [yeah]. Something where I need to keep track of the people who are buying my product but let’s say it’s—

RZ Here’s a—I have one. 

PF Ok. 

RZ Inventory management. 

PF Great.

[9:43]

RZ And I ship—

PF Especially if it’s something bigger like sofas, right? [Yeah] Like, “I gotta get the sofas from Italy off the container, put ‘em somewhere, and then [correct] send them out to people.” 

RZ No, I mean stuff that can sit around. And let’s say you decide to expand your business and now you’re dealing with perishable goods. 

PF Ok so now—

RZ You decide to move into fruits and vegetables [right]. So now the data you need around that isn’t just, “Is it in point A and when does it get on the ship and then when does it get to point B.” It’s, “There’s a timing issue. And I need to know when stuff comes on and when it gets off and I need to prioritize around the stuff that’s coming in, is it gonna go bad?” 

PF And the example that we gave—when I threw out sofas makes it a little ridiculous but let’s say you were a chocolate importer or [exactly] importing like carob powder and suddenly you’re like, “God, what everybody wants is these tangelos from Belgium.” 

RZ Right, exactly. 

PF “I could make so much money if I could just move these tangelos quickly.” 

RZ Correct. So, you go back to your—let’s call them the tech group and they say, “Look, the system isn’t cutting it, I need to know when these things are gonna expire so I can order the shipments out, so I can organize them.” 

PF “I have 48 tonnes of rotten tangelos.” 

RZ Yeah. Dammit. So, what happens? They scramble. They don’t scrap the whole system. 

PF Well and this is really worth noting: these are people who are—and they’re building their jobs and their careers on the current existing system.

RZ They really are. 

PF This is the most awkward and painful part of the business which is that wherever you have a system where people are making their living, you have a group of people who avidly defend it as a sacred territory. 

RZ Absolutely. They get good at it. Expertise builds around it. It’s a system that gets ingrained into a business, not only is affected by the business and the culture and the sort of social dynamics, it itself actually influences the culture, the business, and the social—like it becomes a force. 

[11:44]

PF People are very dedicated. They’re very dedicated to their system. And it’s—look these are pocketbook issues. You go in and tell somebody, “We need to rip that up and put something new in and you might’ve just taken away not just their beach house but their kids’ education.” Like that’s what they see. So it’s very—you think you’re going in and saying, “Hey, we gotta keep these tangelos from rotting and we need to get a whole new system. I read about this one in Tangello Monthly that lets you track the hell out of these things. It’s great, every tangelo [integrate this]—ah this is great.” And they’re like—

RZ They do it fast. And that’s the other thing you have to do. The analogy I love to use is the G train, for those in New York City, it’s part of the subway system, will be skipping these eight stops Friday night through Sunday morning and then it’ll go back to regular service. They’re essentially trying to sneak in the work while the trains are still moving, and that’s often how it works. 

PF When I first moved to the city they would slow the trains down and you would like kinda—and then the guy would like to get away from the track, and you’d go by and then he’d go back and keep working. And that was taking weeks. And they were slowing all the trains down. So that’s why they do that, they’re trying to accelerate it as much as possible. 

RZ Things can’t stop. The message here is business can’t stop. You have to keep going while you’re building the thing. There’s another factor: people come and go. New managers come in who are really into mobile; people quit that had particular knowledge that you need to move away from. So, the technology is used, the programming language is used, the preferences shift over time [mm hmm]. So, when you look back, and there are other factors—we’re only covering a few of them—when you look back . . . at a legacy system that’s been around seven or eight years, it is really no different—you know, I’m not a geologist but you know when you could look at the rings of the rock to see what’s happened over time? 

PF Or just the layers. Rings of trees, layers of rocks. 

RZ Rings of trees—

PF Sedimentation. 

RZ Sedimentation. There is literally a history embedded in that code base. 

[13:50]

PF Well that’s the thing: very often—let’s say the carob powder importer has been around for 20 years and started by importing nuts, right? There’s a legacy there. They used to do it on paper. Then they did it with like a calculator. Then they got their first computer, it was like a PC. 

RZ Yup. 

PF It was like daily accounts receivable. 

RZ Yup. 

PF Then they got that first system. And that—that was a big—and like you’re on top of that. 

RZ Yeah, yeah, and a lot of times that’s reward. Look, business—

PF What does that mean, ‘that’s reward’? 

RZ That’s how you get re—meaning you can’t just stroll in and say, “Look, this is getting ridiculous. I need a couple of years.” 

PF Well you can’t. 

RZ You can’t, right? Because they don’t understand what that means and, “What do you mean a couple of years? I need this in 60 days because—” 

PF Everybody survives is by the quarterly—

RZ That’s right. Your performance, the favors you do, the quick fixes you do, the prioritization. Saying no is . . . responsible and politically very dangerous, right? 

PF See, and what you’re talking about here is this dynamic that is—it’s where the larger culture of business—and not just business—organizations of all kinds. NGOs, religious orgs, whatever. Meet the very special culture of engineering. Like remember there was a point where all the engineers in the world decided that deadlines were bad and started writing about that on Hacker News and [Rich laughs]—Like, “Deadlines have nothing to do with software.” 

RZ Yeah. 

PF And you would say that to somebody whose, you know, gota quarterly number of widgets to ship. 

RZ Yeah. 

[15:26]

PF And they’re horrified. It’s like you told them that nothing—that like Washington, D.C. isn’t real. 

RZ Right, exactly. So, this is life for awhile, right? And then—

PF It just gets—either the—well, there’s always the possibility that the legacy people will figure it out; adapt; and figure out how to support tangelos in the system. 

RZ And they often do. And then there’s are little fixes and little fixes and little fixes. 

PF And say yeah, like actually very often, like 70% of the time but that other 30%—

RZ Starts to eat away, right? And it starts to get really rickety and you’re eight years in, or 12 years in, sometimes your 30 years in, 25 years in, and then it finally—finally the leadership says, “Enough. It’s too slow. It’s inflexible. Our competitors have some system that is 130 times faster. It’s time.” And it never comes from the engineers. It always comes from the top. 

PF This is the subject that when I wrote that issue of BusinessWeek, the What is Code Issue. It was all about that, it was: can I explain the technology world to middle managers? Because this is all they see, they get so angry [yeah] with tech and tech doesn’t understand cuz it’s like just driving. It’s gettin’ stuff done. “Why are they so pissed at me?” [Exactly] And it’s because you’re sitting—they’re seeing—they’ve been given the goal—and I’m not talking about—like I’m talking about like a fairly senior technologist talking to a fairly senior like salesperson. 

RZ Yeah. 

PF No one—these are both fairly serious, well along in their career kind of folks [yup] and the tech people often are just like, “Well, you know, I’m doing my best here: I’m keeping the system up; I’m doing the work; I did actually install those three things that you said you really needed.” And the person on the other side is going, “I’m gonna get my ass handed to me in January.” 

RZ Yeah. 

PF “Like I’m no longer in a position where you can tell me no.” 

RZ Right, exactly. 

PF “You are putting my entire career and my life at risk, I’m already motivated to see risk, and you are sitting on top—” And the tech person is like, “What am I gonna do?! I’m sitting on top of this giant rickety pile.” 

RZ Right. 

[17:29]

PF “And I’m doing the best I can with second-rate people cuz nobody wants to come work at my thing.” 

RZ Right.

PF And—

RZ “Here I am.” 

PF “Here I am and you’re just gonna yell at me all day but the reality is unless you can free up a couple million dollars, there’s no way to fix this, and no one’s ever gonna free it up.” 

RZ And, right, exactly. And then every so often, it happens. Where—and I don’t know the difference, to this day, I think maybe I sort of do, I don’t know what a CIO is to this day. 

PF Chief Information Officer. 

RZ No, I know what it stands for. 

PF No, I’m just—most people probably have no idea what it stands for. 

RZ I think it’s the diplomat between the CTO and the CEO maybe? 

PF I’ll give you my explanation. 

RZ Oh geez. 

PF I think it’s really specific which is that especially in the late 90s there was—the idea of knowledge management as core to a large business took hold [ok]. So you know who’s amazing at knowledge management? A place like McKinzie. They have giant consulting firm as white shoe as you can get, just very well respected. They kinda run the world. And they were famous up until, you know, probably relatively recently, up until like five years ago, for having a Lotus Notes installation that approached like galactic levels of complexity [Rich laughs] with every—all your email; all your information; every piece of information about the global economy, macro and micro, that you could ever imagine; every report; every summary was sort of nestled away in this giant searchable database and it was created preweb. It was a giant information appliance. 

RZ Right. 

[18:59] 

PF And so what people saw was that that level of knowledge brought about by information technology almost as the internet was coming along, they were like, “This is a tremendous advantage.” 

RZ Ok. 

PF “If I can have dashboards; if I can have information [got it]; if I can have this kind of insight into my business; and this historical record [right], I can do amazing things.” 

RZ Ok. 

PF And that is very different than keeping the servers running; that’s very different than getting everybody their email—

RZ Purchasing—

PF It can be part of it. Yeah. That’s the Chief Information Officer. 

RZ Got it. Ok. So they’re not necessarily technical, they understand just enough about technology to sort of tie to business acceleration and success. 

PF I think it’s just because knowledge management is two letters. 

RZ Got it. 

PF So you don’t wanna see K-M-O. 

RZ Right, you know, I just learned something, right here on the podcast. So I understand it now. The CTO is more about execution

PF That’s right. 

RZ And look—

PF Well and it gets very blurry cuz the CTO at a big technology firm—

RZ Is doin’ more than just executing—

[19:58]

PF Yeah, is often like—like, you know, the CTO at Amazon was someone who was executing and building Amazon Web Services which is a core business so—

RZ It’s blurry. There aren’t perfectly clear lines. No doubt. So, let’s accept that the C-suite, the higher ups, the higher-higher ups have decided it’s time. They gave them the money; they might’ve hired someone, that’s also likely. 

PF My thesis is that there’s learned helplessness on both sides when this is going on like the people on the non-tech side have gone, “Well, I can’t get it out of them and that’s how it is and they keep telling me.” And then the person on the tech side is like, “I have to work with what I got. I got this, you know, Peopleware installation. Um I have—” 

RZ “My cue of stuff that they want—” 

PF Yeah. 

RZ “—these little bits . . . is endless.” 

PF “You’re telling me to fix the logistics in the factories [correct] and at the same time I got a Jim over here telling me that he wants a new web platform for all the salespeople to be supported by apps.” 

RZ Yeah, everybody’s coming at everybody—

PF “I need another eight million dollars.” 

RZ Correct, so—

PF So everybody’s helpless and then somebody’s like, “I’m actually not gonna be helpless anymore.” 

RZ And it’s power, right? [Right] The power dynamic kicks in where they say, “Ok, I’m going to support you. [Mm hmm] Paul, you have been blessed as the person that’s going to bring us the modern platform.” 

PF Well and let’s be clear too: this is business. Power is the ability to open up a checkbook. 

RZ Yes! 

PF Or to get someone else to do that—

RZ You got the budget! You got the budget. 

[21:30]

PF So you can scream and yell at your tech people all day long but if they don’t have the money to do it, they’re just gonna shrug and walk away [yeah] but if one day you show up with a checkbook and they still say no, that’s when the major legacy situation starts to show up. 

RZ Yeah, exactly. And so, let’s be optimistic for—let’s draw an optimistic scenario where you’re given 12 million dollars and you’ve been told, you know, “Give me a plan.” Actually, you’re not given 12 million dollars. They tell you, “Give me a plan and give me the costs of how we’re gonna get off this thing.” And you come back and say, “I need three years; I need 40 people.” And so the math gets done and they allocate 12 million bucks so you can get it done. This is big. 

PF Now if you work in financing, that’s actually not an insane request. 

RZ No, no. Absolutely. You wanna stay modern, you wanna be competitive. 

PF Like we live in a world of little zippy apps gettin’ developed in three or four person teams but that’s—

RZ Oh these aren’t unusual numbers. 

PF Everything you’re describing, when I’ve worked on bigger things, and especially when you talk to people in just large industries, [yeah] finance, insurance. 

RZ Totally normal. 

PF Yup. 

RZ So, they sign off; we’re ready to go to work. Now, here’s what . . . almost always happens. What almost always happens. The ideal setup here is to build a hanger outside of the office building [mm hmm] where this gets to happen and it’s insulated by the winds of business because the pull of business and day to day and month to month business needs is incredibly powerful. I call it the hijacking. What happens is you thought you had your 30-person team, right? And then somebody really influential in sales walks over to you cuz they just know as the tech guy, right? And says, “I need a couple of things [yeah]. And I need ‘em in a month. I’m sorry. I really do.” And this person, they could pick up the phone and have some really influential people bring pressure, right? 

PF You need to get out of mind. 

RZ You need to get out of mind. 

PF Out of sight, out of mind. 

[23:32]

RZ Rarely does an organization have the discipline to truly cut that out and say, “There’s a wall here. Do not talk to this team.” 

PF It’s also really tricky, right? Cuz the people who are on that team put themselves in a position of great risk where if they don’t succeed, they’ve been out [oh yeah]—they’re associated with the success of that project. 

RZ Not only are they associated with the success of that—That project is, for awhile, and that’s inevitable cuz you need time. It’s disassociated with business. 

PF And everybody’s rolling their eyes at it—

RZ Yeah like, “What are you doin’ over there?” 

PF Yeah, “Over there on the princess farm doin’ your nonsense.”

RZ Exactly [laughing]. Exactly. 

PF I mean the archetypal example of this was the first Macintosh. They called it Texaco Towers, it was a building that had been a gas station that Apple had acquired. And like the team was over there, in a remodeled gas station office thing. And they hung a pirate flag. 

RZ Oh really? 

PF Yeah. And were like, “We’re gonna go build the Macintosh,” and there was actually already this effort called the Macintosh, they completely hijacked it and just sort of—that’s where the Mac came from. 

RZ So they really, truly separated themselves—

PF Cuz otherwise they were being asked to kind of like, you know, “Let’s further the Apple II vision.” 

RZ Right. “Fix this. Let’s get a new version out or whatever it may be.” 

PF Right. 

RZ See that’s—that takes—I’ll tell ya: one of the big advantages of an outside shop is we’re outside. We’ve had product leaders and decision makers come to us and say, “I need to do this not in my—inside my walls.” 

[24:59]

PF This is what happens, it’s often—and what happens is people get a little budget freed up and they’ll be,  “Go solve that.” And, you know, she’ll come on down to Postlight where she’s talked to us a couple times and this EVP of whatever, she’ll go, “Look, I can’t get my side unlocked because they’re on like a 36 month strategic journey and I have this thing I need in 8 months. And what about you guys?” And we’ll be one of, you know, three shops that’ll pitch because she can’t unlock her own resources. That’s very normal too. Like this isn’t—it’s important to articulate to people cuz people don’t often—I remember myself growing up I never had any experience at this altitude. This is just kinda how it works. Like the fact that—it’s with the blessing usually of the CTO, like, “Oh, you gotta go outside for that. We’re not gonna get that for you.” And there’s some eye rolling, there’s some like, “Yeah, it’d be really nice we could but uh this place can’t handle that.” Like people in higher altitudes at orgs tend to know the limitations and [sure] where things are going and they’ll go like, “Go get one of those shops.” 

RZ Right, right. And—and it’s wise because not only are we gonna be less distracted by the business, we can’t be influenced by it very much. Nobody who I have not met at Acme is gonna pick up the phone and say, “Hey! I heard you’re doin’ this other project for uh Diane, I needed some help with this other thing.” He doesn’t know who I am. He’s just not gonna walk in and say, “Can I steal some of this resource?” They won’t do it. 

PF There’s another thing going on. You know, there’s an understanding of business where people talk about the formal system versus the informal system. The formal system is the set of rules, like, “Go talk here, put in this budget, and then in two months we’ll tell you whether the work should go forward or not and then we’ll staff bah bah—” It’s, you know, like it’s a guide book essentially. 

RZ Yeah. 

PF Here’s the hierarchy in the business; here’s your title; here’s how things happen. And then there’s the informal system which is, “Go down the hall and talk to Mike and actually if this thing is gonna happen, he’ll get it done for you.” 

RZ Right. 

PF And that can be really—it’s so—Big orgs are incredibly informal. They’re very like—there’s a way it’s supposed to happen and there’s five million Mikes. 

RZ It’s truly a social network. 

[27:04]

PF And so what happens is that unless you have real ultimate power and even if you do, even if the CEO says, “Let’s change this,” things won’t happen because if somebody has the power to drag their feet and they’re not feeling it, they’ll exercise that power and often the technology team does. 

RZ Or yank you and in a particular direction. Like it can come from the technology team, it can also come from the business side that just keeps pokin’ holes at this project. They just can’t help themselves but say, “C’mon with this 18 month thing, I just need this next week. Can you just do this for me?” 

PF That’s right. 

RZ And that distraction is incredibly costly. 

PF The other thing that happens internally is that it’s very hard to get software built in general, so there’s a lot of, “Let’s do another deck; let’s make the business case; let’s do this thing; let’s do that thing.” And that puts you in a position to just get nibbled to death. You can just do proposal after proposal internally

RZ Yeah. 

PF And people will come for you to keep you from getting your thing built because they’re worried that they won’t get to build their thing. So one of the reasons we exist is that you can come to us and say, “I gotta get this done.” And that’s one of the reasons that agencies exist. Agencies are kind of unholy like you should be building your own stuff, by yourself, inside your own company, right? Why would you go out if you didn’t have to? 

RZ Right. 

PF But sometimes you have to. Cuz we’re incredibly motivated just by the dynamics of the business, we have to ship this thing for you. 

RZ You’re firing us every day. 

PF Yeah, that’s right. 

RZ We don’t work for you; we don’t have like connections to that guy I always have cocktails with whose got the influence and the budget. We don’t have any of it! 

[28:36]

PF You’re spending money on us, it’s ultimately in your best interest not to spend that money if there’s any way you can avoid it. 

RZ Sure. 

PF That’s the agency business, or the studio business around product. It’s different with advertising but with what we do it is very intense. 

RZ Oh yeah. I mean this kind of business, and again not to pat ourselves on the back, it’s hard! You’re constantly under threat because that’s the nature of the business. 

PF I mean I think I’m probably one of the best businessmen who ever lived, really. 

RZ I think you might be. 

PF [Chuckling] It’s very clear when you look at my life. 

RZ Paul, we’ve been complaining for the last 25 minutes. 

PF I don’t think we have! 

RZ I wanna give some tips. 

PF This is the thing: we’ve been talking about the actual institutional dynamics that get into software. 

RZ These are real. Humans are a bigger influence on software than software. 

PF That’s correct. 

RZ The tools, the code, the platforms, the libraries, the frameworks. Humans are the biggest factor in the success or failure of software. 

PF I would say also that’s like especially because the way the industry is going is I can tell you what you’re gonna build, no matter what you’re gonna build. You’re gonna build something that, if it’s gonna be on mobile, it’s gonna look and use the Android toolkit or the iOS toolkit. And if it’s gonna be on web, it’s gonna follow like four or five best practices and do this thing with Javascript and this thing over here. I know what you’re gonna build, it’s all down to like that final 20%. That’s where all the work is. Where it’s just your business, just your idea, just your design. 

RZ And that’s where there’s a lot of vulnerability. 

[30:06]

PF Yeah! That’s right. 

RZ So should we give a couple of tips? 

PF Yeah, my God, we made people listen to our big story. 

RZ This is pretty rough, yeah. 

PF Yeah. 

RZ So tip number one—

PF Well, wait, before you do that—

RZ Oh ok. 

PF [Chuckling] Sorry. Before you do that, tie this concept back to legacy cuz when people think legacy they think, “I have a big Oracle Database,” or, you know, “We’re doin’ something with Filemaker from 1998 [yup] and we need to update it where I have a thing I have to type into an orange screen.” 

RZ Yeah. 

PF So what is all this drama that we’re talking about have to do with legacy? 

RZ Well, I mean, at some point the switch is gonna be flipped and you’re not gonna use that old system anymore because—

PF [Jinx with Rich] Because. 

RZ Because of all its limitations, its speed, its functionality, all the things that have become such pain points for that system. 

PF Well the real trigger to me is when it becomes a noticeable negative impact on the ability to grow the business. 

RZ Exactly. That’s the boiled down rationale for getting off a thing. 

PF If your business isn’t in a growth mode [oh!] you’re just gonna keep—you’re gonna use that VT320 terminal until it catches fire. 

RZ Yeah and it’s not harming you. Like if the blocker is technology, you’re gonna go after it and you need to get off. And then the getting off is huge. It’s a big deal. We haven’t talked about data. 

[31:32]

PF We should! 

RZ That could potentially be its own podcast. 

PF I think we should. We’ve gotten—if we’re talkin’ about legacy and we’re making this a little bit of a theme, we’ve talked about the social dynamics that bring up legacy issues. 

RZ Yes. Maybe we should end this with some tips, Paul. 

PF Tips about the social dynamics. 

RZ Social dynamics. 

PF That’s right. 

RZ Yes. Tip number one: don’t hide. Expose and share the progress. 

PF Who are you talking to? 

RZ I’m talking to that tech group that was blessed with that big budget to do the new thing. 

PF Ok. Don’t hide. 

RZ Don’t hide. Don’t say, “See you in 18 months.” The more people feel like they are somehow influencing this big, big endeavor, the more they’ll actually start to become defenders of it. 

PF So you need to get people invested in your progress. 

RZ Exactly. 

PF It’s gonna be good for them. 

RZ It’s gonna be good for them, and if you drop, “Whaddya think?” into the conversation the dynamic is—it’s no longer, “I’m about to force this down your throat in a year.” Like it’s more like, “Here’s how it’s progressing; we’re feeling pretty good. We learned a couple of things when we were wireframing, what do you think?” And so—

PF People love—love to be consulted. They love to be—

RZ Exactly. 

[32:48]

PF And it’s not just love, it gives them a sense of investment and connection and it brings down their sense of risk. Cuz if you’re like, “Look, this thing’s gonna sink or swim based on your input, so I need your input.” 

RZ Yes. 

PF “We may have had some fights about it in the past, that’s beyond us, and we need to sit down and figure out how this is gonna work for you.” Then they don’t—what they’re hearing is, “My life and future are dependent on this, partially, but also safe. Someone is building this to help me get the things down that I get paid for.” 

RZ That’s very true. That’s right because a lot of these people are experts in the old system. 

PF Mm hmm. 

RZ And they wanna feel like they can transition smoothly to the new one. They wanna feel comfortable with it. 

PF One of the things I always say is a product should at least pretend to give you a superpower. 

RZ Sure. 

PF And that’s true on these big replatformings too. 

RZ Oh, for sure. 

PF It should be like, “Mike, you know, you’re gonna be able to hit like 20,000 people where before you had to call five.” 

RZ Yeah. This is hard, what we’re suggesting because [yeah] tech people hate . . . I’m gonna generalize  here—

PF Oh I hate politics, hate politics. 

RZ They just hate all the nonsense and like, “Are you really—I have to go ask that guy? He is such a pain in the ass. Do I have to take notes from this guy? He’s a jerk. He thinks I’m useless.” 

[33:59]

PF And it’s also they don’t like—Nobody likes feeling disliked, right? 

RZ Correct. 

PF Most other C-level positions, you end up getting used to the fact that you might be at war with everyone else in the organization—

RZ You’re C-level! Outta the gate! I mean—

PF It’s a little tough but like that junior CTO often still really wants a hug sometimes. 

RZ Yeah. Right. Tip number two: don’t drop the big bomb. Don’t . . . assume that this all happens in one big fireworks parade release. It’s not gonna happen that way. And what you actually need to do is drop a bunch of little bombs. And here’s what I mean . . . if the group you’re servicing is 3,000 and it’s broken up into ten departments, give it to the first department. 

PF Mm hmm. 

RZ A) You’re limiting your risk because you’re not betting it all; B) You’re gonna learn a couple of things; and C)—and (C) is the most important thing—if you find momentum with that first group, your “campaign”, quote/unquote, takes care of itself. It just sort of catches fire and people are like, “Ok, when am I getting on this thing?” 

PF You know it’s tough for people though. This is very risky for your ego. The most satisfying thing in the world—and this was something I had to learn a bunch of times—is you wanna just drop something, like you know, in the chat and just be like, “Well here you go. Here’s your whole new life.” 

RZ “Later.” [Laughs

PF “Check out what a genius I am!” 

RZ Yeah. 

PF “Bye!” And that’s the worst thing you can do for your organization is prove how smart you are [yeah]. You have to as an engineering leader accept that you’re own intelligence and ability to change the org is meaningless [yup] unless the people who are using your software are excited and coming along. 

[35:48]

RZ Exactly. And you don’t need all of them to be excited, you just need 10% of them or 5% of them be excited and guess what? They go to lunch with other people. It starts to show. It starts to reveal itself as better. It’s like, “Man, am I next?” 

PF Right. 

RZ “Am I gonna get this thing?” And that is way, way better than you, every month or so, telling everyone, “It’s gonna be really great.” 

PF It’s hard though, right? Because engineering never gets a chance to like walk in, drop the contract on the desk, and say, “Hell yeah! I did it!” 

RZ No. 

PF “Here it is! You were lookin’ for ten million dollars, I went out and found you ten million dollars!” You’re a cost center [yeah], nobody particularly likes using your software. They like to whine about all of the bugs which are often part of the platform that you have to use [yup] and you have one chance here, one chance to really show everybody all of the incredible things that you can do, and we here at Postlight are snappin’ it out of your hands and saying, “No [Rich chuckles]. No, no, little bombs, little bits, little bits of impact, small teams.” 

RZ It’s like a campaign. It’s like you know you’re going from state to state and you’re trying to build momentum. 

PF It’s right. It’s just I know how much it hurts people. Every time I’ve told people about this, right? 

RZ Yeah. 

PF They just get so down. Cuz they’re like, “But you know I think a big release is better for everybody. We gotta get everybody in a room and we’re like, ‘Here’s what you’re—’ Otherwise it just like trickles out. They won’t get excited.” 

RZ Right. 

PF It’s not true. 

RZ So those are our tips, free of charge. 

PF Two tips! 

[37:14]

RZ Two tips. 

PF But deep tips. 

RZ We’ll continue this soon. There’s more to talk about when it comes to legacy. It’s an incredibly big part of business and it’s an incredibly big part of technology. If you have—I’m not one of those—I’m not Forrester where I could say, “This is 63 billion dollar problem,” quote/unquote, but it’s everything! Everything. Modernization of everything is just how it works. It’s inevitable. And you don’t have a choice. It’s like—

PF All large problems in technology are legacy problems. 

RZ I think that’s right. 

PF All of them. All of them. 

RZ I think that’s right. 

PF You know and honestly we talk about certain things like scaling. Scaling’s ultimately just a really fast legacy problem. You just got their faster cuz more people wanna use your product. 

RZ Yeah. 

PF You know, Twitter’s a great example. It was built on Ruby on Rails which is not a fast platform but it’s very agile, you can build things really quickly with it. But it didn’t scale and that’s how—Twitter got famous because when it would go down they would put up this picture of a whale held up by little birds called—and it’s called the fail whale. 

RZ Yeah. People thought it was cute. 

PF They thought it was cute but it really branded Twitter as like [sighs] an eye rolling like kids startup thing [yeah] because it would show this cartoon when it was down. It wasn’t a serious communication tool like the phone system, let’s say. 

RZ Sure. 

PF And all they had was an accelerated legacy problem. They had to rip out that Ruby system and I think they replaced it with a language called Scala which is like based on the Java virtual machine, it’s very fast. 

RZ Yeah. 

[38:46]

PF It has a couple of other things going for it and I think like—but it’s the same set of problems that like shipping and logistics has, just—

RZ Upgrade it. 

PF But in that case it was over like from year to year and month to month as opposed to five year chunks. 

RZ Yeah. 

PF But that’s just because they were experiencing insane social network hyper growth—

RZ Good problems to have, as you say. 

PF Yeah as opposed to trying to get their margins from, you know, 8.7% to 9.5%. 

RZ Yeah. 

PF So, a different kind of speed. So, look, we should—we’ll come back to this topic but um—

RZ This was a business slash technology podcast. 

PF See I love talking this stuff. This makes me so happy. 

RZ Do you really? 

PF You know, when I was a kid I used to read Ad Week and Forbes and Fortune. 

RZ As a kid?!? Like how old? 

PF Like 16. 15. 

RZ Oh, that’s weird. That’s another podcast we’ll cover [music fades in]. 

PF [Laughs] You’ve been listening to Track Changes, the official podcast of Postlight, a digital products studio, 101 5th Avenue. You wanna get in touch with us? 

RZ hello@postlight.com 

PF That’s right, so we’re saying, “Goodbye,” but if you wanna get in touch with us say, hello@postlight.com. Give us a good rating on iTunes if you’re so inclined. We’re trying to be helpful, let us know if you need anything. Talk to you soon. 

RZ Bye [music ramps up, plays alone for nine seconds, fades out to end].