Get in touch

Postlight joins Launch by NTT DATA! Learn more.

Software is like people: On this week’s episode of Track Changes, Paul and Rich sit down to discuss software baggage. We chat about how a simple software idea can morph into layers and layers of product features and about how implementing use-cases can lead to confusing and bloated software. We also share our many HR software gripes.


Paul Ford You know, when I used to do that, maybe ten years ago, it was like unleashing a bunch of flying monkeys because you’d be like, “At the core, isn’t this fundamentally just a search engine?” And people—

Rich Ziade [Crosstalk] Which can be very insulting and demoralizing—

PF [Crosstalk] And then people are, “[Three high pitched squeals] Oh my God! No!” [Rich chuckles] [Music plays alone for 18 seconds, ramps down] Ok, Rich, listen: one of the things that we do a lot of—a lot of here at Postlight is complain about software. 

RZ I don’t know what you’re talking about. 

PF [Chuckles] Yesterday, at this exact time, I was sitting in here and I was reviewing [music fades out] candidates for a Director of Content position that we’ve opened up here at the firm. 

RZ Ok, and we use a tool to do that. 

PF We use a tool. We use a—let’s not name it cuz I don’t wanna name all the software as a service because—I’ll tell you why: because everybody has a good growth marketing team that is gonna DM me on Twitter and say, “Hey! Heard some interesting feedback about our product. We would love to talk to you.” And you know what? I don’t wanna talk to you. I don’t wanna do your product work for free. 

RZ It’s not just that. We might hurt some feelings. 

PF We’re gonna hurt some feelings. We’re gonna say some words and the things that we’re talking about apply to just about every software as a service product. 

RZ Now, this is an HR recruiting tool. 

PF So people apply directly through the tool. That’s the nice aspect of a web platform: they fill out a form; they upload their resume; and there it is. Ok? And then I log in—I always have to log into it, always. I’m logging into all the time. 30% of the time I’m using this program, I’m logging in. That is a true marker of a 2019 web application. Especially when it won’t let you login unless you use some other authentication method and you have to do it twice per session. 

RZ Just to give you a sense—I don’t know how accurate this is: recruitment software is going to approach three billion dollars. 

PF A very large addressable market. 

RZ Yes. A very large market—

PF That’s in the US. 


RZ That’s, well, it’s—software knows no bounds these days. 

PF True, true. 

RZ But they’re saying—and this is based on Fortune business insights—I don’t know if that was Fortune or Business Insights or [Paul laughs] Fortune Business Insights but they’re saying by 2025, it’s a three billion dollar market. 

PF HR is vast, right? So this is not a surprise. 

RZ This is not a surprise and this is growing, apparently. 

PF Right, so this used to be, you know, your giant HR management system that only big companies used . . . but now there’s this second tier for companies like us. You know, we’re around 60 people; we’re always hiring because we work in tech, and we’re growing. So, we need a system like this. We also—like a lot of companies—fire six people, might interview somebody before they come and work at Postlight. So it’s two things, right? It’s document and information and knowledge management about individuals; and it’s workflow. 

RZ I think that’s it. 

PF That’s it. 

RZ And look, let’s also highlight: there are products out there—and you know how I always know—like there are products out there that are just mega size. Like, you know how you know when you go there and there’s a lot of people smiling, staring at computers but there’s no pricing. When there’s no pricing—

PF That is true. You hit the product page. There’s a couple of different shots, sometimes in a carrousel of people doing absolutely generic business things. Like, there are two guys at a white board and then the next picture is 25 people just looking up towards the camera. 

RZ [Laughs] Yeah! 

PF And you can’t figure out what the hell they sell. 

RZ Yeah. 

PF That’s very likely your large company HR service provider. 


RZ And what they wanna do is talk to you. There’s no pricing yet. And what they’ll do is they do a—they’ll bait you—

PF Well cuz pricing starts at hundreds of thousands of dollars. 

RZ It’s a relationship, right? It’s a long sell. What they wanna do is get to know you. So they’ll say, “Book a demo.” [Yeah] And what they do is make you fill out a form or schedule a call—

PF Or the white paper download is a good one too. 

RZ White paper download is another way to get your contact info.

PF Look for the top five trends that are gonna influence HR and hiring managers internationally in 2020. 

RZ You get the paper and the next thing you know the software is somehow in your inbox every 60 days. 

PF You know what else they do? Is they like to get some report from, you know, like a Gartner type organization [mm hmm] that would normally cost 20,000 dollars to download but they put in the white paper and they add like a little bit of their own cover too. 

RZ Yeah, “We’re helping you do the research you need to do to make a big purchase for your HR.” 

PF That’s right. That’s right and, “By the way—” 

RZ Now, we—we’re not that big. And there are a lot of—I don’t know what you’d call this second tier but it’s essentially, it’s Saas. And you do see pricing. There are tiers. It’s usually per user, per seat, or whatever. 

PF Let’s frame it on our—right, like if we spend a couple hundred thousand a year on HR software, it’s more expensive than having people do it; it’s more expensive than doing it—like you get—that would be really, really bad for us cuz we could hire less people with the money we were spending to manage our HR. So we’re not gonna do that. We’re gonna buy something that probably costs, you know, 150 dollars per user per month, or something like that. 

RZ Correct. And we’re on our sec—it’s worth noting: we’re four years old, and we’re on our second tool. Our first tool—[Paul chuckles] by the way, let’s talk about our first tool which was email us at jobs@postlight and then [yeah] we were essentially routing PDFs attached to emails. 


PF Forwarding to an HR Director. 

RZ Correct. 

PF So that was tool zero. Tool one, I’ve actually erased the name of it from my head cuz it was so terrible. 

RZ It was bad. It was incredibly disorienting; you couldn’t get around the app; you didn’t know where to go next. I wanna say that tool two solved it all. 

PF [Crosstalk] To be clear: there is an expectation. There is an expectation that web software, you don’t really need to learn. And that’s what—maybe we should’ve gone to a six-week training course, you know, but it was bad. 

RZ No, no, no, it was catastrophic and the truth is like I’m looking right now at our current tool and I’m seeing—hey, is there like an owner’s manual like somewhere? No, it takes you to a knowledge base which is very normal move but like can you help me like give me the five-minute tour? Like is there something—

PF This is a place where we’ve taken a step back, where it’s like: “Here’s a video to market it! And here’s the how to get started!” And then that’s it. There needs to be a Software As A Service Microsoft Help. 

RZ [Whispers] Microsoft Help. 

PF Remember that? You’d get some app, and you’d be like, “Oh, here’s the little weird pseudo-HTML thing that tells me everything I need to know [yeah].” It told you what you needed to do to use the software effectively. 

RZ Yeah, and look, I like the idea of zero training now, Paul. 

PF Nah, because you know what they do? You know what’s even worse though? Is like, “Ok, you’re not gonna give me any help because it’s all so obvious and intuitive but then you’re gonna pop-up a person in the bottom right and try to upsell me on stuff.” And now I’m starting to lose my mind. 

RZ Yeah. 

PF When you’re like, “Hey! Did you know that with bah-bah pro—” And I’m like, “You didn’t even get the baseline right and now you’re—” 


RZ Well when you go into help, you’re trying to solve a problem

PF Yeah. 

RZ You’re not really going through, “Ok, I’m going to put three hours of learning it.” Usually it’s like, “I’m stuck. Maybe if I hit this little question mark at the top right, I’ll get this solved.” 

PF Ok, so—

RZ Alright! So, wait! 

PF But wait, wait, hold on, you—

RZ Finish your story. Like you’ve got tool number one was not good at all. 

PF Alright, so we upgraded, we migrated everything, and that took a week. 

RZ That’s a lot. 

PF Like we had to spend our own development resources. So many thousands of dollars down the hole to get onto the new tool. 

RZ Ok, now we’re on tool two. 

PF Tool two! Suffices. 

RZ Well, let’s get one thing out of the way: it’s definitely better than tool one. 

PF Tool one, you just were like reloading the page all the time and work was wasted. Tool two does everything. It works. 

RZ It does everything it advertises to do. That’s true. 

PF And it lets you search through resumes. 

RZ Yes. 

PF It lets you upload PDFs. There are little edge-cases, like the PDFs don’t always show up in the browser, that’s not great. And there’s other things like that but they’re very edgy. Like one out of a hundred resumes you look at, something is clearly wrong on a technical level. There are understandable glitches but the UX is complicated. It’s hard to follow along. There’s all this workflow. Things fold and unfold. And we were talking about it and I was complaining for the 5,000,000-th time about this piece of software. You complain about it too. 


RZ Yeah. 

PF And you said, “It’s just a set of features.” 

RZ It’s a collection of features. 

PF And that’s real because what is—we said it earlier, but let’s say it again: a [sic] HR system is a document and knowledge management platform and a workflow platform put together. I have users in groups and I have tasks that people must perform around resumes which are the absolute core object. The profile of a person is the atomic unit of meaning . . . in this system and I do things with those resumes to move them along a path; turn them into interviews; respond to the interviews; so on and so forth. So, conceptually the platform is really, really clear. What I’m building on top. 

RZ Yes. This is a classic counter to what you’re saying which I hear people saying if they’re listening to this, like, “You’re grossly oversimplifying it. This is a very complicated piece of software. This is not a to-do list manager. It’s not just that.” 

PF Let’s be clear: the thing I said is still true. It doesn’t mean there’s not a million edge-cases. 

RZ Yeah. 

PF But the thing I said is true. There’s two things that you do. You create objects that represent people, and you assign documents and knowledge around them, and then you move those from one part of workflow to another. And yes, you might send an email—

RZ Pops in a chain. 

PF—You might reject a candidate; there’s a million things that follow on from that. 

RZ Right, so to clarify cuz not everybody here recruits people. We’re talking about Betty’s resume comes into the office [mm hmm], there’s a first call that happens in HR—what we call The Bozo Filter. And it’s just should this move along? Step one. 

PF Well cuz sometimes you get people who are literally applying for the wrong job. 


RZ Yeah, yeah, it’s a mistake. And so they say, “Ok, pass this step one.” HR Manager lines up four people at Postlight to meet the person, that’s round two. And then if they pass through that, and then each of those people fill out score cards, and then they pass this to the next round for final decision or whatever it may be. That’s the workflow you’re talking about, right? 

PF That’s right and the resumes stay on file so that we can search and return to them. They’re connected to the jobs. Like there’s things like that. 

RZ Correct. 

PF But it’s data, metadata, and users and groups, and workflow. 

RZ Yeah, and look, you’re bringing up something it is an incredibly powerful tool to constantly oversimplify things when you’re trying to solve software design problems. When you try to oversimplify things, you bring a certain sort of false naivete to the discussion, it forces you to bring scrutiny to the integrity of the overall product. 

PF But here’s the frickin’ thing: of course there’s a million following things you do. Of course you can’t just like throw Elasticsearch at it and cross your fingers. 

RZ Yeah, yeah. 

PF Just sending the emails about the resumes is a brutal problem that involves either five cloud services [Rich giggles] or, you know, personal spam configuration—

RZ Well, let’s not get into implementation! You’re talking about useability and just the thing being easy to understand. 

PF Even below that, I’m talking about the conceptual platform that you’re gonna use to build this thing. 

RZ So I’ve got a name for what you’re describing, Paul. 

PF What? 

RZ Backlog driven product. 

PF Ok, wait. That’s not what I’m describing when I say the platform is really simple actually. 

RZ Oh. I got excited for a second. 

PF Here’s the thing, Rich: that platform idea is right there, right? And we can see it. You and are—


RZ When you say platform, are you talking about software or are you talking about the product? 

PF I’m talking about what’s underneath the product. You and I see it, we complain about it, but what we’re really complaining about is a process that they use to build this thing which is that you can see. You can see and feel that it is a set of features meant to respond to different use cases for different people who need HR problems solved. 

RZ Glued together. 

PF That’s right. And then they back that onto a couple—they went, “What tools can we use underneath this in order to solve these HR problems?” 

RZ Yes. 

PF And software just emerged into the light almost by accident. And that’s not, frankly, how we see the world. 

RZ We meaning you and I. 

PF You and me. That’s why probably one of the major reasons we work together is it’s not just an intellectual exercise. If you figure out what needs to go underneath, like what the foundation of the house needs to be . . . the product is much better. It’s more consistent because we just assumed that people can learn what is actually underneath these things [mm hmm] and then you wanna empower the user, right? . . . There’s two ways to look at this. One is, “I’m gonna give you a set of concepts that you can very quickly intuit and understand, and then you’re gonna be able to do things with those software concepts made real on the screen.” Twitter is an amazing example. “I’m gonna give you a little box and you’re gonna put some words in it and then these behaviors are gonna emerge as more and more people put words in the box.” 

RZ Yes. 

PF “And you’re gonna retweet and you’re gonna like.” Like these are operations on objects but that’s the platform. It’s a very simple platform. And then you add search to it, and like there’s a few other things going on. 

RZ Software is like people. I have little kids. And I’ve met their little friends. And some of them are monsters. Some of them are terrible. They’re just—they wanna destroy your house and they wanna do evil things and they’re really mean. But even at five—

PF Mm hmm. 


RZ I see something very pure there. 

PF Oh! 

RZ I see something there that is still just this kind of wonky personality but, you know what? It’s fine. It’s like they’re cute and they’re actually—there’s always kindness there. Even when—like a four-year-old is kind, fundamentally. Even when they’re mean, like—

PF You know why my wife was happy last night? . . . Cuz my daughter was just as terrible to me as she was to her. [Rich laughs] Right? . . . It’s just it was good to her to know that my daughter was just being annoying to everybody equally, right? 

RZ Yeah, yeah! 

PF It’s just—yeah. 

RZ So, when you have that idea, we’re gonna make a beautifully elegant piece of software that moves a candidate along a chain and it’s smooth and there’s very friction around how it flows—

PF And what I’m thinking is—

RZ Lemme—lemme go through this though. 

PF Ok, keep going. 

RZ And oftentimes it starts that way! There’s a good chance that tool number two version one point oh, had a pure aspect to it, a very pure—

PF Sure. 

RZ—a very pur—as Twitter did! 

PF “We wanna solve HR because it is a pain. I wanna make an inexpensive tool. We can do this better than everybody else.”

RZ Yes. 

PF “That lets us move thousands of resumes through an organization and makes it really easy to apply. We can do that.”


RZ And they do—let’s say they did. I don’t know one point oh because we have whatever version it is. Let’s say they did. 

PF Cleary they got there cuz we’re—Ok, they got somewhere. 

RZ The child is five and there’s a pure aspect—and then you see that—you lose touch and you see that kid when he’s 19. 

PF Just to elucidate this metaphor as it’s going along [Rich laughing]—keep going. The beautiful native state is like . . . “Here we are. It’s full tech search that works beautifully, that gives me great relevant responses. And I put documents in and I get them out.” 

RZ But what happens is they land 50 clients and the feedback starts coming through, and patterns are starting to emerge around the customer—not call ‘em complaints but feedback. 

PF Well, you’ve got these Product Managers who are writing up use cases and user stories based on the input from your actual customers. 

RZ I’ve heard this ask like 11 times already. 

PF Yeah. 

RZ “We gotta put it in.” Ok. 

PF What they really want is a way to schedule five people for an interview. 

RZ Yeah, exactly. And it’s like the VP of Product is like, “Look: I can’t just take every request and put it in the software.” They’re like, “Fine. Ok.” But you know what I’ve heard this like seven times. And by the way, one of the clients is Boeing. So we’re doing it, whether you like it or not. 

PF [Crosstalk] They wanna print all the resumes that come in on a daily basis. [Rich laughs] Like that is a classic—

RZ Some whacky—yeah, exactly, exactly. 

PF “As a large aerospace company, HR and Hiring Manager comma I would like to print every resume that comes in on a daily basis comma because I need it for archival and legal reasons period.


RZ Yes, and, “Looking forward to chatting about renewing the license [laughing] next week!” 

PF That’s right. No, no, but I mean that’s my user story, right? 

RZ Exactly. 

PF That goes in and like people—now, all of a sudden, there’s this artifact . . . that doesn’t actually connect to the platform, like isn’t part of that original vision. That original product owner ideally now has been probably promoted [laughing]—

RZ Also! All these things have been logged, right? 

PF Yeah. 

RZ So what do we log? Probably Jira. 

PF [Crosstalk] But also the guy who built the first version or the woman is not the CTO and they’re just trying to keep this hurricane under control. And now there’s 50 Product Managers because they got some venture capital. 

RZ It’s a huge success. Yeah, I mean and this stuff’s getting filed away—

PF It’s a huge success but it’s also—now it’s got venture capital and it’s a giant, gaping maw that is being filled with user stories that have to be turned into product features or they won’t be able to find a hockey stick of growth. 

RZ That’s the theory, right? 

PF Yeah. 

RZ Like that’s the hope that if you get these features in, so I can keep going. 

PF The core hot little product like that actual key platform underneath where you could look at it and go: “Wow, this is a great resume management tool.” 

RZ Mm hmm. 

PF That is about to get covered up in layers of sand, garbage, and children’s toys. 

RZ [Laughs] Not necessarily though! Because there’s a couple of things that happened. First, there’s one of three possible paths. One is yes, just layers of garbage get layered on and on and on and like Jira looks like an Allen Ginsberg poem of user stories [yeah] and it’s just all these shards of needs. Like just layered on top of each other. And whatnot. So path one which is the most common and, frankly, is the outcome for our product that we use, tool number two, is you can see, it’s been caked on. The features—they had to check the boxes and it’s been caked on. That’s path one. 


PF Well, and you know—and the other pure example here is like the Spotify desktop app. 

RZ You know, I wanted to go to Spotify with this conversation because Spotify definitely like you could tell there’s no cohesion to how the product evolves—um—

PF No, I mean, the Spotify desktop app is like you get a big frying pan and then one team puts on an egg on the right; one team starts to make pancakes on the left. 

RZ Connect us back to our 19-year-old who was really cute and pure as a four-year-old, they have what’s called baggage now. 

PF Yeah, also known as body hair. 

RZ That’s path one. It’s a bad path. You end up with essentially bloated, feature-driven, what you call backlog-driven product, right? Path number two is you have a VP of Product that always forces you to test it against the integrity of the whole thing. 

PF Now, hold on, first of all, path number one represents about 97% of software services—

RZ 97%!

PF Like this is real. So, let’s just say what we’re describing here which is that . . . a product comes into the light; a small team has built it; it’s built on understood APIs; and it’s tight and it runs good and it’s like a [sic] early motorcycle. Like just [makes engine revving sound]. It’s fast and it kinda burns a lot of fuel and it’s cool. And then, in order to grow, it requests and user stories start landing, those turn into Jira tickets and that becomes the entire flow. So then people start literally putting chewing gum on it to hold—

RZ And look they’re making money and they’re checkin’ boxes, right? And they talk about these feature adds. 

PF And this is not bad. This is just everything. This is the baseline for how software gets built. 

RZ No, this is—In a hugely successful product what happens is this: they say, “You know what? It’s time to redesign it. It’s time to shed all of that cruft. All of those weird decisions that were made over ten years—


PF [Crosstalk] You know what also happens? This is real magical too. 

RZ By the way, this is the equivalent of the 19-year-old starting therapy. 

PF That’s right [Rich laughs]. It’s not just that. It’s that they don’t just redesign it because everyone’s like, “No, no, second system’s syndrome. Too dangerous.” Second system syndrome is like when you decide to rebuild the thing, everybody tries to make the perfect unified everything and it’s a disaster. But you know what they like to do, is that’s when they introduce the new language or framework on the backend, and on the frontend. 

RZ See this is thing. Like, their chance to get in there, right? “We’re gonna rebuild it.” 

PF Twitter did this with Scala like [yeah], “You know, we’re getting off of the Ruby,” and they’re moving towards a functional Java virtual machine programming language. 

RZ Yeah. 

PF And it’s always a magic bullet. “We’re gonna solve this problem.” 

RZ Look, it’s a huge undertaking, you look at software that’s version 12 and what you’re looking at is essentially those layers of dirt, right? It’s those layers of rock that if you were to bore right through the earth, I mean it’s just time passing and that was a hurricane 20 years ago and—

PF Everybody always wants to solve it with the smarter tool. You know the smart way to do it. 

RZ What?

PF Ok, so a good example: let’s say I built a giant platform in Python with a bunch of like HTML and JQuery on the front. 

RZ Ohhhh. 

PF It’s a mess, right? And it’s a mess in 2019. 

RZ I wanna just give a shout out to JQuery. When I first saw it, I was like, “Wow, that’s really neat.” 


PF No, and it’s like Flask on the backend and, you know, just [ok] . . . all good but it starts to bump up against all kinds of limits if it’s released—

RZ [Crosstalk] Ok but you’re conflating two things now though, right? We’re talking about features being caked on but you’re talking about also the actual tech is dated—

PF I know but it all gets perceived as one big thing at the company. 

RZ That’s terrible. Isn’t it? That’s a terrible, terrible thing. 

PF So then you have two solutions: one is, “We’re gonna rewrite the entire Python stack in language—” 

RZ Don’t say closure. If you say closure, we’re shuttin’ this podcast down. 

PF Let’s—language Wiggle. Wiggle will be the name of the language. [Ok] We’re gonna make up a language and Wiggle will be—does everything—

RZ Multithreaded—[laughs] . . .

PF Oh it does everything, it’s—you can scale it to 50 different machines at once. 

RZ Of course. 

PF It doesn’t have any particular support on cloud services but that’s ok . . . And you know just sort of like you make the most future looking stack cuz you’re like, “We are driving into the future.” The thing that really seems to work is like you carve off little pieces, write them in lower level programming languages that are kinda boring. They break a lot and then you fix it. 

RZ When they hear, “Redesign,” the engineers show up. And—

PF Oh yeah. 

RZ They’re like, “Oh, well, here’s our chance! Cuz you know what? That was eight years ago and that framework isn’t cool anymore.” 

PF “We’re gonna get rid of all of our visual language and start with the new brand.” 

RZ Yeah, and look, to be clear, this isn’t just—Like we’re being glib about it but—

PF This is just humans. 


RZ—engineers—technology does get better. And so yeah, you should take a look at things but it is still fun to like say, “Oooh! Look at this! Completely clean slate.” 

PF Here’s the problem: there is no solution to this. Everyone has an answer. 

RZ Yes. 

PF The earlier they are in their career, the more concrete the answer is. You know the sign of a truly serious product leader, engineering leader, design leader? . . . Is you tell ‘em your problem and you say, “I think I’m gonna do this.” And you think it’s a good idea and they go, “Well, it depends.” Because you’re gonna make tradeoffs.

RZ Of course. 

PF So there’s this like feature agglomeration based on user stories and Jira tickets. 

RZ Yes. 

PF That’s one way it can go. But the pure product, where we’re gonna really keep the focus, means it actually slows your growth. 

RZ Yes. 

PF Look at Gmail. I mean it’s huge. But my God. Where are we after 200 years? . . .

RZ Pretty much—

PF Where we started. 

RZ Where we started. 

PF A little slower. Now you have an inbox that dances. 

RZ Yeah. 

PF Filters are still hard to use. 

RZ Yeah. 


PF Inbox was like a weird side thing. 

RZ Let’s pause for a second. Can we name a product that we’ve watched it grow and evolve over years and it didn’t get worse? And it just kept getting better, in fact. 

PF That’s a tough one . . . Aspects of Google Search, certain programming languages. I think Python’s absolutely getting better. 

RZ But that’s not a fair—because like I mean our case is like software that like is hearing, you know, the shouts of all the big clients, right? All the big licensing fees. And, you know, Google doesn’t have to worry about that. 

PF Yes, and you know where it actually—you know where things get better? . . . This is a little abstract. At the data layer. Postgres is better than it used to be. Elasticsearch. Like those sort of things like where the use case is focused around true stability and speed. Those things get better. 

RZ Is there enterprise software that continues to get better?

PF Mm. I mean—

RZ That’s almost impossible! 

PF It was never good. 

RZ Right. 

PF I mean I will say like if you look at SAP now versus SAP ten years ago, it is designed like a modern software application like—

RZ Sure, they modernize it. 

PF But they’ve caught up to 2010. You’re still—

RZ Basecamp? 

PF No, I find that it’s actually gotten much worse. 

RZ Ok. 

PF Slack is not as good as it used to be. 

RZ Ohhh. 

PF Don’t you think that’s true? 


RZ It’s getting a little cluttered. I would say. 

PF It doesn’t have a hypothesis for the future of work. It is just holding onto its enterprise. I mean I love Slack. They’re our friends.

RZ Yeah, yeah, yeah. 

PF This is hard! I don’t—

RZ There are like five different ways to get to settings. There’s like five different settings menus. 

PF It’s seriously—If I had the answer, I’d go ahead and fund the replacement. Like go give somebody 20,000 dollars. 

RZ It’s really—I think what we’re saying is it’s really, really hard. The pressures of continuing to be up-to-date and—

PF How would you fix Slack? . . . You are now the Product Manager of Slack. 

RZ I would look at usage. Like on a granular level. And take things away. 

PF Yeah. 

RZ I love taking things away. 

PF I know you do. 

RZ That’s like my favorite—

PF One of your favorite things to do. Yeah. 

RZ [Laughs] We have this very cool add-on to Slack called Dash that lets you essentially have—create channels that expire. 

PF That’s right. So it’s decision—

RZ Decision-driven and there’s like always time kind of putting pressure. And when you create one, you have to put what day it expires and what time of day it expires. And I hate the time of day. And I want it gone. 

PF Yeah. 


RZ And I’ve been complaining about this for five months now . . . and it’s tearing this company apart. 

PF It is. It’s brutal [Rich laughing]. It just stops every conversation and it comes up constantly. 

RZ But like how would you revisit that recruiting tool? How would you revisit Slack? I would look at what are people like wondering why that button’s there for? When 98% of your users are wondering why that button’s there? Get rid of it! 

PF Well, see, here’s what’s tricky. Let’s put Slack out of the equation because Slack has 650 billion users including like everyone on Mars, right? 

RZ They probably have great data though around usage, right? 

PF With Slack, there is a case to be made that Slack has actually had like the local optimum for Slack. Right? Like I’m sure that if you said, “Let’s take this away.” Someone goes, “Yeah, that will cost us two billion dollars in revenue.” I mean it’s not—

RZ [Laughs] Well that’s the hard part, right? 

PF And it’s hard in both directions. It’s hard to take it away, and it’s hard to add. But our HR system is unequivocally a mess. 

RZ It’s a mess. You know what we’ve learned to do is we’ve learned to navigate it . . . in a bad way. I’m saying that in a bad way cuz I know the three clicks that gets me to where I need to get to which is terrible. 

PF The other thing, too, there’s a huge difference between a tool like a Slack or even a Gmail where you basically spend hours in them, and so you’re willing to optimize around them . . . Unless I’m an HR Director—

RZ They’re also more generic, right? 

PF And they’re using more conventions that are basically like of the operating system or of—they’re their own environment. 


RZ We had a review workflow tool . . . that allowed us to review each other and we’re [oh!] dropping it cuz it was just this bloated, bizarro tool. And we’re just not gonna bother. And we’re gonna use like Google Forms or something. 

PF Well, it was one where when we did the math on how much it costs and what would it take for—

RZ Also, nobody used it. 

PF Nobody used it and if a human coordinated the same level of effort, it would be better for our HR person to manage cuz she would be directly involved. 

RZ Yeah. 

PF It was this thing where it’s like, “No, no, a human is much better than software in this case.” 

RZ Yeah. 

PF For both her and for the company. 

RZ Let’s end this with a piece of advice. 

PF Ok. So, wait, wait. What would we do to fix this HR product? . . . See I don’t actually think you do. 

RZ Uh, I think—Look—

PF They’re fine! 

RZ The cat’s outta the bag, right? 

PF They’re gonna get bought by—

RZ Is it the cat’s outta the bag or the horse left the barn or something? 

PF No, both are fine, actually. Yeah. 


RZ Ok. The cat’s outta the bag. Here is the number one piece of advice I would give . . . is when you got that cue, that sort of big pile of user stories that have come in from, you know, 200 different clients and you’re thinking about the next version of the product, as a product leader, step back. 


PF Yeah, that’s right. 

RZ Step back out of these discrete, one off . . . almost like a to-do list of features that you need to check off. And step back and think a little more cohesively about what fundamentally is going on in the product and how you can—is there a way to think a little differently about the product and knock out six of those tickets at once? 

PF Can I throw another idea out? 

RZ Yeah. 

PF Like pure efficiency point of view. Pure I want growth and I want quality . . . The product is six, seven years old. It’s a collection of features with a core underneath. You know what I would really do? 

RZ What? 

PF I’d just create a new product. 

RZ That’s devastating, though, right? 

PF No cuz—

RZ That’s redesign and the rebuild—

PF No, don’t do it that—

RZ Is it the same product? 

PF Don’t do it that way. Like I would almost timebox it . . . even if you’re gonna throw it away. Put a team of three of four on it for six weeks. Say, “Go find every cloud API, every simplifying service. Rebuild this thing and make sure that we can migrate this data schema to the new thing.” 

RZ A new experience. 

PF Just do it. “Get me like—draw it for me in wireframe form and—” 

RZ You see that sometimes—you ever see like, “Do you wanna try the new . . .” 

PF Because you know what? 


RZ And you can hop over and feel this new thing. And then if you don’t like it, like that’s a very—

PF Fall back to—get out of beta. Go back to alpha. 

RZ It’s a safer kinda way to test the waters and see how people react. 

PF Because I don’t really like—

RZ They love code naming that shit. The new thing. 

PF Yeah that’s right. Gallion! [Rich chuckles] Here’s the thing: unless you are truly committed, it’s just too hard. It’s Jenga. And you’re pulling piece out after piece out—

RZ It’s really hard. I think once you do it, it’s hard to unwind it. 

PF And that’s o—

RZ So be careful when you do it. And when you get those changes coming in, you’re like, “You know what’s fundamentally wrong here is the way we signal out that something’s hopped in the workflow. That’s why these 11 tickets are there. So let’s talk more fundamentally about what the problem is.” That is what—I mean, I think a great prod—That’s the difference between a good product leader, who is someone that’s gonna move along tickets—to a great one. A great one takes those six and puts ‘em to a boil and then it like melts down into one solid block. 

PF Well, look, they’re building new abstractions that they can then later build other things with. Like you’re making building blocks, you’re not actually building features. And that’s—a good product manager is someone who’s like, “I need to create new Legos.” Not, “I need to make a little guy drive around in a car.” 

RZ That’s right. 

PF Now, look, let’s really end this. Let’s do what we wanna do which is how much—

RZ Name these actual products? 

PF No. [Rich snort laughs] That I desperately wanna do but we’re not gonna do that. I just came to you and I said, “I am tired of every HR system in the world. And I have lots of money. I wanna build my own. You’re an agency, how long’s it gonna take for me to get something better than this piece of crap?” 


RZ Six months. 

PF “Really?” 

RZ Yeah. 

PF How big’s the team gotta be? 

RZ Four to five. 

PF “Four to five people; six months, you can get me something that competes with this?” 

RZ Absolutely. 

PF That’s real? 

RZ Yes. 

PF That is actually what’s great and terrifying about this industry . . . is that I think that that is true. If we sat down for about two weeks, we could figure out exactly a six-month plan to obliterate that piece of software? 

RZ Yeah. And, to be clear, by obliterate, it doesn’t mean we’re gonna have feature parity. I don’t want feature parity. 

PF No, no, no, no. 

RZ That’s not what we want because there’s some bizarro-ness here. Like there’s some bizarre, really strange things going on. 

PF I do wanna make sure that people could eventually target this with some of that data because we need to rip that content out and put it in our world. That’s real. 

RZ We also, just to be clear, have a world-class set of talent here at Postlight . . . at 101 5th Avenue. We’re a digital products studio. No, I’m—I’m joking but this is the Ten-X engineer, right? Like we’ve got talent here that—when I say six months times five people. I’m really saying the equivalent of 18 months of a team elsewhere. Not to sound too arrogant but it’s true. 


PF Look, I’ll stand behind that. I think the other thing is—here’s what’s really specific: you don’t have to go out and find that team. 

RZ That’s huge

PF That’s huge. Right? 

RZ That’s probably the single biggest thing that stops—

PF It’s 18 [music fades in] month to build a good, small engineering team. 

RZ If you build that engineering team more quickly, they’re not as good. 

PF That’s what I mean—

RZ You hired more quickly. Too quickly. 

PF And it’s not just us, like people who are aligned against specific things—

RZ Totally. 

PF But that’s what I’m saying, like six months, this is what drives me a little crazy in this world is like you’ve got all this bad software and you’re six months away from making it better. That’s the eternal struggle and they should live in fear. Everyone should live in fear.

RZ Well what I like about this podcast, Paul, is we slipped in how good we are as a digital products studio. 

PF Well, this is content marketing dammit!

RZ That’s how it’s done. It doesn’t feel like an ad. 

PF No! Not at all. Not an ad at all for Postlight, a digital products studio at 101 5th Avenue—

RZ Here in New York City [chuckles]. 

PF Anyway, look, Rich—

RZ We love building platforms like this and thinking about big product problems and we wanna be your product partner. Hit us up! 

PF That’s right. 


PF If there’s any piece of software that you would like to see destroyed, within about five to six months, [Rich chuckling] a team of five people can destroy it for you. 

RZ Yes, we’re doing work for some amazing clients right now. Some of the biggest companies in the world. So, talk to us. 

PF Alright, friends, let’s get back to work. 

RZ Have a lovely week [music ramps up, plays alone for four seconds, fades out to end].