If You're Asking When to Kill a Feature, You Already Lost | AA269

So Om, a guy asked me an interview [UH] how I know when to kill a feature. And gave him the real answer. And just know I didn't get job. You told him the truth. I don't know. So here's the answer. you're asking when to kill feature that you have built already, you, you already lost. Welcome back to Arguing Agilent. And if this is your first time, welcome. I'm your host product manager Brian Orlando and this my -host enterprise business agility leader and the reaper of the road maps, Mr. Om Patel of the Reaper of the Road Maps? Today you are thought was funeral director features one other. Like everybody asks the same thing. When do I kill this feature? Wrong question Om! Oh yeah. Wrong question. Right. So, so you're asking when to kill. You're behind eight ball already that's wasted something by now.[UH] we're we're going take that question apart and[UH] [UH] into four questions that you should have asked instead. And then We're talk about some the things that I have here screen asking when we can kill something obviously a tell that don't have really great discipline in in place or don't have the phases software development the sun setting or they sorry the exploratory phase and then the [UH] exploitation phase or the maintenance phase whatever to call it. And then the sun setting phase. don't have good or like alpha beta production probably not something that they talk about at your shop either. few things like that. then, [UM] number two, we're talk that nobody can kill the CEO's pet pilot or the CEO's pet feature or whatever. That's where lot people that might listen podcast are this is great, but it's not realistic to me. That's right. That doesn't work the real world. It's product management doesn't't work here. That's right [UH] yeah, that's right. We haven't tried really, but it doesn't work here. And then we're going to try talk off ledge about that. There's no killing things. There's just sun setting. OK, great episode. so if asking question when to kill it is tell here's another trap. OK, the two things that you would use make that call are the two things you can't trust. The metrics got gamed the moment that you tried track those specific metrics. Absolutely. Yeah. So oh, and one is, well, our biggest customer would throw [UH] they throw fit. They'd riot, you know or other or other things that nobody can actually disprove. So you're between a you're between a qualitative and quantitative rock in hard place Sorry. One, feel the other one really hits you. That's right. Oh my goodness. That's right. That's. So that's this category. so data not help[UM] But there are some arguments. There are some definite arguments[UH] So let's get into the steelman. there cynical take this one [UM] isn't there? Yes, there's certainly cynical take. But this one might actually take some people by surprise because they always say the numbers are the numbers. That's right. Numbers don't lie. don't lie. Exactly. Except they do. Yeah, I've been in shops where I say this and think said this on our on one our podcasts about the numbers are numbers. know, mean numbers don't lie. Just use numbers to cut features. That's cynical take. If you think about that. I was, I was how that cynical? That's the way is until were prepping package. I'm oh, actually that is cynical take because now we're not, we're just being, we're just being quantitative we're ignoring qualitative. We're also ignoring a bunch other stuff that you can't see in numbers. Exactly. And the sources of numbers matter as well. Right Where are you getting this stuff from? If it's not your own generated source, then it could be tainted. So again, these numbers, but are they really? Yeah. Can you trust them faithfully? That's the question. You know, other, the academics point here of you got to just instrument everything and then you can, know, can, you can be more data driven or know, data decision making whatever is. going say, know, just instrument everything and then we can decide based on data because know have to maybe that be our definition done. Everything gets instrumented And normally I would agree that. That's great, that's great idea concept. But I've also been bunch startups are like barely hanging on. Sure, sure. mean[UH] again, talk about opportunity costs. mean, how much invest in instrumentation? Right? You know, could just throwing things out there. The customers can and see. I've been, I've been at quite few companies that had no instrumentation just like developer level logging. I'm can't dashboard. got get developed. You basically to pull developer off what they're doing. Yeah, To go figure out anything. I'm like, guys, believe not, it's a hard fight. Maybe this maybe the logging thing can be separate podcast because to dig too deep into now. But the logging thing instrumentation so that people that are non -technical can go look for signal themselves without, you know, stopping and stopping a development team member or whatever. Yeah, that's an interesting one that requires [UM] that requires commitment intent and know, and some expense. Right. You're actually going to take some time trouble to do this. Right. And there's nuance there too because can go, mean, can whole team could be doing this for a half year. Depending the product. Yeah, depending product. Yeah, yeah, yeah, yeah But but yeah, the academic is taking nuance out this and just instrument everything super the academic who is easy. Isn't It's taken little extreme view right there. How long want keep them logs forever. Exactly. 90 days. Like those logs could be big. They can. Yeah. Especially you're logging at a lower level. Yeah. the other point. The other thing point out now is Campbell's law. The metrics like when metric becomes the target they become meaningless. Basically ceases be useful. Yeah, yeah, yeah. That's good time to point out that Campbell's Law is like full effect here besides fact that there's bunch other stuff on anyway yeah, the practitioner is a low usage can still be load bearing And the practitioner here has a they don't have great they're any arguments with that point is what I'm saying that I've hey it doesn't used lo. But the people that are using the people matter. Those the cases that I've made product manager. Yeah. So in this case, maybe the people that are using it, mean, there smaller population and you're really looking for stronger signals build something bit bigger and more complex. Right? So it's not the end all product but it's interim product. Maybe you want that feedback. I'll tell I'll give another. I'll give real scenario of from my time in defense tech is we built a product that was sort like a how to describe it. was sort like a how to orient yourself to the information environment, like the information environment oh, what's going Internet today? Like where go to figure out what's going on Internet today? what are the trending videos and headlines and hashtags all that kind of stuff. Like believe not, like that stuff has rhythm to it. Sure. Where like especially have tools that let you go see like how hashtags grow and spread and become other hashtags because the when you dig into it, you might find out that source of those hashtags are what looks like organic movements online is really like state actors under hood. Sure. know, absolutely And sometimes they take other hashtags and they insert a different topic basically they like hijack wording Right the product was orienting you to what's going on today and then tracking back to, know where did start? What was source? its origin? Where they trying to go with Potentially like analyst comments where they're trying go, that kind stuff. And what we found out, what we wanted was we wanted, oh, we want people logging every day and spending, know, hours on it looking around. What found was what people would do is they'd log in. We'd look their, you know, time zone adjusted. They'd log in in the morning hours, you know, seven, eight, nine o in morning. Spend maybe 15 minutes the tool and then they wouldn't revisit tool either until next day or until like late afternoon if was a heavy media day whatever. Yeah. And when I actually went and went to talk to some those users, we found out that they really did value tool. But they don't, they don't use it like we were thinking is like read every single thing. They would log in, orient themselves to whatever they were looking at that day, whatever they were interested that day because they didn't just have my tool. Right. They had half dozen other. Yeah, right. Right. In addition to searching Google doing whatever themselves. Right. so 15 minutes as long were doing it, know, every Tuesday, Thursday whatever their cycle was, was actually fantastic because these are very busy people that are trying to chase, you know news stories and hashtags and inauthentic, you know, activity over Internet every day. So they log in, orient themselves, figure out what things are, know, write what things wrong whatever and then go chase leads for rest their time. And that's how they. So basically they started day by checking me. know, I was like the morning newspaper that they checked with before they started rest their day. So was, it a great compliment that didn't have spend ton of time. Right But but will tell you as product manager that time we didn't understand our users usage of the product. So we had change way we built the product once we learned that oh, the 15 minutes twice week or three times week whatever in the morning is actually good signal. Yeah. And how can we leverage the little time have to make biggest impact? It required bunch like UX changes other things. You know, it sent on different road I'm saying. Yeah, so you've just highlighted, you know, something touched on earlier. The numbers, right? Yeah, the numbers aren't everything. You've really dig behind meaning there. So yeah like I know going along in category one but like know know does help with that having a a having a designer on your team really helps with that because they they they're ones that[UH] want to pull all those threads when they see conversations like that. And also they're ones talk people about, you know, how using it, what's workflow, that kind stuff. Whereas developers or the salespeople or executives whatever, like they may not, know, they may assume like course that's what they're doing because whatever. then you know get to whatever next assumption is because they don't really have to test their assumptions and live that world. Must be nice [UH] nice Oh yeah. For I want this section called a better, better, better arguments or something like that. needs called that because it's Steelman and then should be better arguments or better points something. Right? Yeah. Because it's kind saying, you know, for but there's no against. There's no against. It's just yeah, it's just like here's the here's bad takes that you'll see here's bad takes that you'll hear and better takes. guess could called better takes. Yeah, that's fine too. Yeah, yeah, yeah, So first one here, I mean really, know, no metrics survives becoming the kill trigger against game before gets used. I'm just thinking here that probably applies to most things rather than just few people will game metrics to suit their own purpose. Yeah, mean, the second point here, the signal that you needed came before the build, not from the dashboard [UH] metrics or whatever after the build. Like that's main one. That's main one. when you're talking people face about their problems, it's whole different experience than trying to like surveys talk to people over the [UH] asynchronous over Internet surveys, whatever. Yeah. The signal need came before build. And you were talking people before you intended to build anything, you wouldn't have this problem. You wouldn't be in this boat first place. Absolutely And if you think that you're a super busy back and technical engineers solutions architect whatever and that you're you're super busy with your own people and they're out talking people. So you don't need to either First all, you're wrong. Yeah, well, this that old. know who I'm pointing out right now. Yeah, this old mentality. know, don't let developers talk customers. mean, that's that should be. So 90s. But that is an old mentality. Yeah, right. Yeah, that's right. And think all people that have that mentality, think they're all dead. That's not true. I mean, maybe it's true. don't know. don't know it's true. Thanks coming live broadcast. don't need be like spoken too much. it's fairly obvious that one isn't like the decide what evidence counts before anyone defending the feature or you know, has invested the feature whatever. Like if you're talking customers, then you don't need do this one. Right? Yeah. But you're not. Yeah oh, we're to keep that resume updated. Yeah, yeah, That one. All right, let's move on takeaway here. So stop waiting for data to make the call. Where's my see now? don't have Mark Zuckerberg commander data. don't think I missed out on this one. right. Decide what evidence would count before anyone is defending the feature. so don't don't like don't fixate on data afterwards. Fix it on data before. Beforehand. Yes, absolutely. Right. Yeah yeah. That, that's a, that's. That's probably the strongest one can make this category. Let's, get out here. Let's get here. So tell me the comments, what's the number your team hides behind the one that keeps that dead feature moving along on life support. Let know. So even the perfect number won't save you. The deadest feature is that is that word the most deadest How about that [UH] I know adjectives. The deadest feature as as deadliest in in your the deadliest feature. And I'm keeping all this in in your product. It's the one that nobody is allowed to touch. And that, that's what I hear that one's let's talk about sizzling. Let's talk about Listen, the suffering will stop once we've reached acceptable level of suffering. Thank you very much. so the data can't carry call. The most expensive decision in product is not like some these one to kill it. It isn't doubling down. It's the one that nobody called decision. Just leave alone. It's still we're still piloting it. We're still gathering data to survive. It's fine. We'll keep eye on it. The worst one, the pilot is funded by the CEO. Oh boy, nobody will dare touch that. Right? Nope. That's that's a, it's shrine, not graveyard. shrine. It's a shrine. The graveyard full of graveyards for the next. Bring out your dead. Bring out your dead. Me that[UH] man yeah. It's like listen executive they fund everything. So they get get have their pet projects. They're funding everything. They should. don't touch the CEO VP whatever. Spread project like that. You got to right there. That's a, that's realist. got be. were real. That's a cynic point, right? Yeah, yeah, I get like the practitioner cynic are very close this one. Hey, killing the feature picks fight that you're not win. That's, that's the practitioner. I'm just trying to, I'm not going pick battles that I know. I can't win. That's a practitioner syndicate. practitioner [UH] they're holding hands. They're holding hands this category. Right, right, right. Yes, yes But know, one things to say here is even though you don't feel like you can kill it, feel free to poke it gently What I'm saying is try figure out why the VP is funding this. Like what's behind it. Are you testing new market with this thing? don't know. Right. Try and understand least. Yeah. Yeah. Have some ones. Nothing wrong that. Yeah Look a a tableau dashboard together. Sit down and go through data tableau lost the thing. Yeah. Oh absolutely. kidding me? Yeah, absolutely. OK. Oh yeah. Yeah. I mean, I don't understand academic this category. Preserve optionality under uncertainty. don't understand means. Preserve optionality under uncertainty. So it's like rule out. don't rule anything out. don't pick battles because like you need like don't understand that. That's the cynic and the practitioner. Don't pick battles. Yeah, but the academic saying like ignore it. I think that's academic saying. Oh really? Well, don't like it. Well, it's like keep options open. That's the academic is saying here. Yeah. So we again, the same thing mentioned earlier, try understand rationale behind it. Forget about who's funding it for second and say should this is this best use of you know the funds. Right. You're not to change the VP's mind. You're just simply underlining that with, you know, here's the alternative. Perhaps like a, I don't know, an opportunity cost. But yeah, they might sway you by saying we're exploring something new, a new customer, new market. I'm all about that. mean, I'm all if they have money burn. I'm all about moon shots. Not, not, not giving K3 moonshot. not those. I'm all about like we got some money burn. We might well shoot moon on couple things. And know, as long we're not like rolling the team off when the money is expended or whatever We got crack crew. got to. So got beginning the 18 and bring it for some reason like I'm not I'm not, I'm not doing I'm not doing it. right. OK, got optionality. We're not to commit anything. And if there's free money, then like, hey, don't fight it. Like keep your options open. Right? Yeah. But also, I'm a bunch times I've seen teams like that. They're like slick sexy teams spun up real quick and then burn out and then they're all gone. I understand the viewpoint of like justjust ignore Brian. You know, don't worry about it. Don't worry about this don't feel comfortable with that viewpoint. can just ignore it. That doesn't make sense me. mean, listen, the practitioner saying hey, don't pick fights that you're not win is the same as academic being keep your options open keep your resume updated. Yeah. Yeah. Those people aren't going anywhere anytime soon. Right like you said, they're not ruffling any feathers. Boy ruffling feathers. need you be my eyes ears, eyes ears. That pet pilot survives on rank, not on merit[UH] also not on evidence. Yeah, well, if it's just pet project, then you know, the sooner can prove that it's either going work not work, better it's. Yeah, but it's pet project. But also, it's pet project. But also, you know, again, people don't like hear this. So I keep saying it. Sorry of the person who owns money is real product manager. All day I'm not disagreeing with that. kno. Yeah So you're not kill that. You're not kill that thing. mean usually what happens is that person gets gone somehow. And then project goes away and then team goes away too. That's team working on Yeah, you're right. Absolutely. So what think about this one? decision with no single owner defaults to whoever has the most to lose, not most to gain. I'm going down the ego route [UH] definitely. Yeah, yeah. if we think about this for second, I'm sure can come up with decisions for products where this actually happened. Right? They went down cul de sac and never came back. Yeah, yeah, I like this. The whole let's keep on thing like I don't really like It's, it's, the org is saying that we don't really know who decides this and we're going to kick this can down the road. That's what sounds like me. sounds like the Clint podcast of we're not, we're not really sure what role intersects with whatever role. And also sounds like don't have great product management practices. That's also, that's probably very true. If this happening frequently. Right? Yeah. All So let us know comments if you have a pilot that's been that's been almost decided and how long it's almost been decided How many months? How many years? I don't know. So you tell us. All right. I know. I know. I need a winner is what I'm saying. Pick winner So we talked about the pet pilot surviving on authority alone you keep is bet and the cost is maintenance. But[UH] Every feature that's not the real cost [UH] mean is real cost because it's real cost also not only cost. Yeah, other cost is what you will never build because you kept that feature around. you can't see one feature time. OK, can't make decisions with being super close to seeing one feature time [UH] you, you can make better decisions from a strategic altitude. You need be in that rarefied [UH] you like my macho man fingers. Yeah, yeah So you need, you need be the product room where happened. right. need you and you can't do that when you're getting handed road map. OK, absolutely. Yeah, That second cost you mentioned, that is the opportunity cost that refer to. Right. You're not building something that you should be building. Perhaps. That's right. That's right is great this great like title slide. like the portfolio view is the only view like I like. I really feel strongly about this one because I've been bunch shops where know, given just the feature that we're talking about and just the customer we're talking about. Yeah, makes total sense. But if you were to like mouse wheel back three times now suddenly nothing makes sense anymore. what **** talking about? Like what talking about? Like the the the hey, just just just ship what sales needs. OK, sales says they need a PDF feature for this customer to able to land the ship. Like how what going take two, three days? Just get done, Brian. Why, you fighting this? Why why you not team player? not team player? You're being difficult. Yeah. So sales won't say, know fax this in color type feature. Right. Right. Nobody got color faxes but that's OK. They want to give them because this happens frequently though at shops where they're handed features from you know hand down like go. You're build this. I feel you just, you just highlighted exactly what I'm talking about. It's like you want me to build a fax feature which [UH] which which turn requires me build a time machine to jump back to 1972 to able fax somebody [UH] get out. What you talking about? How what fax you kidding me? How about just email you instead? You know, this email fax service where just email and then it faxes the fax to you from the email inbox. And that is 25 year old technology. My goodness. Yeah, yeah. just had to ship it. Sales as they needed to ship it and get out my office. There's just no, no bigger picture you there. Right? Maybe one customer did even ask sales for that. You never know. But, you would lose that fight 100 percent time until you come to a portfolio meeting where you're like, I got build fax feature and everyone looks you and you're why are you looking at like that? Who you? Yeah, exactly[UH] are you? And 1975 called they want you back, get back, go back to the future whatever. yeah, they're weird because your sponsor whoever is going that meeting and get torn up by the people with the funding. That's true [UM] think from a financial viewpoint, a portfolio approach makes lot sense as well because obviously you know money not unlimited. Right So where you invest? You can't look that unless you're looking at the portfolio. Otherwise it's be my feature, your feature, her feature, know, his feature. Right? That doesn't make sense. Yeah. So the academic here is going say, well, we just have to score every feature with rice or whatever other things. Like 27 different whatever other framework. Yeah. Whatever framework. Just ship need or the academic square feature on rice. score every feature on rice like frameworks like that. As I, soon saw academic, academic who also by way, the academic in the steelman. We will sub out for the CPO, the chief product officer or VP of product who's never done product in their life. Oh, you need give me a RICE score for everything or this other framework that I read about this week, they will sub these things in because have no real product sense. They've never had to exercise a skill. That's how you know you're dealing with an academic, someone who's never actually experienced. Right. Right. And is a lot a lot this category. A lot. But sadly, in broad space, you know, the funny thing about the steelman here is the practitioner's point. listen, we're not we don't decide strategy anyway. What do worry about? And as practitioners[UM] it really is not. It is that someone hands you backlog or someone hands road map. Yeah. And you're just to go run with it. Really? Yes. Oh my God. Yeah, I work lot people [UH] product managers like this. Yes. Yeah, I have to. Sadly, sadly, sadly right because I'm like, yeah, but if you just stand up[UH] usually you just stand up once or you make case once and you really hammer at home, like I'll back up. The other product manager back because they know this is right way to do it. They know you need get evidence. You need be testing your assumptions before you roll them in and and and can make some real progress or you just don't fight [UH] What is it? always just don't have option to speak[UH] Yeah, understand that. But if you're working with more senior product folks, they'll, they'll like much like the volleyball like highlights. They'll, they'll put ball up for you. I sure will. You know, they'll ask the questions. They'll ask the uncomfortable questions. And if I'm do any YouTube isms, they'll ask the uncomfortable questions in those conversations. No one else is asking that. No one else is asking me because because the academic doesn't want go. Well, we got a score feature with rice. what rice is score the things most impactful low? It may well do that because you're not[UH] you're just simply going with objectivity of some kind [UH] academic objectivity probably and not exercising like said your product sense. Right? So [UH] the rice framework breaking on rice framework. I just want to point out The rice framework is reach impact confidence effort. So in the reach it's you measure how many people will be impacted by the product or specific feature within time frame impact is you gauge the level the benefit or change an idea will bring users rated on scale from minimal to massive. Confidence is reflect how certain you are about your estimates for reach and impact, which is totally made up way and effort is ating total time resources required to execute the idea, which again, totally made up. Of course, of because again, whatever you tell me is time estimate. I'm to cut that half and report that up. So yeah, yeah [UH] that's how real companies work. This has been, this been real this. I think started on the practitioner of like, hey, you're not even strategy room. Why, why would you worry about that? And then we ended in the real, real world where we stop getting real start getting real. is real world? What's the real world? real world is where everybody should be strategy room. That's the real world. mean, it is [UH] like startups small companies and stuff like that. Like the real world is we, we make decision and we're all room. We all talk about it. Yeah Yeah. Crazy. we have? I feel like have some better points we've waited 45 minutes talk about. We're not the room. Is the problem to fix? Not excuse. I fully concur with that. You're wrong room you're not that room[UH] the room where happens. That's talking about. Yes, that's right. me, me go get Thomas Jefferson to fix all this. if you're not the room when you're talking at the portfolio level, the funding portfolio level. Boy, like I don't be one Oh, I'm make because don't be one to tell people like if you're not the room when they're talking about portfolio level concerns and you're in product that I'm not say don't matter. I'm just say you need you should ask why you're not that room absolutely should somehow find yourself in that room. OK, whatever takes because really that is place need be if you're in product. Otherwise[UH] dareI say, keep that resume updated. You should, you should say it early and often. Feature It's product cut out the only conversation that, that matters. Yeah, yeah, yeah. That prizes them. Yeah, yeah, yeah. I a, this is a bad category. mean like it's good category. It's good category that we're talking about like reviewed one time. You probably could lose. mean or win or whatever. Like sometimes winning losing in product. Hey, got everything ever wanted and nobody uses feature anyway. Right? Now you're the problem. Right? Right. So did you really win[UH] I feel like win when lose [UH] Waterloo. So don't know why I'm singing Abba right now. I think the takeaway here is to stop defending features one a time, price the whole portfolio. And if you're not the room, you to fix that before you start getting any kind of like kill this feature, kill that feature kind of list. Yeah. You're not operating strategic level that if you're not room anyway, you're just simply building stuff. That might wrong stuff, but you're building stuff. The room where it happens. Yeah. For last time. So in the comments below, are your product people in a room where happens or do you handed the roadmap until and told just get it done. Just get them done. Right So every trap so far is the same trap. You built it before you tested it. Let's fix that. Let's do that next. All right So as you may have heard, you never had problem killing features. Every dead feature traces to one move. You build it, hoping to find a customer later. You build it, you build it and build it and that they will come. That's right. That's called hope driven development hope hope driven development is the real thing. got two fixes. got two fixes. got tree fitti. Sorry, no, don't. don't. I have to fix this. I have two One, no idea gets built out until it fully earns it with signal. And to every feature you own lives in one of three phases. Exploring, exploiting, sunsetting, expiring. Sorry, I'm sorry. going the next Oh yeah. Oh, are we exploring exploiting expiring? Nice. Nice. I like that. like that. That's thing. there's no phase called kill and there's no phase called Limbo. Oh, that's good phase. right. That's right So feature you can't name face for is one that's killing you. Boom. But listen Om, oh, we got AI now. We can just ship it and see. We can just ship it. You don't need listen. You don't need all this stage gate every feature stuff that the academic talks about. don't need all this. We can just ship it and see. We can test it in the real world this is this this spray and pray. Right? Just ship it out there. Hope something sticks. Daddy Dario told me. I don't need, I don't need to be the room where happens anymore. can just fire forget. Well, he's not room or happens. Maybe that's so weird hope hope Kimi K3 not upsetting Daddy Dario right now. hope he's losing lot sleep over right [UH] So you're just simply doing that and if something sticks, whatever you're shipping out there by using AI isn't stand upright anyway. just looks like it looks like feature It's not to work like feature. Right. Because you didn't put in all of things that make it successful. Yeah If had built it intently, like for example, all ilities. Is scalable? Is it safe? What movie was that from? That's Rain Man. It's definitely not be. I think it's Lord Rings on there. it's not be first all because shipping it is. Oh boy. Oh boy. I want like thought about during week between when we knew going do this podcast. I really thought about had a whole different podcast. I was thinking about the whole time of like just pat the vibe coding kids head be like you're so great. Like I like you. You're cool. Like I'll my top pocket take home. And then grow so great because like when you ship something to see works, you don't to worry about maintaining it. You don't worry about scaling it. don't worry about if like every customer uses it every day, every hour every day. And mean, have worry about like database contention locks and stuff like that. You just ship it and throw it out. Oh my God good, good, good boy. You're good boy. Have a dog. it breaks, go build something else. Yeah, yeah, yeah. Just move along like never mind all that rap damage you're going get right you know, don't worry about Listennobody answers the customer[UH] like tickets anyway. Like we wait for marketing to say that we're going viral for people saying that our service sucks. Like that's what wait for to fix bugs now in the modern age like that. Yeah, that's the saying to ship it, to ship and see we We got AI, don't worry about this. Funny as I actually know people like that they come to some meet ups. Just ship it. We don't need to make it work. Just ship it. Make it work as in like real world. I mean, I get is like before you would test the feature by making a wire frame and walking through it like like like magic. Yeah Like clicking a click through wireframe. Yeah, that looks like your app. The customer can't really tell that it's not completely functional shipping. But that's difference. Yeah, but it's, it's Figma. It looks like your app. They don't know. And mean, you get that aha moment. That's what you're looking for. You're looking trigger. That's how you should be getting signals but not. Here's product I just whipped out through vibe coding. And you know, there go. It's released Are we whipping things out? that we're doing now? guess we need calm down using AI too who. I think the, the practitioner here is the one have out of all these. It's one that will take most seriously. That's what I'll say. Someone I'll take most seriously because practitioner is to say, listen, we have roadmap and it has dates on top. I know it's a gantt chart in disguise of a roadmap but it has dates top, not phases. OK, I have to meet these dates. So get my office. Like I understand you saying I need test idea, but I got dates. got deliver to. So deliver those[UH] you know, deliver those things by those dates. But then are they land? Because don't know. Right. So that's the problem roadmap has dates. DAtes have habit of just arriving. They will arrive and you'll deliver. But then what happens after? don't know who's working that place because I'm not working there when the roadmap has dates top. the phases we did. We did podcast about this Wow. Arguing agile one 93 product life cycle phases from MVP to sunset. a guide for product managers. That was very long title by way. Yeah, a while ago too. But was, was December 11th, December 11th, a day for which we'll will live no, said. No, that was December several of but was arguing 193 December 11th, 2024 where we talked about the different phases which should the phase that we coined here well. There were different phases by way. But I'm click because maybe put the MVP growth maturity sunset. had four phases MVP growth maturity sunset. but these phases don't apply to entire product. They could apply to a feature. That's right. You get MVP a feature and you can MVP feature with Figma. Yes, yes, kids. Figma still thing. You can still use it. know, Claude has artifacts whatever now you know, cloud design whatever have now. But you can still show basically clickable screenshots that look like a product. The customer won't able tell. Right. But enough nods from customer, turn around use Figma make to get you all way up to the [UH] you know development level or any whatever. Turn that design into code. Whatever your favorite AI product is you can use that. So your roadmap has dates. But I guarantee your roadmap, this a product level decision. it's like it's not really roadmap, but You need be at the portfolio level like we talked about the last category to step through at the portfolio level. Is this investment meaning product right place we're spending money is it MVP meaning like premarket? Is it growth? Is maturity or it sunset? Right? Growth means we're still spending bunch money. We're trying acquire more customers. We're trying expand it. Expansion growth. Or is it maturity? Which is I've been on lot products maturity phase which is like cost cutting. You know trying keep cost in control. Right. Try not to spend too much money. A lot capitalization happening in maturity phase or is in sunset. You know, we're trying run an enterprise with two developers. Yeah. Been there to hopefully you're you're weaning customers on to the next product. That's right I mean that's point sunset off ramps. You know, sunset is highway that goes to nowhere eventually with bunch off ramps. Hopefully, hopefully. But you're right. asked. That's right. That's right. And end road just the end of road. Boyz II Men are nowhere found. That's right So have some better points? All right. have some good takeaways here. I think what I want to leave the listeners with, I think leave the listener with here is you want to move your organization product, you want to[UH] if you're in on move your organization towards testing each idea before you build it. That's, that's kind of the whole purpose of podcast. So if I have introduced this at pretty much every organization I've been at that I've been product capacity is before you decide that you really build something. You need to explore as a test, which makes your qa people more valuable because they can, they can participate in the pre software development testing of ideas. Yeah, And after you build a thing, they can be involved in the metrics and the test, the actual software testing, know, afterwards and the normal traditional stuff think and the gathering of metrics the like the business launch even. Yeah, post launch kind stuff. So be involved lot different places And it's sad that like the QA as discipline kind going away because they're so flexible on the different places that can be [UH] that they can help with. Yeah. Yeah. Absolutely agree with this first point you just made about getting the signal first [UH] signal no build I like that unless you're being told by the CEO. Yeah, and that is same thing. mean, but completely understand. But if you go and sit down and treat them like person who this is not their like software development is not their area expertize. Absolutely. And just walk them through. Hey, have this idea. Here's the typical way that we would test this idea against the market. Walking through, know walking through lean startup basically what I'm saying. Yeah. Walking through the organizational of the goal like your operations folks should be able help here too because they should be experts on goal. But your product people haven't readily set up and your operations people haven't read the goal. Like I don't know, like go keep resume updated. Like don't like that point, you know, don't know like the kill here of like oh this idea isn't good. going get of it. Like that's that's it's sunsetting. That's not really sunsetting. I'm describing. I mean could be sunsetting of like there's no money this anymore. extracted all money or the market has commoditized and moved on. But it's sort like we want run tests on the idea to make sure that we're not making a bad business decision on one on on the front and on the back end. It's we want to have an off ramp to other products or other, know, services whatever. Yeah. And your your customer support people are great conduit for getting those signals to. Right. Like what the real pain points now that the products mature and people are using it So you don't miss out on those avenues as well. Right? blaif there's takeaway here stop asking when kill features. Put everyone into the appropriate phase exploring exploiting or sun setting or back and listen to arguing agile 193 the MVP growth maturity sunset. Those are four phases. I actually like those better than se are. But listen, whatever can get started with, do that. OK, whatever get started with. like having MVP phase in and of itself because MVP is like a, like a pre explore phase. It's like haven't built software yet. I'm testing the market. Yeah, think some people get confused with that. They actually think MVP is a first version of the fully functional product as opposed to just enough product there to get signal. Oh boy. Well, if you if you like MVP, I can suggest you arguing agile 141 from December 6th, 2023 [UH] minimum viable product. so every feature that you have, you can label it with one these MVP sorry. You can label one of these phases. One of these phases we've outlined other the the explore exploit would you say expire or the other, other, know, the MVP whatever other phase that we have And you want to get them on real customer signals. Yeah, And then you know, sunset by date and pick date. Pick date. Listen, we pick date to to implement this stuff. not pick date to get rid it? Just saying. So think date will probably driven by the signal, strength signal anyway. Right? So you're not like pulling date air. mean, listen, don't make enough money off this feature by X date, just cut it. yeah. Your attention better off on another feature another area. Maybe, maybe, know, maybe so we have a final sticky for you. That's do. we opened with the answer gave on the interview. Yeah. Is one kill feature like that's a tell that have other issues going on. OK, yeah. So now that you see it like the data is hiding like the real fix for your, your culture, organization, whatever. so we have few things over the point this podcast get get you this hope driven development. to give little sticky. You can put here the side your monitor. can look at it next week as make your product decisions[UH] Number one, write the one metric that would kill this feature and want to check if it's game-able. see someone can manipulate that. Absolutely. Number two, want put a decision date and one specific owner on any pilot that you're doing. Pilot also includes we want to test this feature or non feature against market. Right? Because maybe maybe you're little more advanced and you don't build features for us. You testing against market. OK, cool. But who's the decision? The single decision maker. is the, this is the we're going Amazon, this the single threaded leader. Yeah, yeah, yeah, the person. OK, number three is you to name the bet that you're not funding by continuing to spend effort on this. This one I like lot. We didn't talk about this one. Yeah, yeah. People forget this one all time. Like this is not very common that they name bet. You're not funding because they're just so laser focused on the one that they are working on. Right. Right. This is very powerful. Right. And then last one[UH] phase every feature in [UH] whatever phase that like. We talked about two different ones. I'm leave one here that says exploring, exploiting, or expiring slash sunsetting. Right, right [UH] I like sunsetting. I just like, I like it because it's like we're rolling the sidewalks up. We[UH] we're out here. OK, That sun's come back tomorrow. Oh no, tomorrow. OK, so If you can think of feature that needs the snip, let us know comments I mean if you're not scared. Oh well, if you're there the comments[UM] by all means, please like subscribe.