How to run customer discovery (aka: surfacing PULL fast)

Hey, Rob. Morning, man. How's it going?

Good. How are you?

I'm doing great. Excited to talk about discovery today, but we are also wrapping up PMF Camp. You've got some interesting things going on with AI, Rob, like in general, like what's, what's new? What's going on?

Yeah, so I have a lot of fun things going on. I think I wanted to, we have probably some new people listening now, and so I wanted to share like a little bit about What the heck are we doing here? Because this is not like your typical podcast. Like, what we're doing here is I am trying through kind of all the things I do to figure out why customers pull products out of people's hands and like how product market fit works. And I'm kind of continually trying to develop a set of thinking, theory, practice, all this kind of stuff and kind of like ways to approach this. It is not, you know, it's not complete. still very much in progress. And so what we do here and in the newsletter is try to figure out new things on that way. And so you are here to help me try to explain things that I am running into and things that I am writing from a, you know, otherwise I will like go insane doing this myself. And so that's what we're doing here. And today, this newsletter, I think, is kind of Kind of interesting. It covers a thing that I find a lot of people get stuck on, which is like, yeah, I get this whole pull thing. I just can't figure out how to ask about it. I can't figure out how to surface it on a, in a conversation with a potential customer. Like, what are the exact words that you use? And give me as many different permutations of it as possible. And what am I doing wrong? And so hopefully this newsletter, like it gives you at least 30 different ideas and like kind of like a pathway to follow.

Yeah. I laughed out loud when I read the title of this one, 37, excuse me, 37 Ways to Ask About Pool. It is not a hyperbole. You actually do have 37 different ways to ask about pool. However, it's not as overwhelming as it first sounds. And we're going to get into Uh, talk, we're not gonna go like one by one by one unless we decide to. I mean, we, we could, we could just go one by one. No, I'm just kidding. Um, uh, but what we'll do is discuss, uh, a couple different scenarios that founders might, uh, might find themselves in, um, and how the approach to really discovery, to unearthing customer pool in a way that's helpful for you, that's helpful for the customer. Um, but why don't we, why don't we start by establishing a, you know, maybe a little bit of a point of departure. I wanna hit, You know, how do people, you know, what are some of the common failure modes or whatever that people do today that you see? But maybe like before that, who is this one for? Like, what, you know, why is this one on your mind?

Yeah, so this one is, it's just like a thing that has been bothering me for a while in a bunch of sales calls that I've been watching. And I've actually gotten some feedback recently from people who were saying like, I didn't actually know what the purpose of discovery was. I remember someone was like, it's just the thing you do before you pitch your product. It's like, you're supposed to make them comfortable or something like that. And so I find discovery is, sales in general is totally misunderstood. Discovery in particular is like this thing that I'm not sure why, like, I don't know what I'm doing with my hands. In discovery, like, it just, it's a, it's a thing that we don't really understand. And so what people tend to do in discovery is like, you know, in the beginning of a sales call, you're supposed to ask some questions and you're supposed to get some answers back. And there's like lists of questions online that you can ask, but it's not super clear. What are you looking for? What's the difference between discovery and qualification? And so what tends to happen is I, as founder, just like have a list of questions because I feel compelled to ask questions. And those questions are like, tell me about your current process. What, maybe there's pain points, right? And like, how would it ideally work? And we surface a bunch of things that sound like they are relevant and are actually not relevant because, right, like you could have an inefficient process, but it doesn't matter if you're not trying to change. And so what I found is discovery is— what is discovery? It's about surfacing if they have pull, and if so, what is it? What are they trying to do that they can't do that they need help with? Which sets you up to describe your product in a non-offensive way to them, in a way that they will actually rip it out of your hands. And you have to surface pull. Because if you're not even, if like most calls that I watch don't even try. And so if you don't try to surface pull, well, it would be like the person would have to on the other end be like, yeah, I know you're asking about that, but like, can I just interrupt you and say this isn't super relevant? What I'm actually here for is XYZ. And they're not just like not going to do that. And so there's a lot of waste. We miss a lot of things.

Yeah. And I don't know where the prevailing theory of how to run discovery comes from, but I suspect some of it is, it's on the surface level related to the ICP. There's a lot of qualification, like somehow, like, and in particular, some of the qualification stuff is basically irrelevant until you have your case study that you're repeating. And I think what I was doing with my early sales calls was I was trying to qualify based on like what I sort of thought the person would look like or the type of organization that would buy.

Yeah.

from me without actually knowing why they would buy from me. It was like, it was more like demographic level. And I've seen that in some of the, uh, some of the teardowns that you've led as well with other founders making those same mistakes. But you list, um, you know, 3, uh, ways that people run discovery today, all of which are basically incorrect. One is current state questions, another is ideal state questions, and then, uh, you mentioned specifically qualification. Questions. Rob, why don't we, why don't we hit those real quick? Not to go super deep in them, but how do you know if you're doing any of these things and why are they the wrong approach?

Yeah, so I'll give a couple examples here. Like, current state questions are, tell me about your process and your pain points and your team and your da da da da da. Like, what is the current state? Ideal state is questions of, how would you ideally like it to work? Like, how would you ideally like your current state to look like? And then qualification questions are, you know, do you have budget or things like that, right? Like, are you, have you, you know, whatever the questions are of like, are you serious? The problem with current, we'll talk about current and ideal state questions, is that those aren't the relevant questions to focus on. What we're, they can have an annoying current state. They could have a very specific future state and have no desire to to exert any energy to get from current to ideal state. And so in that case, you are learning about like static versus dynamic, right? You are learning about, like, not, you're not learning about what force they are trying to apply right now, which is what we're actually looking for when we're asking pull questions. It's what are you trying to do? Where are you stuck? Where do you, what's the progress you are trying to make?

Yeah.

And then on qualification questions, right? If you ask somebody about their budget before knowing what they're trying to do, before they can articulate that, they're gonna say, my budget for what? Like, what are we talking about here? And so those, it's like super, so you're like, there's BANT and like MEDIC and MEDPIC and all these kinds of qualification frameworks, which often have like need within them. But But like the need is what causes the other things in the framework to be relevant. And so, you start with the need, with the trying to change, but unable to change right now.

Yeah. And the current state questions, you know, some variations of that or permutations of it, like it can be about size of the organization, like channels that they're selling into, like processes that they run or software that they're using. Um, you know, an ideal state question has the feel, has the feel of like, it's almost like you're asking that question, hoping that they are going to describe your product and then you're going to like pop out from behind the curtain and, and really—

Or there's like, oh, I didn't even include, there's a category here of like gotcha questions, right? Where I'm trying to logic you into wanting my product. Or I'm trying to like devise a list of questions where if you say specific things to these questions, right? Is blank a big pain point for you right now? Wouldn't it be better if like X, right? Therefore, you like, it's, you've basically logic'd an argument there and they can just be like, yeah, but also not trying to change here. Like, thanks a lot. Peace out.

Yeah. Yeah. I feel like that's like every Hollywood sales script that, you know, has the pre-canned like result of the person being like, Oh my goodness. Like, you're totally right. I should buy that. Well, let's jump to, now we've established what not to, or like what, you know, some folks might be doing. And again, this is mostly on the pre-PMF, like, but I, correct me if I'm wrong here, Rob.

No, no, no. I wish it was mostly pre-PMF. Like, I've been helping companies, $1, $10, $15, $20 million plus ARR, and like, they don't Figure out what customers want in the call. Customers are often buying despite what's happening in the sales calls. And so they have kind of like a list of questions you're supposed to ask. They don't learn what pull is. And so it's just like a, I don't know, like, like there's a percent chance that people will buy based on people who have previously bought. And I can't, right? You, you just kind of like, see, do they show up to the second call? Coin flip, right? Unclear because we're not actually surfacing the thing. And it's crazy. Like you can, you can, you can basically scale a sales team with like decent close rates, having no clue what pull is and basically being counterproductive in these sales calls. And so I've kind of like actually been intervening in a couple of these where it's like, it's like, oh, wait a second. We don't need to be asking half of these questions. All we need to be asking is like, this kind of question. And if we hear this kind of thing, cool, they're not gonna buy. If we hear this kind of thing, they are gonna buy. And it's like, makes everything way simpler. So I wish it was just pre-PMF, but I have been, I assumed it would be, but I've been shocked at what happens in sales calls post-PMF.

Hmm. Um, I'm gonna hold on. I've got some questions about the post-PMF thing, but I think what's most helpful here is like, Rob, I want you to help like basically Fix the way that I think about discovery. And there's—

Yeah.

And what we're gonna do here, like, let's just start at, you know, the highest level. Like, what we wanna know is what are we actually looking for in discovery? Like, what's the ultimate objective? And honestly, like, there's probably, there's, you lay out 37 different ways to get there. Like, everybody's gonna kind of figure out the right combination that like matches them. But, you know, you say that you're looking for 2 things. One, what are they trying to do, the customer or the prospect, right now, but they can't do with their current tools, methods, et cetera, such that they would buy something right now? And then 2 is, what would they do right now if there was a way such that they'd buy something right now? So both of them have this result of purchase-shaped movements.

Yeah.

I don't know about aspiration, but like motivation. What is like going on behind these? And can you actually help me understand the difference between like that you're teasing out between the two?

Yeah. So it's, this is probably just more of a difference in my own head than actually needs to be like spelled out in a ton of detail. But basically we're trying to fill out the pull framework. We're trying to figure out what's the thing they're trying to do right now that they can't do with their existing tools, methods, options. That's it.

Yeah.

Right. And why can't they do it with their existing tools, methods, options? Right. It's, I'm trying to minimize the amount of time I'm spending on meeting notes for client meetings. And that's unavoidable because if I don't take meeting notes, I'll be thrown in jail. And, but like, I have so many other things I would way rather do. So every time I take meeting notes, it's like a little part of my soul gets beaten to You know, whatever. And so that's, that's, that's pull. And I can't use a generic AI note taker tools, methods, options because it's like unclear if it's compliant and doesn't integrate with my CRM and da da da da da, right? Like, and so you have, we're trying to surface that kind of thing, trying to do X right now blocked. And there's kind of two categories of that that we've talked about before. It's the thing that is actively, like, if I think of a Trello board. There's like the backlog on my Trello board of all the things I would like to do, but I'm actually probably not going to do. There's things I'm prioritizing right now that I am totally fine doing myself or with my existing tools, methods, options. Cool. Neither of those categories are relevant. There are 2 categories that are relevant, which is trying to do right now, but can't do with my existing options. And so I am kind of like grinding against them. That's the like jump example. And then the other one is like, would do right now if only there was a way, right? But I don't have a— like, if you watch me, I'm not like— it's the Lovable example. Like, I wasn't making a website before Lovable because I just didn't wanna hit my head against the wall with one of those like point-and-click editors. I have no design sense. Like, none of that will work. And so you wouldn't see that, but it was on my to-do list in this category of like, if only there was a way that was not such a hassle to make a website. I was in both of these, you're looking for like the coiled spring, the person trying to do something or would try to do something right now, but existing options are not good enough. That's pull.

Yeah. And importantly, pull is not the amount of pain that Somebody is experiencing. It's not a problem. It's not a big problem. It's not a business case. Like, I think there's like a, you know, whether a founder or somebody uses the acronym PULL, like we've seen folks that have an intuition about PULL. It's like they're looking for something that somebody would pay somebody to do.

Right.

And like, you can build a business off of that kind of instinct. But the popular dialogue around what are your biggest pain points that you're experiencing right now like that? And if one of the pain points lines up to my problem, like there's this aspiration that results in the purchase. Like I think that's probably a common one. There's probably others as well. Let's talk about pull. You write something really interesting and you've been talking about it more with the founders That you've been working with in the last PMF Camp, which is about the difference between sequencing your— I'm going to use the word interrogation. That's not right. It's like questions around pull, like constructing pull piece by piece versus a holistic question or a holistic discovery that unearths pull that is complete at one moment. And, you know, just real quick for if there's any new listeners, pull is an acronym that that Rob came up with that we use to understand the circumstance in which a customer will purchase from you. And it's, there's a project that's on their to-do list. It's urgent or unavoidable for some reason. They're looking for options, but those options are lacking. We can link to that in the show notes if you want to dig deeper into it. But, you know, one approach to that, Rob, is, okay, what are the projects that are on your to-do list? Why, you know, choose XYZ project? Why is that unavoidable? What are the options that you've looked into? Why, you know, why are you satisfied with those options? Like, why are they lacking? I've run, I've run calls like this. So I need you to help me understand, Rob, why, uh, why does this not work? And then we can orient to like, what's a, what's a better approach there?

Yeah. So, so the most common error mode I see in discovery is PULL framework has 4 letters. So I'm going to ask about letter P, then U, then L, then L. The challenge with that is that, think of, think of a product team's roadmap, for example, right? They have a, a product team's roadmap is a bunch of projects they are going to do. There are reasons why each of those projects is unavoidable. They're on the roadmap. They have dates on them. They have specific reasons for them, but their existing options for getting basically all of those projects done. Are do it ourselves, and that's good enough. And so what you, what you realize when you look at projects that way, and if you approach this sequentially, is there are a lot more projects than there are pulls, like completed pull frameworks. And so if you just ask about projects or priorities, you're probably gonna find something that has good enough options.

Hmm.

So what tends to happen is you ask about priorities, you find priorities, and you find out they have good enough options 10 minutes in. It's like, okay, what do we do now? Like, I, let me just Hail Mary a demo and see what happens. Right. And so that's because you approach it as asking these questions sequentially rather than the opposite of this is not to ask about projects, priorities. It's about to ask, it's about asking about all of pull at once, which is not a, what are your priorities? But it's kind of like, what are you trying to do where you're stuck and your existing options are not good enough?

Yeah.

It's like you kind of fit the entirety of the pull framework into each kind of question that you're asking to get something that is like, oh, okay.

Yeah.

So it's not just any priority. The thing I'm stuck on and haven't been able to find a way is X. And then we can fill that out in more detail. But what, but like the opposite version of that, of just asking about priorities leads you to not fill out, leads you to find out that It's like, our existing options are good enough. No pull framework. If you just fill out P, U, and L and not the second L, guess what you get? No customers. It sucks.

Yeah. And you build up, you know, we've talked about how we anchor on things and it's like, if you head down a path, it's like increasing optimism that that is actually like, oh, I found the thing. It's unavoidable and it's a project in particular. And then, you know, you end up with the weird outcome of being ghosted, or it's just like, it kind of peters out because you actually, you didn't ask about it. Let's talk though about how you surface it holistically. You give some examples here that there's— we'll get into those. I feel like the one that you've consistently recommended that I use in literally 100% of my calls now is like, why did you take this call?

Yeah.

Which is like kind of a, it gives them an opportunity to surface pull without you even really asking about it because they might just volunteer. I've been looking for this. You proposed something that caught my attention and it lets you dive straight into that project. Is there anything else that's valuable to add on that kind of like the, on the, why did you take this? Why are you here? Why is this worth your time? Like, what would you, what else would you add about that?

particular opening?

Yeah.

So, so again, what we're trying to do is we're, we're giving them space to explain their pull to us. What are they trying to do that they can't do and need, need help with? And so the, the, um, first, so I, I think of this, you're, you're trying to surface something pull-shaped. It doesn't need to be super specific at this stage, but it needs to be, oh yeah, we're trying to do this, but, but specifically stuck here. And then you can double-click into it. So you're surfacing something pull-shaped. And I think of it as basically like you have the— ideally, they say when they get on the call, they say, Rob, shut up. Here's what I'm trying to do. Here's what— like, and they just fill out the entire pull framework for you.

Yeah.

That would be nice. But they have some reason in their head for taking the call and you wanna surface that. It might surface pull. If it does, cool, just like fill out the rest of the framework. If it doesn't surface pull, then we have a set of fallback questions we ask. So it's like, okay, cool. Why'd you take the call? You reached out to me.

Cool.

That's not super helpful. Then I'm gonna take some, some of the fallback questions here.

Yeah.

And so that's the way to think about it is like you, you have a couple different, you have like a series of chances to surface something pull-shaped. And then double-click to kind of fill out the details of it. And so why did you take the call says, is it, and I have like a bunch of different permutations of it. Like, I'm sure you get a bunch of people reaching out to you. Why'd you decide to take this call with me? Right? Like you can have a couple of different permutations of that question to see, is there something pull-shaped that it surfaces?

Yeah. Yeah. And after you do it a couple of times, my experience is like, You'll, you might cycle through a couple of those.

Yeah.

Um, uh, like, is there something you're help— you're hoping to get out of this? Like, and one of them will start to work like a little bit more frequently than, than other ones. Um, let's talk about some of these, uh, or there's, there's another template that you said, I'm just gonna, I'm gonna read out. I'd love your reaction if there's anything you wanna add, but I just thought it was really very direct. Um, and there's, there's one thing in there I wanna dive into a little bit. Um, and you write, you know, this is for the example of somebody that runs QA, uh, or is selling like a product or a service or software related to QA, you, you, you offer this one. As it relates to your QA process, is there something in particular you're trying to solve right now that you're hoping somebody could help with? And that actually has all of pull embedded in it. Rob, can you talk us through like where the elements of pull are there? Just like elucidate that for us. And then there's also the framing, which I thought was really interesting. As it relates to your QA processes or something like that, why is that there? Like, can you just kind of break this apart for us and help us understand how to use it?

Yeah. So, so this is, so think of if they say, if, if you get on a call with them and you say, why'd you take the call? And they say, I don't know, like I had nothing better to do on a Tuesday. Like, all right, cool. Then you go to fallback 1, which is to ask an open-ended poll question. Which comes before a closed-ended pull question, which comes before a Hail Mary, right? Like there's, there's kind of these fallback. And so the open-ended pull question is basically saying like, okay, here's roughly a pull framework for you. Um, fill it out for me. It's an empty pull framework. And so what I'm, what I'm doing here is I'm trying to include all the elements of pull. Is there something you're trying to do Right now, P and U, thing you're trying to do is P, right now is U, that you're hoping someone could help with or that you're struggling with or that you're trying to, trying to switch out vendors for or something like that. Right. And that's the Ls.

Hmm.

And so we have P-U-L-L. But if you just say that broadly, is there anything you're looking to do right now that you're hoping someone can help you with? Like, that's a very broad question. I sell QA software or I sell whatever it is. And so I kind of want to narrow it and say kind of anything within your QA process, anything within your accounts receivable world, anything kind of specific. And you kind of, you can get specific there. And then you can always later, if you get something there, say, but okay, I know QA is like a tiny part of your job. Is this actually a thing right now you are trying to do, or is this Right. So what I'm trying to do is I'm trying to get something pull-shaped in our domain, and then I can zoom out and see if it's like actual pull versus not. But I'm asking a question with all of the elements of pull in it.

Yeah. Okay. So let's say that, or maybe one thing to add on there, like that, you know, you talked early about you don't want to surface just priorities because they may not be looking at— I really liked the second half of that question that you're hoping someone can help you with. Like, that's, you know, immediately it like kind of flips the brain into like, oh, okay, there's like a potential transaction. Like, what do I actually value? It makes it so that it's more worth your time and their time. And, you know, like a broader comment about this is like that we've talked about in the past, we haven't visited lately, but it's like sales calls should be helpful and they should be focused on helping the customer achieve something that they're trying to achieve, but can't, rather than the, you know, I'm trying to like maximize revenue. Like at its heart, done correctly, you know, these, it should be like, oh wow, like you can actually do that. Like it should be almost like a relief if there is pull that you match. And let's hit the, you mentioned the case study poll question. I think you just referred to it as the closed-ended question. I'll read an example here and then, Rob, how do we use this one effectively and when is it the right time to use it is something that's on my mind because it's a more complex question delivery. What you wrote, the example you give is, usually when I talk with CTOs, they often say they're trying to get E2E testing off their team's plate, but winding up with an engineer spending a ton of time in Claude plus Playwright. Is that not something on your mind right now? I don't know what that is.

I put that in there just to see you read it.

All right, Rob, thanks for teeing me up for success right there. Anyway, so you're gonna— I know what Claude is, so that makes one for 2.

Yeah, so what we're doing— so, okay, so what we do first is we give them an empty pull framework and say, Fill this out for me. Is there anything here that you can fill out for me? This time, if we don't get anything from that that's useful, they say, nothing that's on my mind, just kind of curious to hear what you're up to, right? Okay, cool. I have nothing in the pull framework. Then what I'm going to say is, okay, cool. I gave you the empty pull framework. You did nothing with it. Let me give you a full pull framework and you tell me Thumbs up, thumbs down on this. And so what we're doing is we're asking a question that gives them a full pull framework and seeing how they react to it. And most of the time, they're— if you get here, they might say, oh yeah, that's exactly what I'm trying to do. Cool. Tell me more. Let's fill out in more detail for you. Or they might just say, eh, yeah, I guess more or less, which means no.

Yeah. What do you do in that? Or I guess, is that the, I move on to the Hail Mary? Like, if you're still, I mean, can you run through, let's say that ideally we've talked about this, you have one case study. Frequently founders will have a couple case studies that they know exist. They might be running after one particular anecdote or, but they know that a couple people have used their product for different ways. Do you just like— Oh, okay. Like, that doesn't sound like it. Here's another, here's like another thing that a CTO bought from me for. Like, do you just kind of like hit a couple of them and see if there's something where the light bulb goes off?

Yeah. So, you can, you can, like, if they surface something that is totally irrelevant to you, you're like, oh, that's new. Or like, oh, we don't do that. You can then use the case study question to say, okay, that's super interesting. We usually hear this other thing from CTOs. Is that actually not relevant for you? And they say, no, that's not relevant because XYZ, and it's actually this other thing. And then you can decide what to do about it from there.

Yeah. And, you know, just one thing from, I had an interaction yesterday with a potential customer and that a thing that I had to find myself resisting was arguing basically with them. It's like, I have to like access this like Zen mode part of my brain that just accepts Even if I sort of disagree with their general approach or their conclusion, you just have to accept that in their mind, like they're at a spot where it's totally fine. And so like, you're just kind of like looking for the thing where wherever they are, as you talk about it, like applying force against the status quo, or they're looking for other options, whatever it might be, something where they're actively not satisfied. Let's hit real quick the Hail Mary. Is there anything that you would pay someone right now to help you solve or accomplish? That might be like the scariest, like most awkward question that you recommend we put out there. Rob, could you, like, why is that even a question that's worth asking? And when is that the type of question that you, that, or maybe if you could give us an example of like how, where, like where that works.

Yeah, so there are variants of that question. This is usually something we're in, kind of break glass in case of emergency territory where we haven't learned anything yet in the conversation. And it's getting to like 10, 11 minutes and, and like it's kind of starting to get awkward. And so what you can do here is you can just like ask, this is, this is like the just in case. I, I think of this as like YOLO territory where, where we're entering now, which is You know, if I don't ask this question, I'll feel bad about it, right? And, or like, I'll learn nothing on this call. And so you can preface this kind of question with like, I wouldn't be doing my job if I didn't ask, which is a great preface to be able to ask literally anything, right? Like, can I date your daughter? Or like, whatever it winds up being like. I've always wanted to say that one. Let me use that. But But what we're doing here is we're basically just saying, okay, cool. Open pull framework, nothing. Closed pull framework, nothing. Tell me if there's anything you want to buy right now. Is there anything you're trying to buy right now at all? And you can ask that a variety of different ways. What did you come to this conference to buy that you haven't found? That's a great question. There's also like the inverse of it, which is, okay, so it sounds like there's really nothing you're looking to solve or buy right now where you need help. And so, but like, these are, these are like, and I've actually seen these questions have led to kind of like massive pivots that have caused startups to take off when they feel like they've been pushing. You just like, you're just going down the wrong path and you're like, okay, cool. Is there anything you're looking to pay someone to solve? And I know like, I can name like 3 startups that have gone zero to a million ARR where they started like the They asked this question and the person was like, oh, thank you for asking. This is a thing. This is the thing. And that thing happened to repeat.

Yeah. And it's actually, you know, with your first startup, the customer like gave you the answer to that question, which you weren't asking, but they were like, would you, you, if you would have said, is there anything you would pay me to do? They would have said exactly what you just Alerted like, oh my goodness, like, yes. But instead what they did is they got tired of waiting for somebody to ask that question and they said, can you build this thing for me? And then that's when, you know, your story of $0 to $4 million ARR in 2 years. But hey, Rob, there's 2 other things I want to hit real quick. One is you haven't surfaced anything pull-shaped yet. You've gone through, you know, you've gone through that kind of like 4-part Or 4-step thing. It's like, oh, you're not sure what to do. You say that founders can either use inversion or playing dumb. Playing dumb is a little bit more obvious. It's like, okay, I think I don't understand what's— they've surfaced something pull-shaped. You don't understand why it's unavoidable.

You don't—

you're like, it's like, be skeptical rather than resist the urge to be optimistic about they're going to buy. It's like, Have the customer prove to me that they will actually follow through on this, that this is really something that they're looking to transact around. And so playing dumb is a way that you talk about doing that. Like, okay, I don't get it. Like, why is Claude not good enough? Why is Playwright, we're just gonna bring that one right back in.

Yeah.

Yeah. Not good enough or whatever. You know, in my case, in the media world, it's like, okay, like why do programmatic advertisement campaigns like not work for me? Why does Meta not Meta ads not work for you. Let's, you know, that one is just like, get them to say in their own words, very more explicitly than you probably are comfortable doing, like why they can't continue doing the thing that they're currently doing. Let's talk about inversion though. How do you, what is inversion and how do people successfully use inversion?

Yeah. So, so playing like one of the, I just remember a founder conversation where he's like, yeah, It, like, I asked them why Playwright or why Claude wasn't good enough and they didn't give me a good answer. And so I just moved on. It's like, no, you need to know, like, if you don't know why it's not good enough, you don't have a pull framework filled out that's legible to you. You don't know what to do about it. You don't know how to describe your product. You don't know how to position it. So, so these tactics are after you've kind of got something that feels pull-shaped. Fill in the— it's a way to fill in the details to be like very specific. And so playing dumb is one way, or just like being curious, like, maybe I must be missing something here. But then inversion is like, okay, why is this urgent right now? Is a kind of like one version of a question. The other version of the question is, is this something that actually needs to happen now? Or Is this something that could just wait? Right. And so we've inverted it from having them give you a positive reason why it's urgent now, from having them kind of like basically try to convince you that actually, no, it is something that is urgent now.

Yeah. And the way I did this in a recent one, just as another example, was like, honestly, if it sounds like they're all set, just like call that out. It's like, Wow, that sounds like you have, you're like totally set. You've got the expertise on your team. I don't really think that we can help. Am I, do I understand that correctly? And sometimes they'll be like, yeah, I think that's kind of the case. In which case you're not trying to push something onto them that they obviously don't need. Or they might say something that's more surprising. Like, actually, there's this other thing we haven't talked about. Like they're not ready to let you go yet because there is something that's kind of like on their mind that they that you might be able to help them with. It just hasn't come up yet. But, you know, it's almost like a push them away a little bit and see if they come, see if they reengage, like see if they chase you a little bit, like, or do something that's like kind of chase-shaped and then understand like what's going on there. Let's talk just like maybe, and this is a long one, sorry about that everybody, but how this changes post-PMF is also in the blog. And this is basically, you've got a case study, you're growing quickly. You give some examples of different startups that are post-PMF and the discovery changes. It becomes way more direct. It can even become supply-focused if your targeting is good enough. Like you give the example of one company that's, I guess they developed AI agents and they just list what the AI agents can do. And see if one of those, like the customer pulls it. But you have to have the, I think the risk here is that you, we want to jump to the post-PMF flavor of discovery before we actually have PMF, before we have a really strong sense for pull. Anything that you would add on the kind of like how it changes and when it's the right time to move to a post-PMF approach to discovery?

Yeah, it's basically when you can predict. With like almost 100% certainty that this person in front of me will buy. And you know what the cause is, what are they saying? What are they doing? What's happening in their world that is going to basically cause them to buy? Then you can change your approach because you understand causality now and you can just go straight for it and you can ask some more specific questions or do kind of like different things. You can do a variety of creative things. Most people here are not in that stage. So I just like, We'll lay it— lay out a couple of examples so that you, like, the important thing to know here is that once you've found pull, like a lot of the things that we talk about are methods that are, that are kind of like the principles carry over post-product market fit, but the, the methods themselves are kind of throwaway. They're, they're, they're things you jettison once you find product market fit and you design for like aerodynamic fit. to pull and to maximize the volume through your case study factory. And so after you find product markets that after you can kind of like really know, okay, I know exactly who's going to buy exactly when, your approach will, the specifics of your approach will change while the underlying, we are surfacing pull and testing it, is the same.

Yeah. And there I've made the mistake before of Mistaking having a strong anecdote or a strong belief in a hypothesis for having product market fit, basically. And I've directed my discovery questions at identifying if they align to an anecdote before, you put a very high bar there. Do you know with like basically 100% certainty that they'll buy if something's true? And I've jumped to something that was more, yeah, like a post-PMF approach to discovery. So I—

Too early.

Rob, maybe just like a real quick recap, because I think that it's like, for me, what I take away from this is know basically like how I'm going to walk, how I'm going to get through discovery to try to unearth or surface what Pool actually is. You give like, if somebody just like basically goes through your blog and chooses like, okay, here's like the 4 questions I ask this one, Doesn't work. I asked this one, it doesn't work, but they're, you know, having a sequence that you understand and what you're trying to do if pull is not surfacing. How else would you say founders should use this post or like use or rethink, reimagine their approach to discovery to have it be effective at surfacing pull?

Yeah, I think the main thing is like, one, actually look for pull. 2, ask about all of pull. And there, these are methods and different talk tracks and different approaches in here that you can do to do that. But if you have those 2 things in your mind of, I'm actually looking for pull and I'm looking for the whole thing first and then filling it, filling in more detail about it, then you're gonna be generally okay. And everything in here is like my suggested tactics based on what I have seen work, which is the why'd you take the call and then all the fallbacks. If that doesn't work. So yeah, just like take this as a, the strong recommendations are look for pull and look for all of pull at once. The kind of, you know, do whatever you want on the how you exactly actually make that happen. Yeah, but actually look for it.

Yeah, awesome. Well, we will leave it there. Any final messages that we have for folks?

I don't think so.

No?

Awesome.

All right, we'll catch everybody next week. See y'all.