002 How Much Should I Charge?

Okay, first of all, bless your little heart coming back on day two. Oh my gosh, thank you. Feedback, give me some feedback. Drop a comment, tweet me, LinkedIn me. I want to hear what you think. Today we are talking about how much should you charge for your accounting services? Oh my gosh. It's like the ultimate question, right? Ugh. Wrote a video about it yesterday. I've got some thoughts. Probably going to tick people off. Share kind of my mental framework for all that stuff right now. Matt Metris, bless his heart, put a bunch of language models like GPT to the test on preparing US tax returns. Just about the ugliest thing you could imagine. It did some of it well, did a whole lot of it not so well. And then I got a friend, I don't want to out him but they're like, holy hell, Microsoft looks really good right now with the AI stuff they're doing with Copilot that they announced. Should I move from carbon to vanilla Microsoft? Oh my golly. Let's talk about it. We're gonna talk about it on today's Jason Daly. Should we call it that? I don't. Okay, let's talk about setting prices. This is for some reason people are super opinionated about how you set prices in an accounting firm. And I think part of the problem is that social media just absolutely squashes any sort of context into dust. Like there's just, there's absolutely no context on social media. And for me like the right answer on pricing is so dependent on what stage your firm is at. And I think online we like to, online I think we like to signal an aspirational version of well, here's my really cool mental model for the best way to do xyz. Meanwhile, you got clients whose bank fees are broken and they weren't really respond to your emails. And the reality of like running a firm is usually a little more sloppy than that. But I think that's not particularly helpful oftentimes for people who are maybe still in the early stages of firm running haven't been exposed to that non traditional model. And so it got me thinking, why don't I do a pricing video that runs through I guess kind of the different stages of firm running. And here's sort of the framework that, that I landed on three different types of firms your day zero firm which is like usually a side hustle, right? Like they're kind of exploring it on the side. Maybe it'll become a full time thing. People generally in our space don't like set out and say Ta da. I just quit my job. I just started this business with no clients. That's kind of. I don't feel like people do that much. So it's probably a side hustle that is like tier one firm. Tier two firm is a specialized firm that's beginning to niche. Tier 3, I wasn't sure what to call it. I think I used a high leverage firm. Is that what I called it? Yes. High leverage firm could be. I don't know, it's probably a cooler branding there, like an infinite firm or like a firm that is enabled by the distribution of the Internet. So. But let's start in the beginning. Day zero firm. These people are price takers. Let's be real. They probably suck at what they're doing. You're really slow. Like, that's just where we all start. Because you haven't figured out all the ways to do it wrong. So the reality is, like, you're not unique. You're not one of one. When a client comes in, they're gonna benchmark your price against the person across the street. So you're gonna bill hourly. You just are. And I honestly think that you could actually win on hourly. When they say, what's your hourly rate? And you're like 100 bucks an hour, they're like, oh, great, the person across the street's 150 bucks an hour. I choose you. You may in the end end up paying more because you're really slow and you don't know how to do this stuff efficiently. But for people in that situation, like, they're just trying to put food on the table. When you are, I think it is easy to forget the mental state that you are in. And like, your first six months of running a firm, like, you are absolutely terrified. You're trying to figure out how to make this work, how to not have to move back in with mom. Like, it's. There's a whole lot going on when you're starting out and will take absolutely any kind of work. So honestly, like, at this stage, I don't have a problem with hourly, especially if you're still slow. Sounds like a win to me, right? But the more I thought about this, honestly, I think like 80% of firms are still in this stage. There's no specialization. There is nothing to differentiate them from every other firm out there. And so you're up against substitutes. You can't really value price. You can't really command your own price because everybody else doing the same thing. Like, how does the client know that you're any better than anyone else. So that next tier is you found a niche or three. I said I don't think you necessarily need just one. And it doesn't have to be industry specialization. It can be a bunch of different things. I think there's a lot of dimensions on which you can specialize. But the upside there is you've eliminated, you've eliminated substitutes. So you should be value billing. People talk about pricing calculators a lot. I'm of the school of thought that you should have an internal calculator, but that calculator is basically the minimum you would take a client at. It's not a hard and fast rule for what you will take a client at. It's just the threshold that you won't go below. So I don't see a problem with the calculator, but. But I don't think you necessarily need to price just based on the calculator, especially if you're trying to delegate pricing across a team. That can be a helpful kind of construct. So I don't have a problem with it. But the plugging the nuts and bolts in and spitting out a price, I do think you miss out on revenue by doing that. The other big thing for me, when you're at this sort of second tier and you've got some specialization, usually you've built that specialization through referral networks. But man, like, this is the time to start thinking about content creation. And I know, I know you're an accountant. It's the last thing in the world you want to do. You didn't go into accounting to be an influencer. And I understand that I might be biased here, but hear me out. Here's the thing about running firms is the way that we've run accounting firms has really not changed since the Internet came around. We have cloud apps and integrations and all that stuff. Well, some of us do. Tax preparers do not have cloud apps. Still living in the dark ages there. But the way that we find clients for most of us really like, hasn't changed. Like, we're still doing business with the people in our neighborhood and our little like in person network when that's not really the way that we work anymore. And like, I think I have obviously a different perspective on this because I am putting out tons of content and I've kind of seen the other side of that. But like, content is kind of like a lighthouse to attract the people that you want to attract. And the reality is if you go down your client list right now from A to Z, there's A lot of people on that client list for whom the problem that you solve is not that painful for them, and so they don't really value you as much as somebody else could. So this is kind of what I've gotten to. I was running a content creation series in my accountant community, like, doing this pricing video. I think here's what it boils down to, is at a certain stage in running a firm, your job transitions from doing the work to. To finding the people for whom the problem you solve is painful. Is 10x as painful as the people you have now? And those people are out there. If it's hard to find them, oftentimes it's because you don't have a specific enough problem that you're solving yet. But at a certain point, that becomes your job, because the people that you've bumped into in networking events and like, the mom and the. The cousin and all the people that you collected on your client list, think of how big the world is, which is bigger than we can fathom. Like, there are so many people with so many specific types of problems that, like, we genuinely cannot fathom it in our puny human brains. But, like, your job becomes in that vast kind of sea of people that you can't comprehend or understand. Your job is to put out the stuff that will find the people for whom that problem you solve. Is 10x as painful as the people that you have now? Because you know what? They're gonna pay you a lot of money to fix those problems. They're gonna be so appreciative of what you do, as opposed to Steve, who's like, I guess I gotta do a tax return every year if nobody's gonna do that. If all you're gonna do is sit there and be the one doing the tax returns and all that stuff. That's gonna be a really. What's the right way to say it? Imperfect client list not optimized. It's not gonna be like, you think of all the people that are out there. The client list that you stumble into is not going to be the clients who most value what you do fundamentally. And that's where a lot of our frustrations lie, where people won't pay you what you think you're worth and all that. And why some clients love you. And other clients are like, ah, this kind of sucks. It's because people take different amounts of pain. They feel different amounts of pain in the problems that you solve. And until you start putting out content and leaning into what the Internet enables, you're not going out and finding the people for whom what you do is really, really meaningful. And so I understand that. It's easy for me to say as someone who's like made content for years. And I think what's hard about that is it can definitely be a slow burn. It is counter to our personalities as accountants. But I think, and this is easy to say as we look at the changes we've made in our lives, in retrospect, it is much easier to stick to what you do because you've seen it work before than to trust that something that you haven't seen before is going to work. But what's really exciting about that thing that you haven't seen before, especially when it comes to creating content on the Internet, is you cannot even comprehend the serendipity that content creation enables. You have no idea who's reading that stuff, who's plugging into it, when that one amazing connection will come out of nowhere. That honestly opens up a whole new chapter of your life of possibility for you, whether it's firm running or anything else. But if you never put yourself out there and you never start producing that stuff, you haven't enabled that serendipity like that. Like those timelines are not open to you. And that's what I love most about, honestly about doing what I do, is the incredible people that I meet on the other side of it. It's the reason I'm doing this stupid podcast or whatever this thing is right now, because it's not as if, oh my gosh, people hadn't been able to hear me enough. It's because there's so many things that I'm excited about and rad conversations that I'm having every day where I'm learning and getting value from that. Like, I just think we need to open those conversations up more and see, see the positivity that's happening in these day to day exchanges between people that are finding success in running firms and stumbling into new things that they never could have imagined, like myself included, doing silly stuff like this. Anyways, that's the second tier firm, specialized firms, and they need to start looking at content to get to the third type of firm, which I love and makes the most of what the Internet enables. It is a high leverage firm. I've talked a lot before about transitioning from one to one services to one to many services. The Internet enables this in a really amazing way. Really obvious examples we see are like Internet think people selling digital products and making outrageous amounts of money. Well, like it or not, if you run a firm and you got a bunch of clients, you are an influencer and your own weird, extremely nerdy way. And all of your clients will still need one to one services. You're never going to one to many a tax return. Like, yes, people need stuff that's specific to them, but what is the right mix of things that you sell? 1 to 1 versus 1 to many. 99% of firms, every single thing that they do is fundamentally low leverage. They're solving a single problem for a single person, right? But when I make a YouTube video, like a whole bunch of people can get value from it. And I have to think there's a marriage of that, the high leverage that the Internet enables and the expertise that we have and deliver to our clients. I have to think there's something in the middle there that yields a more profitable accounting firm. And so one of the examples I've talked about in the past is running mastermind groups where with your clients, so very basic example, when you've got a bunch of clients that have the very, have very similar problems, you actually become a really unique hub for that type of person. Like if you're not online, if you're not engaging in communities and stuff like that, you probably, if you run a firm, it's probably a really isolating thing, right? Like I go home at the end of the day and, and my wife's like, honey, you're doing great. You're just, you're so good at this stuff and you're trying so hard. And I'm like, I'm like, thanks. But I'm really like, you know, I could be terrible at this. You honest, you have no idea. And it's like, I get what she's like trying to make me feel good. And it does make me feel good. But at the end of the day, the only people that fundamentally understand what you do are the other people that do what you do. And when you are the collector of a very specific type of person, there's actually a tremendous amount of value to be unlocked by you facilitating those connections. By being the hub, you don't even necessarily need to add value. The value that you're providing is being the hub. And if you have a specialized firm, you're actually a really unique connector of individuals. And there's some discomfort there with like, oh, what if one of their competitors or they don't want to, I don't know, they don't want people to know that they're one of your clients or something like that. I do think there's ways around that. Like the Mastermind group example I gave. Like, you know, you have people pay 500 bucks a month to come sit in a one hour session with you and 10 other people. Half of that is learning content that you give, but then half of that is people being able to pitch in anonymous questions. Like if you get 10 of your very similar clients into a space and they can hear each other's stupid questions in an anonymous way, that's really valuable. Like that's a, that's a really cool thing that you could enable for them that they couldn't get anywhere else. Because you could speak whatever your domain expertise is, tax expertise, CFO type stuff, you could speak that into them in a way that everyone can benefit from. And that's a really, really cool unlock. So the coolest version of a firm these days to me is that high leverage firm. Whether you're selling digital products or doing Mastermind stuff like that, your business actually starts to look more like a creator business. Like what I do. Like you've got a newsletter like you do partnerships. Maybe you're building adjacent businesses that support the type of people that your firm serves. But I love that type of business. I was super hot on that, really pushing on that the latter half of last year. I know Brandon hall talks a lot about this, trying to build his digital product business within his firm. Um, I would love to hear if that's something that, that you are working on and what your experience has been like selling products and dabbling in one to many services rather than just one to one. Because surely a hundred percent of our revenue coming from one to one services, surely that is not like the ultimate way to run a firm in the Internet age. Right? Right. Am I crazy there? Okay. Matt Metris, bless his heart, yesterday on Twitter, super cool. He ran a whole bunch of language models. And I promise this isn't just going to be Jason railing on and on about AI things. I absolutely am like tinfoil hat AI guy online now because I'm just far too excited by it. But it comes from a place of like understanding how overwhelming it all is and trying to keep accountants like out in front of it all. But Matt went out and ran a whole bunch of tax scenarios by I think it was GPT4 actually several different models. So he gave it a few different scenarios. I'm a single taxpayer in the US with a 7 year old child. I have one W2A $70,000 box 1, 8931 in box 2. So that'd be the federal Withholding box. What's my federal refund? So this is a fairly simple, like turbo tax level return, which not to just cruise right by that, to say that there's a free thing that we all just got access to that can now do TurboTax level returns. How many tens of millions of dollars, maybe hundreds, went into building TurboTax. And like, we all know if you're in the US and have ever filed a tax return, like just how ugly that is. But sure, Matt, maybe this free tool can totally do that. And it did some stuff right. It did some stuff wrong. I talked a bit yesterday about what kind of makes this hard is limiting the context of a language model to say, I want you to use this year's, I guess, right answer or this year's context. Because language models are trained on everything. In the case of stuff that's year specific, like tax rules, it's trained on all the years tax rules. So if we don't specify a year, I don't know, it might guess I might pull a hodgepodge of all these different years. But honestly. So none of these, at the end of the day, none of these answers were right. You may have seen in my AI video we did bounce some scenarios, some 2001, 20, 21 scenarios off of GPT4 that it got correct to the dollar. Very simple scenarios, but still terrifying for the reason of like, okay, what does that displace TurboTax and what does GPT5 look like? Like, it's pretty incredible, but surprisingly, like it did a reasonably good job. And what I think is really interesting is with language models progressing faster than any of our other software, there's many of the functions of our software that language models like ChatGPT will ultimately displace. If I can just pop out to this free little Chatbot and say, here's this CSV. I actually need this column and that column and that column laid out different ways. Can you do that for me real quick? And it gives you the file back. Like it can already do that. Like I've, I've, I've shared some of the ways that I've done like data manipulation with ChatGPT. It's really good at that. But if you got an app that does that for you, or you normally do that in Excel, you're just going to chuck that stuff into the language model. So if that stuff is getting better at a rate that's so much faster than any of our other software, when a new language model comes out, it doesn't come out with a Press release or a list of new features like your favorite software tool does, right? Hey, cool. Now you can do XYZ that you couldn't do before. That's not how language models work. And it's actually a real pain in the butt because you don't know what it can do and not do reliably until you do a whole bunch of testing. So if you've seen I've been talking about like a tax benchmarking framework, where to me the only way to know what can it do and not do reliably is to have a bunch of standardized tax pro provided like benchmarks of scenarios you can run through it. Does it get it right? Does it get it wrong? It's going to be hundreds and thousands of different models. There's going to be models optimized for tax or whatever that specific use case is for you. In my mind, when there's a new model, like there's not going to be a product like a feature comparison, it's going to do some things better, it's going to do some things worse. So who's it going to be up to to figure out what works and what doesn't work? Cause obviously the day that this can do a simple tax return, holy smokes, that changes a lot or whatever it is your domain expertise is if that new model comes out and boom, it can just do that thing that it couldn't do before. You wanna know right away, right? But how to figure out if it can do these things reliably? It's a really interesting paradigm shift. And so who are the people that are gonna do that? Obviously I'm doing a lot of that right now because I don't have anything better to do. But like it's interesting to think, like who will be, who will be the people that they call it capability? What is it? Capability overhang, I think is when a language model is capable of things that we haven't yet discovered yet. So like we're gonna be discovering for years things that the models we have today can do that we don't yet realize that they can do. But when you get a new model and you're in the business of X and it changes its ability to help you with that business, who is the body that will then figure out what works and what doesn't work and whether it's good enough and what situations it works for with situations it doesn't work for. And that ambiguity to me is not a reason to not be able to use it. Maybe it's a rationale for kind of adjacent tools that use language models, but in maybe a more structured way, like maybe there's a crossover of those two worlds where that product is ensuring the output of it in a way that maybe the language model itself couldn't. In fact, a really good example, just coming back to preparing tax returns. If you're gonna ask a language model how to prepare a tax return, it needs to be mandatory that you tell it what year you're talking about, for example, right? So like if you had a dedicated app for doing tax prep, there's gotta be a dropdown field there where you pick the year. And that has to be a required field. And then when the prompt goes off to the language model, it injects that year into the prompt that you gave it. So it's, you know, it adds a prefix. So for the 2021 tax year or something like that, and then it gives you the prompt that you gave it. So maybe that's the solution is like kind of these hybrid tools that are taking the best of both worlds. But I'm really fascinated by that because I don't know who the body is that is going to suss out the abilities of those models for people. Is that professional organizations, is that the aicpa? Is that something that doesn't exist yet? Is that something that software companies try to make proprietary? Ooh, we found this one thing that it does really, really well. We hope nobody else finds it. We're gonna stand up a landing page and with a little bit of prompt engineering, we'll let you pay to have access to this cool thing we found. I don't know, I'm not sure what that looks like, but it's fascinating to me with each of these new models as they come out, how they're all capable of different things and you just don't know until you test the heck out of them. So, like, I feel like that having a framework and a benchmark and a test bank of standardized things to run through those models, like that seems to me like the only answer. But it's really fascinating. It's just like I love the idea that you don't even know what's going to work when that new model comes out. Okay, last, a friend of mine said something the other day, and I don't want to necessarily out them specifically, but they raised a question that I think is a super relevant question as the mainstream folks like Microsoft and Google and all that are just absolutely going ham on AI applications and investing so much money into it. The big thing is like Microsoft Copilot, like, that's huge. If you're not familiar, Microsoft Copilot is now basically, they're building GPT into the Microsoft Suite. So in Excel you can have a chat assistant and talk with your spreadsheet and ask it to semantically, like via words, do certain things. Really cool applications around email, like summarize for me the emails that I got over the weekend and add tasks for them to my to do list or something like that. Doubly cool because Microsoft has a, has a Bing mobile app that lets you do all this stuff over voice. So like, imagine if your voice assistant could do all this stuff for you. But the really cool thing about Copilot is sees data across those different silos. From Word to Excel to email, it can see all that stuff. So if I've got a meeting in 10 minutes coming up with Brogan, I can say, help me prepare for this meeting that I've got in 10 minutes and it'll fetch any documents that I have in my file storage that mentioned him that have been updated in the last, you know, three months, maybe since the last meeting I had with him. It'll pull any recent emails, maybe. I got an email from him this morning that I haven't seen yet and that's something that I need to see now. Maybe he's got an upset tum tum and unless it shows me that email, I'm not going to know. I'm going to hop on the meeting like an idiot when instead I could go to the pool, you know. But I love how Copilot is now pulling data across all of those different things because there are so many contexts in what we do that that is helpful for from the client asking, Eric, can you send me a copy of my 2020 tax return? And I gotta go fire up that old donkey CCH engagement where that perm binder is. I gotta sync the binder, I gotta pull that out, I gotta save the PDF to my desktop. I gotta drag that sucker into the email client. As opposed to how things ought to be these days, where if I drop that email question into ChatGPT and I say write a, write, write a kind response to this. That's like one sentence. It's like, here you go, Steve. That's not hard. But when that system is connected to my file system, it can go out and fetch that tax return and suggest it and say, I'm pretty sure this is what I want. If all this looks good, just hit send. Here's your reply and here's the file Attachment. That's a really simple example. Tax returns. But oh mama, is that going to be such a huge time saver for the way that we work? The problem is now, and this is what my friend was getting at. He says, I use Carbon. I really want what I just saw from Microsoft. When is the right time to make the jump knowing that Microsoft's investing in this more than anyone else, versus waiting for Carbon to implement something meaningful here? And it's a really hard question to answer. It's why? Actually, a lot of my content, the stuff I talk about online now, is really pushing the companies in our space to move on this stuff. Because honestly, if the tech companies in the accounting space do not jump on this, there has never been more incentive to just leave the companies in our space than there will be in the next 12 months when Microsoft does stuff like Copilot. That is such a big step up and so powerful overnight. Man, I want all my files there where it can see it. I want all that stuff there. The fact that it can pull in meeting transcripts from teams from the last time I met with Brogan, the fact that I can search across all that stuff from one place, this may be what I'm most excited about with AI. Just imagine right now the idea of having a single search box that searched across everything across your meeting transcripts, across your organizational emails, across your documents, across your tasks. That is really easy with AI. The problem right now is all of that context for us lives in all these different places. And so if my files live in one place and my practice management system doesn't manage my files and I don't have meeting transcripts, and then I use Gmail or Outlook for email, none of those systems can see into any of the other systems. So it suddenly became very important that all those live in one place and that the place where those things live are leveraging AI in a meaningful way. Because that is now orphan data that could otherwise be really, really helpful for me. So as there's systems that don't do meaningful things with those data, the cost of those systems is now sky high. For example, if I'm using Zoom cloud recordings and my transcripts live there, or some other transcript app like Grain or Fathom or something like that, if it can't push those transcripts into wherever all my email and all my documents live and all that stuff, that's a big problem because meetings, for example, are a huge source of context that really we haven't leveraged yet. But like the importance of all that stuff being able to come back to one system all of a sudden got really, really important. And so if the companies in our space don't move on this, I'm afraid we're actually gonna see, like, the market size of software for accountants shrink really fast because other people are moving faster on AI and have more meaningful. Are leveraging it just in a more meaningful way. So when I talk about this stuff online, honestly, I. I don't wanna, like, Fear Monger and be like, oh, we gotta move on this stuff. And you don't wanna be like, the one that's like, pushing too fast and doing things in an irresponsible way. But people are already talking about this. Like, people are already looking at Microsoft Copilot and thinking, shoot, that looks better than the canopy I'm using right now, or that looks better than firm360 or whatever you're doing. Like, that's powerful. And so even if you don't have it in your product today, I think it's fair to ask the question of our vendors. What's the plan here? Because that looks absolutely fire, and I want it, and I don't want to change, and I would rather have it here where I am right now. But if you're not going to do something about it, I got to go because that looks really cool. So, like, I'm trying not to Fear Monger and, like, put all this pressure on the companies that we work with. But if I'm moving to a tool, like, I fundamentally would not move to a tool right now if they couldn't give me a straight answer on how they're adopting AI and virtually any application. I actually think one of the fun things we can do in this show. Oh, maybe fun isn't the right word. Uh, but I do think it's interesting to explore, like, of all the really cool AI stuff out there, like, how can AI make ignition better? How can AI make QuickBooks better? How can AI make canopy better? I think there's a ton of examples. I started doing AI consulting with companies in our space for the last three months or so. It's been amazing. It's been a ton of fun. The reality is there's a huge spectrum and the people that are leaning into it and a shocking amount of people where I'm like, where have you been the last six months? That are like, have a more simplistic understanding of it than, like, the people who just read the stuff I talk about online. So not everybody's gonna get there, which is too bad, because the cost of not doing so is really high when others are doing so and it's enabling really meaningful change. So I hope that doesn't become a negative thing for the companies in our space. So if you are looking at new software as you are reading about this AI stuff, the right answer is not to jump ship overnight and go say, okay, we're whole hog on Microsoft now. I think Microsoft did just enable a huge step, but to the next week say, okay, we're going to uproot everything that we do and put it in this new place. That's probably a little hasty, but you 100% should be asking your vendors and saying, I just saw this thing from Microsoft and here's why. It's really, really compelling to me. What's the plan? Can you give me something similar? Honestly, that's gonna have a much bigger impact than me railing about AI things online is you talking to your vendors and saying, man, I'm seeing this incredible, you know, Fireflies AI feature where I can chat with my meeting transcripts. That seems really powerful. Is there, is there a way that we can somehow incorporate that into the product or. I'm seeing more and more examples of awesome suggested replies and email clients, but today I run all my email through you. You don't have any sort of suggested replies. There a way that we can somehow build this into our product? I think you are absolutely entitled to be pushing our product companies to be looking into this stuff and is going to not only, like, get them on a trajectory that's more aligned with what you want to see, but ensure the relevance of the tools that are made just for accountants as all the other tools around us. Getting better and better, right? Crazy times, honestly, like, really exciting. Let me know what you think. If you're watching this on YouTube, drop a little comment below on your thoughts on what you'd love to see in the tools that you use. If you're just getting this one through your ear holes, shoot me a message on Twitter or LinkedIn or something like that. Any feedback you got for this silly little show we're doing would love to hear it. And thanks for popping by. We'll be back again tomorrow.