Latest / Content, Briefly / Should you build your own productivity tool?

Transcript

Should you build your own productivity tool? · Content, Briefly

• 0:00
  This podcast is brought to you by our friends at Wistia. As content marketers, you know this feeling.
• 0:04
  You spend weeks creating one great, awesome piece of content, only to turn around and ask, "Okay, how can I get more out of this?" That's where Wistia comes in.
• 0:13
  Wistia is a video marketing platform that helps teams create, host, and measure videos, podcasts, and webinars all in one place. But what I love most is how easy they make it to repurpose content.
• 0:25
  Upload a video or webinar, and Wistia can automatically identify the best moments to turn into social clips, generate transcripts, and help you create content for every channel without starting from scratch.
• 0:36
  And that's because everything lives in one platform. You can see exactly what's working, from video engagement and podcast analytics, to webinar attendance and actions people are actually taking from your video.
• 0:46
  That's less time managing tools, more time creating content people actually wanna watch. Learn more at wistia.com/superpath. That's wistia.com/superpath. Welcome to Content Briefly.
• 0:59
  I'm Alex Hillary, joined today by my content friends Eric Jodoin and Ruth Favela. Today, we are talking about building your own internal tools using AI.
• 1:08
  It's something that is a lot of fun to do, and obviously using AI, um, has made building things possible. But sometimes you end up tweaking these tools over and over and over, and spending a lot of time maintaining them.
• 1:25
  So the question that we kind of circle around today is when is it worth it to build your own custom productivity tool for yourself, versus when is it better to just buy something simple and get back to work?
• 1:38
  Hope you enjoy the episode.
• 1:41
  So this conversation kind of spins out of, I think, what I've been seeing a lot lately in our AI show-and-tell, which we have a lot of people come on to showcase personal tools that they've built for themselves for productivity.
• 1:54
  And I think it's a very exciting time for personal productivity tools because the value that whether you're vibe coding, you know, some kind of user interface for yourself, or building a recurring prompt for some process you run,
• 2:11
  it's really exciting because you can kind of design to your own specs, right? I've constantly in my mind, probably for the past six months, battled, hey, should I build out my own personal to-do list, right?
• 2:24
  Like, I have these ways I wanna work.
• 2:27
  I constantly have to squash it down because, you know, something happens where there's some limitation in a to-do list, and I'm like, "Maybe I should just block out an afternoon and, and try to build my own thing."
• 2:35
  I'm curious if, if you all have encountered that kind of challenge. And if you...
• 2:41
  I mean, as I think about this, I'm trying to draw some line of, you know, some, some workflows, some things I do are worth building out with AI and designing my own custom thing.
• 2:51
  And I'm curious, one, if you encounter that same challenge where you're, "Ooh, I could go build this thing for myself," and if you've drawn a line for yourself of when it's actually worth it to dive into a project like that versus not.
• 3:01
  I have, definitely.
• 3:03
  Like, as someone who has, and Eric, I'm sure you've felt this too, worked on a very tiny team or have only been one person, um, there's always that line of, okay, I only have-- there's only so much of me to spread around all of these tasks.
• 3:19
  Where can I automate more, and what can I build for myself to make this go as fast as possible? Well, what if I automate my planning engine, or what if I automate my repurposing engine? Okay, what does that look like?
• 3:33
  That looks like 12 prompts that I have to build, or that looks [chuckles] like- Yeah... um, repeatable, uh, intake forms that I have to build, or that looks like this that I have to build, right?
• 3:45
  Like, what is worth building versus not, right? I always have to balance where do I invest the time that I then no longer have to repeatedly invest the time, and what is worth actually investing the time.
• 4:03
  And for me, that goes back-- that has to come back to, well, what goal am I reaching? Um, and this is actually how I have gone about how do I implement AI actually overall.
• 4:17
  Um, how do I know where and which tools to start with? I've always worked really closely with sales, so for me it's always been like, where am I making the biggest pipeline impact? [laughs] Yeah.
• 4:31
  And working, and working really closely with the sales team to decide, okay, well, we need a, a better workflow intake process, and then creating that because it's gonna save me two hours of time every week and over a quarter that is actually worth it with them.
• 4:50
  Yeah. So I come... Like, I, I mean, if you've listened to me on this podcast, I'm, I'm very automation heavy in general as a marketer. It's just, like, my way of surviving as a team of one.
• 5:03
  I wouldn't say that I've, like, rushed to build my own tools with AI, partially because of, like, how efficiency-minded I am with my own time.
• 5:13
  And part of it is just thinking of your own time, like, as, like, an hourly wage, I think. And it's like, okay, let's say, uh, I don't know, I'll make up a number.
• 5:21
  Let's say I'm worth $250 an hour to my company or, like, some, some number like that.
• 5:27
  If I can buy a $200 a month tool that does 90% of the job, like, that, that's no question to me that I'm not gonna build that myself, right? Like, it, it... I- because it's...
• 5:39
  I know it's gonna take more than one hour of tinkering a month. I just know it.
• 5:42
  [laughs] Uh, [laughs] like, if it's, if it's something I'm opinionated enough that I was going to build myself, I'm, I'm just gonna tinker with it, and I have to, like, resist that urge. It, it's, yeah, the tinkering.
• 5:52
  It's like that's the [laughs] that's the challenge- [laughs]... is because I don't think there's anybody that can [laughs] like resist going in and constantly tinkering on this process.
• 6:01
  If the output you're trying to achieve is, uh, productivity gains, but then you keep on tinkering with it every month, then you haven't-[laughs] Cheap value Exactly.
• 6:10
  Well, I think you're then weighing, do I tinker with this process that I am continuing to iterate and make better, or I- do I maintain a tool that somebody still has to maintain monthly?
• 6:23
  But at that point, you didn't have to build that $250 tool that you bought at that point, right? You have to now pay for onboarding. [laughs] Yeah.
• 6:33
  Okay, there's another way I think about this, in that, um, uh, I, I said this back in the day about automation too.
• 6:39
  Like, th- rule one should be you should be absolutely confident that you are doing some sort of repetitive work or absolutely need this thing, and, like, you're feeling the pain of doing it yourself for long enough that you know that replacing yourself with a tool is going to be worth it to you.
• 6:57
  And so, like, for example, there, there's lots of things where I was like, "Man, it'd be really nice to have this," and then I went and I spent two- like, a week building it.
• 7:05
  And then once I had it, I was like, "Oh, you know what? Like, that wasn't nice to have." I, I never actually... So here's an example.
• 7:10
  I made, like, a dashboard that let me link, like, freelancers to all the blog posts they wrote, um,
• 7:19
  some of the analytics of those blog posts so I could see, like, oh, of all the freelancers I've worked with at Doc, here's what each one wrote, here's the results they had.
• 7:27
  And I was like, this would be a cool thing that I could share with them and they'll be, like, happy about their... I don't know. Like, I had some...
• 7:33
  It, it was a nice to have, absolutely, and I spent all this time making it, and I literally never looked at it again. And so, it, it wa- it wasn't, like, really replacing a strong pain in my day-to-day workflow.
• 7:43
  I just had a cool idea, and I think it's... Like, cool ideas should absolutely [laughs] go in, like, the trash bin for the most part. It's...
• 7:51
  Versus there are other things where it was like, I'm really s- like, um, I built a competitor database with cloud code that I've talked about on this before, where, like, I was finding that was a blocker in every single thing I was working with.
• 8:05
  It was, like, data inputs about our competitors and, like, what feature set every single one of them had.
• 8:10
  That, that has, like, been an enormous unlock for me that still took, like, I don't know, 20 hours of work probably in total and ongoing maintenance, but it, it's saving me an hour of work on every single task I'm working on going forward.
• 8:23
  So, like, there's this efficiency at scale gain that I'm getting there, and, like, it's not always easy to do that math problem in your head before you build the thing, but you should at least try before you just go down the rabbit hole of, "Ooh, this would be a cool idea."
• 8:38
  Yeah, and something I think about there too is just the, the, like, lifetime value of that thing. Mm-hmm. And how set are you in that specific process.
• 8:47
  Like, something I think about a lot is, for example, with these, these podcasts, we'll get really meta here. Um, we have a new process in place, and we have these podcast planning docs that we've been working on.
• 9:00
  But we've iterated.
• 9:01
  It wouldn't make sense for me to build something right now because we're still in the iteration process, and maybe we keep, you know, f- some format for three or four weeks, but then we say, "Actually, it'd be better like this."
• 9:11
  You know, we're still in the testing phase, and I think it's really important, as I think about the math that you're talking about there, is to say, you know, is this process stable enough that maybe there's a few tweaks here and there and, you know, depending on what you're using, you can use chat, like, you know, AI chats to make quick tweaks to it.
• 9:30
  But is the core of this going to stay the same for six plus months?
• 9:34
  Because if it's not, then it's just like, why, why am I doing this to, you know, drop 20 hours into something that's only gonna last for, for two months?
• 9:42
  Which is a kinda hard thing to do because you don't actually know how long something's gonna last. But
• 9:48
  I, in my mind, I kind of, I like that six-month mark of, is this thing gonna be relatively stable for six months, um, before I spend time really optimizing it. I think about the process for planning a podcast, right?
• 10:02
  Mm-hmm. And I think about, like, we were just talking about, or Eric and I were just talking about, like, you know, thinking about topics for a podcast can be a pain. And
• 10:16
  w- one thing that I will always run into is, okay, what am I gonna talk about next time? What do- Mm-hmm... what do the next four podcasts, or what does, what does this webinar series look like?
• 10:26
  Well, the first thing that I would do is automate how I'm gonna plan that. Or I would build an engine to make that part go so much faster so that at least that piece is solved for.
• 10:39
  Uh, or think along the lines of something like that so that I'm automating one piece of the puzzle, not necessarily thinking about the entire engine at that six-month mark.
• 10:52
  But that's just, you know, my thinking in increments. Yeah, and I think another part of that is, uh, let's use the pod- a podcast as an example of...
• 11:02
  Uh, and going back to what Alex said about not wanting to, like, over-process something that hasn't found its rhythm yet. Nikki Zuccaro was presenting at our AI show and tell, and she said, like, "I'm a junior marketer.
• 11:13
  I'm trying to automate things for myself, but I also wanna learn the, like, I wanna learn the job.
• 11:19
  I don't wanna automate myself out of no- like, I don't wanna defer to Claude for every part of my role, 'cause I, I still have to learn how to be a marketer."
• 11:25
  And I think, like, for planning for a podcast- [clears throat]... for example, it's like, you know, y- you might have to do 500 episodes before you know how to properly automate, like, part of that, right? So I think
• 11:39
  there's, there's, like, learning, like, where, where is it valuable to, like, spend time in the mud and make mistakes and, like, fumble around and versus, like, learning what to automate.
• 11:49
  I think you kinda have to do the whole process end to end a few times, and then you can find the piece that you're like, "Okay, I'm adding no value here. Like, let's automate myself out of that."
• 11:58
  And then once that's going, then you say, "Okay, piece two," and then, and then the system starts buzzing.
• 12:02
  And I think that's a good rule of thumb in general that you've flagged, Ruth, here is, like, try to automate the smallest possible thing first, or Like AI, the smallest piece, rather than biting off these huge chunks or like, yeah, trying to build an end-to-end personal operating system.
• 12:17
  I th- I think that's too hard.
• 12:21
  One place where I have found success, and I think I've taken a more kind of iterative approach like that, was-- So my first Claude code project was our one-on-one program, and taking basically everybody opts in, um, all this data is in Airtable.
• 12:37
  Um, the pain point I was trying to solve was to be able to match people and know had they matched before in the last six months, um, how good of a match are they for each other based on their bios.
• 12:54
  Um, and I wanted to still be very involved in that process, but I couldn't do it very easily with just a spreadsheet, like it was a, a hard interface to work with. Um,
• 13:03
  and so I built out a Claude code, um, program that helps me, like basically everybody turns into little cards that I drag on top of each other, and it rejects the cards if they've already matched with each other, and then it gives me a score based on like little algorithm I put together of, you know, hey, these people have similar types of roles at similar types of companies.
• 13:23
  These are probably some overlapping challenges they have. That was, I think, a very good use of automation and very worth it for me because it was like a very big unlock.
• 13:33
  What I've slowly built now, but I did not build from the beginning, was like, okay, now convert those into email drafts and automatically like set those up in Gmail. But I've kind of been adding on one thing at a time.
• 13:46
  And unlike that content calendar approach, it wasn't just like me trying to throw the whole product together in the first go.
• 13:53
  It was pinpointing that specific thing that needed the unlock and then kind of building on things from, from there that could continue to add value over time. I
• 14:06
  might have talked about this in the blog post that I wrote about it, but the, the first-- Before I built it out in Claude Code, I, I was just doing, like I had this big prompt built into to Claude and Claude- Mm-hmm...
• 14:17
  automatically matched people for me. Um, and it wa-- still wasn't getting the right results. Um, so it was optimizing for, you know, my little algorithm, and that's great.
• 14:25
  But there's so many like weird finicky things that are like, I know this person would really vibe with this person, or like, I know this person's had a really bad experience the past two months.
• 14:34
  Let me make sure they, like we get a ten out of ten match for them versus like somebody else that's like generally had a good experience.
• 14:40
  I'm happy with them getting a seven out of ten because that's probably good enough. But, um, there's all these little things that require judgment and, um, I realized that I wanted to be involved in the process.
• 14:53
  It takes me about forty-five minutes to match people, um, and that's so worth it in my mind. So I didn't wanna fully automate it either with like just a matching algorithm or an agent to run the full process for me.
• 15:05
  I think you flag a good example there too of there's times where AI makes sense in the loop, and there's other times where something should just be built as an automation, like something- Yeah...
• 15:15
  with something like Zapier or, you know, Airtable, some trigger from a project management tool, and I think it's easier to just sorta like vibe chat with Claude and make some of these things happen.
• 15:27
  But I think they're less dependable in the long run. And we talked- Mm-hmm... about like the cost of maintenance and things like that.
• 15:32
  Like, yeah, there is some debt you accumulate by having all these Zaps that like you don't remember that exist. Superpath was a really good example of that. [laughs] Jimmy was like the king of Zap.
• 15:41
  There was like ten thousand Zaps running every day. Um- Mm-hmm...
• 15:44
  but I, I do think if you're, if you're sort of new in this world that like automation is sometimes an easier place to start than like vibe coding a whole thing yourself. Mm-hmm.
• 15:55
  Um, and it, and it also lets you-- It sort-- It helps you understand the bones a little better before you start trying to build a massive system. I think though then you are weighing two options though, right? Right.
• 16:05
  Like how do you maintain a system that has all of those Zaps versus if you build something and you have to constantly maintain what you build and update it so that it's not breaking.
• 16:23
  Because as someone who has built landing pages, mini apps, decks in things like Lovable and Replit, they, they break constantly. [laughs] So you always have to come back and you're like, "Okay, this broke.
• 16:41
  Why did it break? Fix it." And then of course the code is gonna break because whatever was fixed then breaks from trying to fix it. How do you decide what to maintain? Absolutely. There's, there's definitely like a,
• 16:59
  I think product debt is the best way of thinking.
• 17:01
  It's the same reason why a, a fresh startup can build product a lot faster than a twenty-year-old s-startup is there's like so much technical debt built into the software, and the same thing can happen in your own workflow.
• 17:13
  I actually have-- I have some automation set up where if I like reschedule something, it cause-causes me all kinds of headaches. [laughs] You know, I have like, like a podcast workflow.
• 17:23
  I'm like, "Okay, it's on this date," and then like all of these hundreds of things flow out of it. And then if, if the guest is suddenly like, "Sorry, I have to wait one more week for approval," it's like, "Ah."
• 17:33
  And, and there's- [laughs]... like forty-five minutes of dragging all these things around and like, so you have to weigh that cost too.
• 17:40
  Um, but it's, yeah, there's, there's definitely costs of associated with over-engineering things. Can I take this conversation a different angle?
• 17:50
  So one thing that was really interesting to me that came up again in this AI show and tell is Nikki, who I referenced before, she was sorta like, "This thing that I built, it was like a, a dashboard for myself,
• 18:02
  and it's basically just Ahrefs data, but I don't like working with a spreadsheet of a..." Like, you know, there's a thousand reports, there's all these buttons.
• 18:12
  Like I just, I just want something that tells me the top 10 things to work on today.
• 18:16
  Or Connor, who had presented as well, showed like a, a focus app he had built for himself, and I think an interesting aspect of this is, like, your own personal working style, and to, to some extent, like, people who are neurodivergent or, you know, v-various, like, various modes of productivity that we all have.
• 18:36
  I think it's... This is, like, another... Like, we, we talked about, like, ROI and efficiency and things like that, and I think that's one part of the equation. But I think another is just, like, how you
• 18:46
  work and, and understanding yourself and building tools.
• 18:50
  Like, you can inter- Like in the example I gave from Nikki, you're, like, interfacing a tool on top of another tool just to be like, "I, I, I can make this more, like, flavored to how I wanna s-spend my day."
• 19:00
  And I, I think that's definitely in the realm of, like, uh, you could have spend a lot of time, like, wasted a lot of time building something really fun. But I, I do think there's some value there.
• 19:10
  Um, like, ha-have you two done anything like this, or, like, what, what are your thoughts in this area of your own personal flavor of productivity? I'm gonna be completely honest.
• 19:19
  I feel like I have always been in roles where I've kind of been the square that's been forced to fit in the circle hole, and I've just kind of made it work [laughs] um, by
• 19:35
  implementing processes that I make work for everyone else, honestly. Um, and that's kind of...
• 19:44
  Like, with AI, I've made it work by, like, being the person that's like, "Oh, but we can do it this way, we can do it this way," and then I get to set the s- and then I get to set the sa- the standard.
• 19:56
  So if someone is telling me, "Okay, but we need to do it this way, and we need to make it fit in this system," then I say, "Okay, cool. But can we do it through this process?" Mm-hmm.
• 20:08
  And if we do it through this process, and I align it with your system, then we make everybody do it, and we both win. I mean, I've been on both ends of this where, like, I've implemented a system for a team, and it...
• 20:23
  Like, I've built it for me, and then I can see immediately that there's, like, a third of the team members that don't speak spreadsheet, right? It's like, "Oh, I've made this beautiful spreadsheet. Here you go."
• 20:34
  Or, like, "This beautiful Kanban board. This will be our process now." [laughs] It's so obvious. [laughs] It's so clear. [laughs] It's so beautifully color-coded.
• 20:42
  And then there's a third of the people on the team who are like, "I don't work this way. Like, I need... Just put, put five ta- like, email me every day with the thing I need to do." You know?
• 20:50
  Like, there's, there's different ways of working, and I think it's easy when you're the person who's deciding the tool stack.
• 20:56
  What I, what I think is interesting these days, and I've been on the flip side too where I'm in an organization, they're like, "These are the tools we're using. Get used to it. Uh, we paid $100,000 for this.
• 21:08
  You, you don't have a choice." But now you could, like, implement your own sort of personal working tools. It's, it's just, like, a different relationship with tooling than we used to have. Mm-hmm. Does...
• 21:18
  Alex, does this resonate with you at all? Yeah. I mean, I think this is... I'm going back to my, um, Todoist example.
• 21:26
  But I re- So pre-AI days, I remember, like, being on a team that had Asana as, like, "Hey, we're doing this team-wide," or marketing team's doing this. I was like, I got into Asana and I'm like, "This is...
• 21:38
  There's too much going on. I have pretty simple workflows. I'm, like, overwhelmed by this whole interface."
• 21:43
  And I think I, um, ended up buying a $4 a month third-party tool to send, that went back and forth between Asana and my own personal, uh [laughs] Todoist account so that it operated the, the way I want to.
• 21:58
  But, um, obviously in the AI world, you can do, um, a lot more with that. And I, I don't know. This is just where I kind of get back to, like, I am scared to just go over-optimize everything right now for myself.
• 22:16
  I mean, I also find, and, and this might be a more of a me thing than, than anything else, but I change what my preferences are over time. Like, within a year- Mm-hmm...
• 22:24
  you know, a lot of what I like doing, I'm like, "Hey, you know, going into this new season, like, these are our focuses right now. I have to kind of shift the way I work." And so I've found that it's not...
• 22:38
  Over-designing for what I need in the moment isn't worth it when I know that the way I wanna operate three months from now might actually be different. Um,
• 22:49
  so I've just struggled to find the, like, will to, like, fully invest in that. The multiplayer requirements of the tool is a big factor in this too.
• 22:59
  Like, if, if you're confident that this productivity tool is only gonna be, be used by you, like, that's an easier decision.
• 23:06
  But once it has to be, if you're like, "I'm gonna have to build a thing, and then if this works, I will have to roll it out across the team-" Mm-hmm... I think that's a much bigger investment.
• 23:15
  And then you get into, like, software development, product manager world, where, like, suddenly everyone- Mm-hmm... has a feature request, and you're, like, balancing requests from the team. Mm-hmm.
• 23:25
  And suddenly you need, like, logins, and it's, it's much harder- Exactly... to roll out a team, a tool across a team, in which case, like, buying is almost always the answer.
• 23:34
  And so I, I think that- It's why every team, [laughs] it's, like, why every team at an org has a different project management software. [laughs] Totally. Yeah. Like your typical, um, organization problem.
• 23:48
  Last question, I think. After this conversation, what is kind of the line for somebody that is thinking about building a process for themselves? What would you say if you were recommending, "Hey, you know, should you
• 24:01
  touch this and, and start on a 10-hour project to build out this productivity tool, or should you, um, maybe think about using your time in a different way?" Step one: What are your motivations?
• 24:14
  Are you doing this to procrastinate on doing actual work?
• 24:18
  Uh, because I think that's a really common mode of if, if I build myself a process rather than just getting into the work- I will have somehow been, um, more efficient. And like I, I'm that way.
• 24:28
  Like, I'd rather build the process for the thing than do the thing. Uh, and sometimes you just need to do it. So I would suggest with try doing it yourself first. Second, [laughs] second is can you just buy it?
• 24:41
  Is your time worth the, a subscription to a tool? Uh, if you work at a SaaS company, you probably have more budget for tools than you think if you can make a good argument for it. Number three is, like,
• 24:54
  what are, what's gonna be the long-term cost of maintaining this thing? And then is this absolutely necess- Is this unblocking something for you right now rather than a nice to have, right?
• 25:06
  Like, is, is it impeding other work where you'll instantly get the time savings back? That's my branching logic for, [laughs] for this question. Ruth, did I miss anything? No.
• 25:19
  I think, I think that's a, that's a good way to put it. I am a very goal-oriented person, and I think I would also include, like, goals. Like, where does it make the biggest impact? Because if you have
• 25:35
  a KPI you have to meet super quickly, then that is also needed, you also need to take that into consideration, right?
• 25:43
  Like, okay, great, where is it gonna save you the most time, and where is it gonna make the biggest impact, and where is it going to create the most efficiency with... and, and also unblock you.
• 25:56
  I've, I've had to look at all four of those. [laughs] Um- No, I'd like a really boring technical asterisk is, like, quite often the answer is not to build your own tool but just connect two other tools together. Mm-hmm.
• 26:08
  And, like, everything has Cloud MCP now or, or- Yes... MCP for opening eyes. Normally, it's like, uh, could I just make Airtable talk to Riverside? Like, that's almost always- Yeah...
• 26:19
  better than, like, building your own UX. Like- Mm-hmm... if you're building, there's so many steps to take before you start building buttons.
• 26:27
  Uh- I feel like what building means today is just my data lives in all of these places. How do I connect it so that it is in one place? Or just between the two things where it needs to be.
• 26:41
  Like, I, I think there's a, yeah, there's the risk of... I know we're getting back in the weeds here. There's the risk of like- [laughs] Exactly...
• 26:46
  like Alex said, is trying to build your own, like, complete operating system versus, like, can I just- Yeah... make this tool talk to that tool rather than inventing a third tool that's trying to do both, I think.
• 26:56
  Exactly. It's like a good... It's a big trap. Yeah. Like, when should you not do things? When you have more than two things to connect. [laughs] Find somebody- It can get really dumb...
• 27:07
  find, find somebody else to connect them for you. [laughs] Awesome. Well, Eric, Ruth, thank you for joining me today, and thank you everybody for listening. We'll catch you next week