Build vs Buy

AI-native teams are already building in tools like Claude, Cursor, and Codex. Now, compliance can happen there too.

Quick aside — prioritizing compliance unlocked serious enterprise revenue for us at beehiiv 💸

Anyway, my favorite security company, Vanta, is hosting a free webinar to showcase how startups can use their new MCP server, agentic workflows, and native plugins to achieve fast, credible compliance without leaving Claude.

You’ll learn how to:

I’ll be there — come say what’s up in the chat.

^{Even if you can’t attend, I’d register, and you’ll get a link to the recording afterwards.}

If I had to attribute my career to anything, I'd narrow it down to two things: being comfortable in situations where I'm in way over my head, and betting correctly on the "build vs. buy" debate.

The second one is a classic dilemma every founder and team faces, now more than ever with the advent of AI. It's exactly as it sounds — you either build a custom solution yourself or you buy it from a vendor.

For example, as your sales team scales, you could build your own CRM… or you could just buy a HubSpot account.

Building your own CRM takes a lot of time and effort, plus the hidden costs of maintenance that you'll own forever. But in theory it gives you exactly what you want, because you built it for your specific needs. And it'd be cheaper than an enterprise HubSpot plan (at least on the surface).

Personally, I think the whole SaaSpocalypse is wildly overblown. Vibe coding your own CRM sounds cool until your team has to build every integration, monitor the error logs, QA every feature, fix everything that breaks, and routinely update the software. Meanwhile HubSpot has nearly 9,000 employees who live and breathe CRM and handle all of that for you.

Oh, and this CRM isn't even a core part of your product or business… so why are you burning valuable time and bandwidth on it as a startup?

It's safe to say we have not built our own CRM at beehiiv. Nor do I ever think we will.

But I've made some big bets on the "build" side of that equation, and I genuinely wouldn't be where I am today without them. Here's the story.

Back in college (shoutout University of Maryland) I built a startup with a few friends called VentureStorm. The premise was to connect non-technical business types with student developers to foster and accelerate entrepreneurship on college campuses (i.e. business school 🤝 computer science school). This was 2014, before Claude.

The first problem we hit: we desperately needed the exact service we were trying to build. None of us had ever written software, but here we were trying to build a tech startup. So we took some online courses on Ruby on Rails and taught ourselves to code. Fast forward a few years and we had a software company with thousands of users. Unfortunately, also only thousands of dollars in revenue.

That was my first foray into software — I was a self-taught developer who never took a single computer science class. I’d describe myself as scrappy and willing to put in the hours to build things that just seemed to work, although I’m sure any experienced engineer would rip my code to pieces.

Learning how to code and building VentureStorm was a formative life experience. Like genuinely changed the trajectory of my life. But it ultimately didn’t scale, and we folded the company shortly after graduation.

Which left me living back home in my parents’ basement, networking and applying to jobs. One afternoon I caught up with my friend Austin Rief, who had launched Morning brew a year earlier. He offered me $3,000 to build some kind of social share feature for the newsletter.

I had $0.49 in my bank account. Despite having absolutely no idea how to build it, the only answer I could give when he asked if I could do it was "absolutely" (see above: comfortable being way over my head).

After almost quitting several times, I eventually figured it out and shipped it. Which led him to giving me more projects, and eventually turned into a full-time job as the second employee that fall.

Back in those days we had a file of HTML code saved to our computer with placeholders. Each evening Neal, our writer, would copy and paste content from his Google Doc into this HTML file, one line at a time. It took over an hour and often produced a few errant characters that required me to debug it… or start over and work well past midnight.

When it came time to hire our second writer, it was obvious no talented writer would voluntarily spend an hour of their night copy-pasting their work into an HTML template. We needed something that scaled.

I got familiar with the concept of a content management system (CMS) and started exploring the available options. There were plenty built for WordPress and other web platforms, but Morning Brew was different. Morning Brew was newsletter-first: the primary output had to be a beautifully styled email that rendered in every type of inbox, while simultaneously publishing the stories to our website.

Meanwhile Neal and the writers had an ever-growing list of requests to design and format the newsletter differently. We didn't want a standard black-and-white email. We wanted to push the limits of newsletter design.

Being the naive, overconfident 24-year-old I was, I suggested to Austin that I just build our own CMS. And to Austin's credit (or ignorance) he trusted me and let me go for it.

Morning Brew was my first real job. I was a self-taught software developer who had never taken a computer science class or built anything at scale. I was so ignorant and inexperienced that I didn't even know what a product requirements document (PRD) was.

I didn't spend weeks gathering requirements, mapping out the constituent parts, or anticipating blockers. I just opened my IDE and started coding.

Each morning I woke up anxious, uncertain whether today would be the day I hit the roadblock that I wouldn’t be able to surpass (see above: no planning). Then I’d have to tell Austin I wasted three months dicking around with nothing to show for it.

I definitely hit a fair share of what I thought were dead ends, but I kept pushing forward. A few months later we launched Oslo — our custom-built newsletter-first CMS, named after a coffee shop in Brooklyn (and yes, a European city).

Oslo was awesome. It had every benefit of custom-built software you'd expect, because it was built for our exact needs at Morning Brew. You could create a blurb for the top of the newsletter, add new stories in a single click, drag-and-drop the story blocks to reorder and format, and most importantly it would output a perfectly styled HTML newsletter for us to paste into Mailchimp.

Shortly after Oslo launched, I spotted a ton of bottlenecks on the sales team. We had a few newsletters beyond the daily business one, and each issue often carried 2-3 sponsors. In any given week the team was managing more than 30 sponsors.

But getting a single sponsor live required weeks of back-and-forth: collect assets, approve copy, confirm every link and image, and finalize all the details advertisers expect when they're spending $50K on a single ad unit. Multiply that by 30 per week.

Naturally, the team was managing all of it in a Google Sheet. And we were dropping the ball — routinely missing assets and realizing the night before that none of the copy had been approved.

Coming off the high of Oslo, I proposed building an ad management solution that would streamline the entire process end-to-end and insert the final approved ad unit directly into our CMS.

This time around with a whole lot more planning, and a companion. Ben was a software engineer just graduating from college, and the first person I ever hired. We tag-teamed this project together, and several months later Comet was born (I can’t recall the inspiration for that name).

If Oslo was awesome, Comet was out of this world (sorry). It transformed the entire sales and partnerships team and helped them scale that operation to $20M in revenue just a year later.

It was also just an incredible piece of software, custom-built for our exact processes. It collected assets from advertisers weeks in advance, had built-in reminders, let our copywriting team submit drafts and the advertiser's team suggest edits, then collected final confirmation and dropped the unit straight into Oslo.

There were probably other solutions on the market, but none of them would have accommodated every last detail our team wanted.

That's the trade-off on the "buy" side. Vendors have to build for a wide breadth of use cases across their entire customer base. The feature you need may never make it onto their roadmap. And you're forever at the mercy of their priorities and their price hikes.

Oslo and Comet were both huge risks, and I was admittedly way over my head leading and building both of them as some self-taught twenty-something. But we didn’t stop there — we also built the Morning Brew referral program that drove over 1M new subscribers, and custom-built our website from the ground up.

Morning Brew was made up of amazing creative talent and produced incredible content, powered by an entirely custom-built ecosystem. I’ll take some credit for the latter^.

If you squint, you can see the parallels to beehiiv. A fully integrated CMS and website, with ad solutions and growth tools. Sound familiar?

Toward the end of my time there, Ben recommended we hire his friend Jake, who he claimed was the smartest engineer he knew. Ben was already the smartest engineer I knew, so that was high praise.

If you don't already know how this story ends — Ben and Jake became my two cofounders at beehiiv.

If we had just bought off-the-shelf software at Morning Brew, I don't think it would have worked nearly as well. I don’t think Morning Brew would have scaled with the same velocity. And I sure as hell wouldn't be here writing this today.

beehiiv only exists, and is on pace to surpass $50M in revenue this year, because we took a few big bets to build instead of buy.

And we were right.

If you enjoyed this post or know someone who may find it useful, please share it with them and encourage them to subscribe: mail.bigdeskenergy.com/p/build-vs-buy

Shoot this desk energy straight into my veins. It’s just like working from a rainforest cafe.

Think you can generate a better office? Reply with your submissions 📨.

Some of my favorite content I found on the internet this week…

Turn on, tune in, drop out. Click on any of the tracks below to get in a groove — each selected from the full Big Desk Energy playlist.

Share this newsletter with your friends, or use it as a pickup line.

👉 Your current referral count: {{ rp_num_referrals }} 👈

Or share your personal link with others: {{rp_refer_url_no_params}}

Enjoyed this newsletter? Forward it to a friend and have them signup here.

Until next Tuesday 🕺🏽

📥 Want to advertise in Big Desk Energy? Learn More