The Email System I Almost Built Twice

By Junaid Ahmed

---

PodGlue has 50 people in the beta.

I wanted to send all of them a weekly email inviting them to a Zoom call. A sentence about what we'd cover. A link. A sign-off. Maybe ten minutes of writing every Thursday.

So I built a broadcast system.

Templates table. Send-recipients tracking. Cron-driven scheduling. Webhook ingestion for opens and clicks. Resend's Batch API behind it. By the time I was deploying the third edge function I had spent the better part of a day on what was supposed to be a one-evening job.

---

What I had blocked out

Kit, formerly ConvertKit, has been on my credit card for years.

I use it for the Hacks & Hobbies newsletter. I have lists, sequences, tags, automations. The unsubscribe handling, the rendering, the per-domain reputation, the deliverability, all of it is already running. I knew this. I have logged into Kit hundreds of times.

But while I was building the broadcast system, that knowledge was nowhere in my head.

The reason it was nowhere in my head is the same reason a lot of founders end up rebuilding things they already own: I wanted the experience to feel native to PodGlue. The admin would compose the email inside the PodGlue admin. And somewhere along the way I'd quietly decided that meant the sending engine had to be inside PodGlue too.

Those are two different things. I'd conflated them.

That conflation cost me a day.

---

The pivot

I stopped, looked at what I had, and reframed the problem.

The job was not to build a sending engine. The job was to give PodGlue's admin a single pane of glass for managing communications with the beta cohort. Kit is the sending engine. PodGlue is the control surface. Different problems, different tools, no rebuild required.

I shipped a Kit integration instead. PodGlue now syncs every waitlist entry and every active user into Kit with a tag, podglue_waitlist, podglue_user, podglue_trial, podglue_paid, podglue_canceled. Admins compose broadcasts in PodGlue's UI; the mail goes through Kit's infrastructure. Tags update automatically when someone moves through the lifecycle.

Half a day of work. None of it was a sending engine.

---

What I keep relearning

There's a habit founders develop where every problem looks buildable because every problem is, technically, buildable.

That doesn't mean building is the right answer.

I had Heather Angel on the podcast, she works with small business owners who are drowning in tools they don't need, building systems they can't maintain. Her advice, every single time: start with what you have, automate what repeats, delegate what an expert can do faster.

"Start with what you have" is the line I kept ignoring. I had Kit. I had been paying for it for years. I started building from scratch anyway.

The first question isn't can I build this? It's is something already doing this well, and am I already paying for it? If the answer is yes, the right move is integration, not reinvention.

I forget this every couple of weeks and re-learn it the same way. I start building something. I get a few commits in. Then I notice the something is already in my Stripe statement.

---

The part I want to keep

The original broadcast system isn't gone. It's still in PodGlue, quietly handling internal operational emails, the kind that need to stay close to the data and shouldn't go through a marketing tool. A welcome email after a workspace gets created. A notification to the host when a guest finishes their intake form. Those belong inside the product. Resend is the right tool for those.

Kit is the right tool for everything else. The weekly call invite. The trial conversion drip. The winback for cancellations. The build-in-public newsletter you might be reading this in right now.

Knowing which one is which is most of the job.

---

PodGlue is in beta with 50 hosts. Every Thursday they get an email from me, inviting them to the Zoom call. It goes out through Kit. It's tagged automatically by their lifecycle stage. The control surface is PodGlue. The sending engine is the one I was already paying for.

I keep building things I already own. I keep catching myself. The catch is the lesson.

Junaid Ahmed is the host of Hacks & Hobbies and the founder of PodGlue.