It’s been two months of running Podscan for real. So let’s take a look at what happened, what didn’t happen, and if I already spent all that funding money.
First, because it’s quite likely the most interesting topic, let’s talk money.
I never disclosed the full amount, and I likely won’t for a while, but let’s just say it’s a 6-figure number. That’s not a surprise, as it’s pretty usual for the Calm Company fund to invest this much into bootstrapped companies like mine.
So, did I spend 6 figures on the business already? Of course not. I spend around $10.000 so far — and that’s all GPU compute. The plan was to ramp up the computational power behind Podscan’s transcription and AI features much faster than I could have done with a purely self-funded business. And that’s exactly what I’m doing.
Experience this article as a podcast, a YouTube show, or as a newsletter:
Before the Calm Company fund, I had a small-ish main server for the app, a single database, around 5 GPU-included cloud servers for transcription, and a small server for my search engine.
Now, I am running a beefy 48-core server for my main app, a 48-core server for the search engine, 16 cloud GPU servers for transcription and inference, and a staging cloud GPU server to experiment with new features on.
I will ramp this up to 24 or so cloud servers over the next month. I am working on summarization and entity detection features that require some intense AI magic, which will require a few resources.
This will likely cost me around $15k/$20k a month in the end. Which means that I have a good year or more of runway just keeping expenses as they are. (And mind you, this is all done so I can get through the historical backlog of transcribing all past podcasts. Once I am through with that finite number, my daily compute needs will be 1/10th of what they are right now.)
Besides compute, I really don’t spend the money on anything else just yet. I’ll be hiring a few experts for design and UX work soon, but that’ll be project-based for now.
So a year of runway just from my war chest. But what’s happening on the revenue side of things?
I’ve just crossed the $1000 Monthly Recurring Revenue mark last week. On average, I get one or two subscriptions a day, mostly on my lowest-tier plan. Churn, at this point, is zero percent. Which bodes well for the long-term retention cohort graph. Month 1 to 2: 100% retention. Not too shabby.
Obviously, that’s not $10k yet, so Podscan isn’t profitable, but it’s technically halfway there. The post-catchup-Podscan will not cost more than $2k a month in pure infrastructure cost, so that’s what I’m aiming to get to as soon as I can.
And I am getting better at acquiring customers. I’ve been forcing myself to be as available as I can for customer research conversations, and I am now over 20 conversations into learning more about my customers. It’s incredibly powerful to chat with people who have a problem and know it. In fact, this might be a good opportunity to mention how much MicroConf in Atlanta last week has helped me with this. Two talks stand out: the one by Rob Walling, diving deeper into Eugene Schwartz’s Customer Awareness scale, and Stephen Steers’s talk about structuring sales calls. Rob taught me that it matters what awareness stage prospects are at, and Stephen taught me how to set an agenda, find alignment, and create situations of genuine connection over someone’s challenges.
The videos recordings for these talks should soon be available on the microconf.com website, so do check these out. Paying $50 for two day’s worth of lectures and workshops by your peers is a steal.
Anyhow: I’m really getting the hang of talking to my users. I ask them flat out what their Job to be Done is, how they accomplish it right now, why they consider Podscan as a solution to it, and what’s still standing in their way of getting their dream results.
Often, these conversations surface some underlying assumption that neither I nor they themselves have ever actively reflected on. I highly recommend recording these calls, or at the very least transcribing them as they happen. I use Otter.ai and Grain for post-processing these calls, and both tools offer “action items” that some AI grabs from the conversation. Highly recommended.
I’m glad I made my choice for who my Ideal Customer Profile is a few weeks ago. It has been helping me a lot in figuring out which feature requests I say yes and no to. I’m doubling down on creating a spectacular API experience for people to build their tools on, which in turn makes it extremely easy to build my own features in the alerting and searching sections of Podscan.
My customer service load is quite low. Every other day, I get a message, which I can usually quickly respond to. It’s not a complicated tool, and the REST API quite literally documents itself at this point.
Let me finish this update with a look at my expectations and the reality of my experiences.
I wanted to get to $1k MRR within a month from last time I checked in, and that happened. It happened just so, hitting the goal one day before the month was up. And in a way, that’s a bit too stressful. If I set my own deadlines and goals, I might set them in a way that doesn’t induce anxiety. So I’ll skip setting a new MRR goal for next month and instead make it about getting 5 really amazing customer testimonials. Those I can put on my landing page, print them out and hang on my wall, and use to fuel my motivation when I need it. Much better than an arbitrary number.
One thing that was a bit underwhelming was my experiment with cold outreach. I’ll have to talk about this in a future update, but let’s say I tried to find podcast booking agencies and send them personal emails. Didn’t go far. Or not far enough. I still struggle with attribution —something that Asia Orangio taught me at MicroConf, too— and that’s on my list of things to fix. I need to know more about where my customers come from. Segmentation and attribution are rising in their priority on my list.
I also expected the business to be a bit calmer than it is right now. I guess it’s unsurprising that spending a week in Atlanta with my SaaS founder nerd friends caused a few things to pile up. Even this week, I didn’t get to all the things I wanted to build because of calls, recordings, and podcast-related stuff. I will need to create more time to get into my Podscan flow state again.
But hey, things are going great, and I’m excited to tackle month #3. I’ll be working on new AI features, streamline the UX, update my landing page, create videos, start a knowledgebase, and so much more. Can’t wait to share all of that with you.
