Great At The Work, Broke Anyway

Well, happy Monday! Hope you had an awesome weekend. Productive week ahead… and a lot of client as well as internal projects to press forward with this week on my end.

So, let’s get right to it, shall we?

Coming up:

• Really good at what you do, but revenue still low?
• How I automate publishing of my video blog (using AI)
• More chaos in the world of Matt Mullenweg and Automattic/WordPress? 🤣

Let’s do this…

Great at the Work, Broke Anyway

I’ll be honest about something… if you gave me a free afternoon, I’d spend it improving a business system. Not marketing. Not outreach. A system. I’m such a dork. 😜

Lately that’s gotten worse, not better, because now I’ve got AI to help me automate and tighten up all those internal processes. It’s like handing a kid who loves Legos an infinite bin of new pieces. I love this stuff. I could do it all day. And for a while there, I basically did.

Here’s the problem… loving the work isn’t the same as growing the business.

I talked to a client last week who put this better than I could’ve. He was telling me about someone he knows… same industry, similar business. This other guy isn’t nearly as good at the actual craft. My client is sharper, more skilled, does better work. He puts more detail into how he practices his craft. But the other guy has more customers and generates more revenue.

Isn’t that a real kicker? You can be really good at what you do… and still end up making less money than another person who isn’t as good at the craft.

But, it isn’t really all that complex, when you really look at it. Truth is… the work it takes to grow a business is not the same as how good you are at your craft. Even the most skilled technician (perhaps with the fanciest website) can have a less successful business than another who isn’t as good at the job. Being good at your craft doesn’t directly pay the bills.

I bring you back to the seven functions every business has. Production… the actual delivery of your service… is just one of the seven. Promotion is another. They’re not the same function, and being excellent at one does nothing for the other. Zero transfer.

But Production is fun. Or at least, for people like me, it is. You get to fix something. Make it faster, cleaner, smarter. There’s a finish line and a dopamine hit waiting at it. Promotion doesn’t work like that. Promotion is uncomfortable, ambiguous, and slow to show results… which makes it real easy to keep “getting to it later” while you go polish something nobody but you will ever notice.

That’s how you can have the most streamlined delivery systems and the nicest website… and still not be making the revenue you think you should be making.

And trust me, I know. I’m a systems guy. There’s always one more workflow to tighten, one more thing to automate. And it’s legitimate work! It’s just not the work that brings in the next client.

So here’s the mind-shift I’m reminded of… marketing isn’t something that happens when the internal stuff is finally “done.” It’s never done. That’s the trap. Promotion has to be a scheduled hat you put on, on purpose, on a recurring basis, whether or not the internal stuff feels finished.

In fact, that’s perhaps one of the biggest mistakes I see solopreneurs make. They hold off on promotion until they feel their website is “good enough” or “ready”. That’s the wrong approach. Promote. Promote hard. Turn on that marketing flow super early. Then, adapt to whatever happens.

If you’re a systems person like me, that’s going to feel unnatural. Do it anyway. The business with the better systems isn’t always the one making more money. The business that shows up to sell itself usually is.

---

WordPress News & Updates

Matt Mullenweg is CEO of Automattic again… already. The board put him on leave, and a few days later he’s back with “full support of the board,” while the CFO, the Chief Legal Officer, and a board member all quietly disappeared. Nobody’s said exactly what happened in there, but c’mon… the guy’s the majority stakeholder. My guess? He didn’t get “reinstated” so much as he replaced the board with people who’d vote him back in, then called it reconciliation. By his own count, that’s coup attempt number five he’s survived. Company drama aside, it’s a good reminder of who actually controls the platform your business runs on. Full story here.

WordPress.org can now block a sketchy plugin update before it ever reaches your site. During the 6-hour release cooldown every plugin already sits in, AI models plus Jetpack Scan now review the actual code changes, and a high-risk score kills the release automatically instead of waiting on a human to notice. It already caught a real backdoor back in July before it ever shipped. Quietly one of the better security moves the ecosystem’s made in a while. Details here.

Update these three plugins today if you’re running them. Super Forms, All-in-One WP Migration, and Gravity Forms all had critical, actively-exploited vulnerabilities patched in the last few weeks… unauthenticated file uploads and SQL injection, the kind of thing that ends with someone else’s code running on your server. My own fleet already caught and patched the Gravity Forms one across every client site last week. If you manage your own install, don’t wait on this one.

FluentCart products can show video now, not just photos. 1.6.4 adds a video tab to the product gallery… paste in a YouTube or Vimeo link and it plays right where your main image would normally sit. If you sell anything that needs to be seen working before someone buys it (a course, software, a template), this closes a real gap. What’s new here.

SEOPress 10.2 makes their AI assistant a lot less annoying to set up, folding it right into the install wizard instead of burying it three settings screens deep. Also worth knowing if you run a multi-location or directory-style site: Local Business schema can finally carry different opening hours per post, instead of one global set for the whole site. Full changelog here.

Perfmatters 2.6.8 added a proper Safe Mode. One toggle turns off every performance optimization sitewide without touching your security settings or code snippets… exactly what you want when a client swears “the site broke right after that speed plugin” and you need to prove it in ten seconds instead of thirty. Changelog here.

My Video Pipeline

As of today, I’ve released 11 videos for The Founder’s Memo, my new video blog right into the BMA YouTube channel. So far, it’s been fun and the analytics on my channel have definitely been impacted by the increase in activity.

Even with my first video, though, I talked about how I wanted to streamline this whole process. Make it easy. After all, the only way I’m going to keep up with this is to remove steps and simplify. I wanted to get as close as possible to “Record video and the rest just happens automatically”.

The videos themselves are recorded on my DJI Osmo Pocket 3. I could have used my iPhone, but I decided that would introduce more friction. The DJI is easy to pick up and record with without any setup. It stabilizes automatically. I don’t have to fight my phone case, or figure out how to tripod mount my phone. I just pick up the DJI and go.

But, here’s how I’ve mostly automated what comes after that…

I drop the MP4 file into one particular folder on my hard drive. That’s where Reed (my trusty Founder’s Memo AI agent) knows to expect it. I then go over to Reed and tell “him” that the file is there and when I want it on YouTube. Sometimes I decide to publish in real-time… other times I’ll pre-schedule it. I do all this by just telling Reed what to do. The rest of this is a semi-automated system that I built.

Reed then runs certain operations on that MP4 file automatically… right on my computer. It uses ffmpeg, which is a free command line video tool. While I would never want to use ffmpeg myself because I hate the command line, AI excels at this stuff. So, Reed starts running ffmpeg commands on my video. It injects the Founder’s Memo logo as an overlay over the video in the lower right. It also pulls a series of 9 screenshots from that video. Reed assembles those 9 screenshots into a numbered grid for me and presents them to me so I can select one. That chosen still shot becomes my YouTube thumbnail.

Simultaneously, Reed is extracting the audio track from the MP4 and then uploading it to HappyScribe (via the API) so I get a full video transcript. Within 2-3 minutes, Reed has the transcript and then fully knows what that video was all about.

Reed then presents to me the screenshot grid as well as 3 potential video titles for YouTube. He then waits for me to choose. I pick one of the 9 stills. Sometimes I pick one of the titles he comes up with, but sometimes I override and specify the exact title I want. If I want, I can even upload my own thumbnail instead of using one of the stills, but I usually go with one of the 9 stills Reed came up with.

Once I pick, Reed goes to work. The video still is uploaded to Canva. There I have a pre-built template for The Founder’s Memo. My still gets slipped into that as the background. My logo is on top and the entry number is incremented. That becomes the video thumbnail. Then, the video is uploaded to YouTube and all settings are set via the API.

Reed then tells me it is done.

This entire process was built with AI, of course. It isn’t as if I specified this whole thing and it just executes. I worked with AI to work out how to do it as we were doing it. I told Reed what I wanted and what the goal was… then we worked it from there. Coming up with ways to do it.

AI installed ffmpeg for me (I didn’t even know it existed, frankly). It helps me get my API connection to YouTube set up. We just worked through it. Whenever there was a constraint or a confusion or I didn’t like the output of something, we made adjustments.

It is an iterative process, not some magical skillset I came up with. Then, when I found it worked, I told Reed to document the entire process into my Vault so the whole thing was documented outside of Reed.

And that’s how it works. I record the video. Very minimal editing. Drop the MP4 file into the folder. Tell Reed. Wait a few minutes. Pick a thumb and a title. The rest just happens automatically.