I think the next generation of great marketers will be exceptional at designing and operating growth systems. It’s a job that will combine channel and audience expertise, with technology and with strong financial and analytical skills. I think that’s a cool job.

And yes tech and AI is absolutely part of it. It improves your speed and cost. Your ability to create derivatives. To personalize. It’s a major lever in the system for sure.

If you want to be a great marketer I highly recommend mastering one channel and then as the next step figuring out how to combine them together. Designing/architecting growth systems is the real job of marketing. It starts with sound financials. And then compounds.

On some level it’s like check out this bozo that just discovered integrated marketing. But there’s a difference between this and activating a campaign in a coherent, omnichannel way. It’s about how you develop channels that messages can ultimately leverage.

Every part of the system should do its main job but also spawn more parts of the system. The whole thing should be a self-reinforcing, infinitely compounding growth engine.

The major revelation for me is that yes you can be natively successful in any channel by just exceptional at that channel. But that is choosing to do everything on Hard Mode. That’s a bad system. The system should create leverage. And make the next thing easier.

Shocker I know, but visibility and engagement with that technical content increased like 3x as a result.

Results were ok but not great. Again, the system. We started using video on YT as top-of-funnel for written content. We clustered our explainer and how-to content together on the web (again see our ai agent experience/hub). Integrated tutorials into the newsletter.

We started building more developer content showing ppl how to build applications, workflows, etc using new AI tools. We managed it like our core SEO program. Developers would search rag + some tool/framework so we’d just do that.

We increased total subscribers by 5x in a year and our weekly new subscriber rate is almost 10x what it was before we made these changes.

We changed the system. Went to one main newsletter. Comes out twice a week and follows industry news, so high relevance. Deeply integrated with our core organic footprint so it’s promoted across over ten thousand pages. Can opt-in from every upper funnel form on the site.

On the newsletter front we used to have 4 topically oriented newsletters. Content, frequency and promotion for each was bespoke. And not aligned to our largest organic publishing programs. More to campaigns. Lots of subscriber overlap, mediocre engagement, low reach.

They sit on another, less technical IBM YT channel with lower reach/engagement. Don’t complement and can’t use our web footprint as well. Aren’t tracking the industry news cycle so weaker fit with our newsletter. Good shows, less good systems.

We have other shows targeting more c-level audiences that have not taken off in a similar way despite also being good shows (we think). The difference is our distribution infrastructure for those shows is weaker.

Smart Talks has a top level brand “motion” around it inclusive of advertising, in-person events, corporate social, etc. MoE is pure organic. Uses our tech YT channel (1m+ subs), our ai footprint on (google “ai agent” and you’ll see how this works) and.

Let’s start with podcasts. We have been building a podcast network and have two pods, Smart Talks and Mixture of Experts that have spent a lot of time on the Apple top 200 in tech this year. And we have other shows that have not.

Growth across these programs has of course been uneven, but we tried enough things where there are clear patterns around where we’ve had success and where we haven’t. I am going to cover three specific examples across developer content, our newsletter and podcasts.

Starting almost two years ago we began to invest outside our core competency for two reasons. First, we thought we could get greater yield from our traffic with stronger complementary programs. Second, ChatGPT arrived. And we needed to diversify programs and traffic sources.

I still believe deeply in that but today I look at it more as one the central pillars of an overall thesis vs being the entire game. This pov is colored by expanding outside traditional SEO and doing more across social, youtube, tech content, github, newsletters and podcasts.

As an example, let’s say I hire a writer for $150k fully burdened (salary, benefits, machine, travel, etc). On day 1, that person is in the red. They cost ~$500 a day and generate $0. But if they write one evergreen article a week for two years. That’s 100 pieces reaching ppl.