Hey there!
Missed you last week, sorry about that! I ended up getting sick after getting back from the company offsite (as did a fairly large percentage of the team, it seems…😬) and slept about 36 hours out of Monday and Tuesday.
So the newsletter didn’t go out.
But I had a bit of a revelation on TalkHomey during the offsite, based on a couple of things.
First of all, my cold outreach got me exactly nowhere. I tried to reach out to quite a few realtors and didn’t get a single response. But as I was talking about it on Twitter, I got some great feedback from Jenn:
That opened up a new realm of possibilities for me. I had been thinking about it in terms of living on the realtors’ websites, but that’s actually a much harder sell. It requires them to do the work of getting search traffic before I would ever see the users.
But what if I could flip that?
Instead of just living on the realtor’s site and being a resource there, why not grab as much traffic as possible from search? Then, I have a bunch of options as far as monetization goes: I can develop partnerships, still go the SaaS route, whatever I want with it!
So I shifted my approach, and developed an AI-native content flywheel as a way to generate traffic.
It starts with blogs. I decided to focus in on first-time homebuyers for a couple of reasons. First of all, they are the least likely to already have a realtor, so they’ll be looking for information online. They haven’t been through the process, so they won’t be in a position where they know what to search for. Instead of focusing on anything too wild, I can do a couple of things, and that brings us to the main topic of programmatic SEO.
What is Programmatic SEO?
In a nutshell, programmatic SEO is a way to generate a large number of similar pages but geared toward very specific search intents. The best example I have is Zapier, who does this really well.
If you search for things like “Integrate ConvertKit with Calendly”, Google helpfully serves up this as the top result:
If you look, that’s not a terribly informational site, but it’s set up to rank highly for very specific groups of keywords, and then this is done at scale for all of the apps Zapier integrates with. Even the graphics are only slightly customized, using the logos for the pair of apps you are connecting, but 90% of them are the same.
This is becoming more common with larger companies, because it’s becoming easier and easier to do. Add in large language models, and you can do some really interesting things.
So I decided to leverage Copy.ai for a bit of help here. I felt comfortable doing this because it’s not a rapidly changing topic. A lot of the things I was looking to generate are relatively “general knowledge” in nature.
So I’ve got a couple of different workflows I’m going to use to generate blog posts.
The first is based on the state the user searches for. This will allow me to target “first time home buyer in <state>” searches. Most of the information is the same, with just a little bit of variation. So I’ll generate those pages for all 50 states.
Next, I used GPT to give me 50 questions first-time homebuyers commonly have. That did remarkably well, and so I created a database in Notion with these questions to start.
Then I created another workflow in Copy.ai to generate blog posts that were SEO-optimized for a given question. And I hooked it up to N8N, along with the Notion database, and BAM! I was able to generate a blog post for each question in the database.
So what do I do with those blog posts? Great question!
I had the same one, so I did some investigation. And I realized Feather looked like the best bet.
Going to be a little honest here. I’ve kinda followed Feather for a bit, but hadn’t tried it out yet. It allows you to use Notion as a CMS for your blog. But it launched shortly after I launched Feather CRM, and obviously stuck around a little longer and had a bit more success, so there might have been a little bit of bitterness inside that kept me away from trying it.
But I bit the bullet, and I’m really glad I did, because it’s an incredible tool. It was super easy to set up, and works extremely well. So I was able to clone the sample DB to my Notion workspace, and then set up another automation that pushed completed blog posts to it, and I’ve suddenly got a powerful toolkit in my hands.
I used this to generate and publish 12 blog posts over the weekend, with more coming. I ended up tossing about 2 out of every 3 that were generated, because there were issues parsing, or the generation didn’t quite do a good job, but it’s really easy to run it multiple times, so not a big concern.
But that’s only part 1 of the AI-native content flywheel. What’s next?
The Rest of the Flywheel
Something else I ended up creating: a 10-day email course for first-time homebuyers that walks them through the overall home-buying process. It’s nothing earth-shattering, but it’s a nice easy intro to everything that happens.
Then I connected it to the Talk Homey blog via the CTA, setting up a webhook in N8N that gets triggered when a user enters their email on the blog, and that adds them to a ConvertKit form, which kicks off the sequence.
And here’s where it differs from the normal blog → lead magnet → email list flywheels.
I don’t want to publish a newsletter about homebuying every single week. And these users will naturally graduate once they buy their first house, right?
So I’m not going to. Instead, I’m going to let them opt-in to TalkHomey, the bot. That will give them the chance to have someone they can talk to about the process, ask questions from, etc. Basically, a realtor before they have a realtor they know and trust.
To get access though, I do ask them for a little more information. This adds a bit of friction that acts as a filtering mechanism. Because I’ll be paying for them to use the bot, but not monetizing them directly, this allows me to hopefully filter out the most serious users and provide the most value to them.
Then I can take the data I get from these conversations, such as what questions they are asking, and create blog posts based on those questions, feeding the content flywheel and making it move faster and faster, while I focus on the quality of the bot, which is the main product. Once I have a dataset, I’ll be able to approach realtors, service providers, etc. with compelling reasons for them to talk to me about potential partnerships.
What I Need
This is something I’ve been thinking about. It might be time to bring on a partner in this, even though I wasn’t initially planning on it. But if I had someone who had the connections in real estate and could sell it, that would allow me to focus on the stuff I know the best. Or the potential alternative is someone wanting to buy in to the business for some amount and hiring someone as an expert. If you know of anyone who fits any of those criteria and might be interested, please let me know! I’d love to talk to them.
In the meantime though, I’m going to be getting this flywheel spinning as fast as possible by the end of the month. Not a ton of time to gather data, but if I can start publishing quickly for the next week, that’s not going to hurt things.
Content Corner
A couple of things for you this week. The theme is crossover episodes.
First up, a couple of my favorite creators got together to chat, and one thing that stood out for me was the passion that both of them have for creating stuff that helps people. You should definitely check out this conversation between Arvid Kahl and Paul Millerd. (Also available on the Bootstrapped Founder podcast)
Next, Danny Miranda talked with David Senra. David is the creator of the Founders podcast, which I highly recommend listening to if you’re interested in entrepreneurship. (Again, also available in podcast form)
Finally, here are a couple of the videos I’ve released in the last two weeks. Will be releasing more deep dives into what I’m doing with everything, so let me know if there’s something you want me to focus on!
(This is my favorite thumbnail, by the way 😂)
See you next week!