Day 16: Crafting a High-Converting Landing Page
Slobodan and CofounderGPT made some great progress on Day 15 with the development of our prototype. There’s still work to be done, but we expect to start being able to test out Knowlo in Vacation Tracker and a few other companies by the end of June, maybe even earlier.
In the meantime, I’m trying to work with CofounderGPT to build an audience for Knowlo so that we have some potential customers for the product which will come from outside of our network.
Today’s update is all about an experiment we want to run to try to increase the visibility for Knowlo. we’re trying to capture more email addresses in order to gather some signals on what our customers want out of a product like Knowlo.
We already have 1 pilot customer besides Vacation Tracker who committed to test out Knowlo in their product, but they prefer to remain anonymous for now. Today I’m happy to announce that we have our third pilot customer: LeadDelta. The CEO and founder of LeadDelta is a good friend of mine and he agreed to test out the prototype of Knowlo with some of his customers. I’m super grateful to Vedran for agreeing to do this. I am also really grateful that I have a good network around me that I can ask to test out new products that we come up with.
I recognize that not everyone has that luxury and that these types of networks take years to build. So I encourage everyone to keep building their networks and as it happens, Lead Delta is actually a tool that helps you to do that more effectively with LinkedIn.
Landing page frameworks
It’s part of startup lore that one good way to validate interest in an idea is to create a landing page which accurately describes the product and then gets people to sign up to get access to the beta version of the product. The idea is you spend a little bit of money on Google Ads and Facebook ads to get a few hundred people to your site, and if a decent part of that traffic (say 5%-ish) leaves their email address, you might be onto something.
So Let’s try to see what CofounderGPT can come up with:
Today we are going to work on our homepage to make it into a landing page which converts better. Right now we just have a hero section with some text, an email capture form, a section with the team and a section with the latest news.
We would like to convert the current homepage into a landing page, with the goal being to explain the product and to convert as many people as possible by getting them to leave us their email address.
Can you suggest a few landing page frameworks we could use to accomplish this?
I already have a framework in mind that Slobodan suggested but I want to see what CofounderGPT is going to recommend:
This is a decent list of ideas and I’m sure some of these might work for us. In the first year of building Vacation Tracker, Slobodan landed on a book called Building a StoryBrand: Clarify Your Message So Customers Will Listen. We both read it, and it made a lot of sense to both of us. So I’d like to try using that approach for our landing page.
I asked CofounderGPT if It’s familiar with the StoryBrand framework and it gave a nice summary of it as the answer:
Great, since It’s familiar with StoryBrand, we are going to try building our landing page with this framework in mind and CofounderGPT will hopefully do most of the heavy lifting.
Here is the prompt I used to get an outline of the landing page with the StoryBrand twist:
Let’s build a landing page with the StoryBrand framework. But let’s try to keep it simple. We don’t actually have a product yet and we will only be launching the beta version of Knowlo in a few weeks.
As CofounderGPT, you know all about the Knowlo product and how it works. Please act as a marketing expert who specializes in conversions. Provide.
Please provide a detailed outline of a simple landing page using the StoryBrand framework. It should include:
– a super catchy hero section of the homepage with an email sign up field
– descriptions of all elements and sections of the page
– include all the text that should go in each section
– include descriptions of any screenshots that you think would be relevant
The first result CofounderGPT produced was average as usual. And some of the text it wrote was pretty cheesy. But I have been testing out a new approach with getting better results. I was avoiding it before because it uses up several of the 25 messages you have with GPT-4 every 3 hours. But it seems to produce good results.
I saw this tweet a few days ago:
And I’ve been trying this approach of asking it to do it better. Except I also add a little bit of extra feedback like “This is great, but I know you can do much better than this by including more information about X” or “This could be so much better if you focused on how it will convert”.
I did this 5 times with CofounderGPT and got several variations of its idea of a StoryBrand framework landing page. And this is also something new I’ve started doing: I get it to do several “do it better” iterations and then I pick out the best parts from each one. The only problem with doing it this way is that you quickly burn through your 25 messages every 3 hours.
Here is the last iteration:
We’re going to use a part of this, parts of some of the other versions CofounderGPT wrote and I’ll add the missing pieces and do the final editorial of the page.
Wireframing the landing page
I’m a visual person and I need to see how everything that CofounderGPT is suggesting will look like on a web page. So I went back to our trusty Excalidraw to do a quick mockup of the idea:
I need to validate with Slobodan if he agrees with this layout, but this is approximately how the first version of the new Homepage is going to look like.
About Us page
To give some more credibility to Knowlo when we bring new traffic to it, we’re also going to add an About Us page. If someone is going to leave us their email address, they may want to know a little bit more about the people behind the project.
Here’s the quick mockup I made for the About Us page:
I’ll write some text for the about section with CofounderGPT next week when we start changing the Homepage and adding the About page. And I’ll also move the Team section to the About Us page and make the roles more serious. It was fun being the Head of Janitorial services for a while, but we’re not sure how some of the people we’ll be bringing to the site will feel about our funny titles.
Updating the website
Now that we have decided how we want these pages to look like, we’ll have to make these layouts on the website. The good news is, the WordPress theme that we chose has all these elements inside it. So next week after I get feedback from Slobodan, I’ll update WordPress to look more like the wireframes above.
I think I’ll be able to do most of this on my own because the template is relatively easy to use and well documented. But if I need help, I’ll ask our freelancer friend to finish anything I am not able to.
More AI-Generated Content
As part of today’s update, I asked CofounderGPT to write and act as an editor on an article about the StoryBrand framework. Unlike in Day 13 and Day 14, we did it really quickly. Here is the process we followed:
- I asked CofounderGPT to come up with an article that explains how startups can leverage the StoryBrand framework to early acquire customers
- Then it wrote a few more “better” versions as we did above with the landing page outline
- The CofounderGPT wrote 5 alternate and more catchy titles than the ones it initially provided
- The final result was copied into WordPress directly
- CofounderGPT then suggested 5 cover images for the post of which we chose one
- I dropped the instructions for Midjourney from this page into CofounderGPT and asked it to write me some prompts using this format
- We created the images, chose 1 and added it as a cover image to the post
- The post was published on Knowlo
The whole process took about 30 minutes today. Definitely within the 60 minutes which is the limit I want to spend on creating AI-generated content.
The final result is an accompanying piece which explains StoryBrand a little bit better for those who are not familiar with it. Check out the whole article.
Time spent today: 6h
Total time spent: 105h
Investment today: $46 USD (WordPress.com payment of $62.09 CAD converted to USD)
Total investment: $959.54 USD
Paying customers: 0
Slobodan is continuing to write code and build the product with CofounderGPT while I update the website and then start getting some traffic onto it.