Launching your idea, from start to finish

I reached out on Twitter the other day to ask what early-stage builders grapple with advertising on social media.

If you’ve read any of my previous blog posts about launching your startup, I really like social media ads as a way to test and iterate on ideas quickly, so you learn what the market wants—and what it doesn’t care about.

65% of the people surveyed either find ads too expensive or didn’t understand why they should even try to tackle them in the first place. Good news, then, ’cause this post is for you.

How to make social media ads worth your money

Do you have a new idea for a startup you’d like to build, or would you like to try out an idea to see how you could grow your current startup faster?

The fastest way to build and launch it is to test the idea out with a few hundred strangers. We’re going to go through the 5 steps you need to build and launch a new idea today. They look like this:

1. Create – Who do you want to test the idea on? Who do you want to avoid?
2. Write – Landing page, core signup loop, and ad text.
3. Design – Come up with the look and feel of your ads and your landing page.
4. Launch – Publish to Facebook, Twitter, or wherever.
5. Learn – See what the market has to tell you. Iterate, if needed.

How do these steps make social media ads worth the money? One, when your immediate expectation for your social media ads is not that they return cashback but rather speed up learning, you’ll be able to get ROI even if you’re not a marketing expert. And two, with a few tricks, you can make your cost per human site visitor substantially less. 🙂

I’m going to embrace the #buildinginpublic mindset and create my first product launch over the course of this blog post. I write about a 9-week experiment cycle I use to build and test million-dollar funnels, and today we’re creating the landing page and ads for the first round of experiments outlined there.

Create – What do you want to say?

Every company has its own needs for what to say here, but we’re going to work out the demand and the need for a product that I’ve wanted CrowdTamers to build for nearly a year now: CrowdTamers AI.

My mission is to help 1,000 founders grow their startup to $1MM a year in revenues in 7 years, and I can’t do that without the kind of scale that AI brings to creative endeavors.

The good news is you don’t have to build something to test it.

With knowledge of who I want to appeal to and what I think they care about, let me create a landing page for an AI product that I could build and see if there’s demand. That way, if I find enough interest, I can go through the expense of actually creating it.

In my case, I begin with a thumbnail sketch of who I want to create an offering for: founders of companies that have about a half-million dollars in the bank with a working product and who need to figure out how to get users.

I know a common problem they have—heck, the Twitter poll validates it again! They want to build, or they have built, but they don’t know how to bring people to try their product. Often, their landing pages are clunky, their value proposition is poorly defined, and they don’t understand their users’ pain well enough to hone in on a good presentation for their product and their company.

What if I could automate that away? AI text generation, some fairly broad use website templates, simple image selection, and design, integration directly into ad company APIs, and even creating automated reporting and feedback loops can all combine into AI powering a lot of the heavy work in writing, designing, and launching ads to test GTM fit. The idea is simple: growth marketing for tech founders, powered by AI.

Write – Don’t what you want to say, say what they want to hear

Here’s a draft of a landing page in Notion, my team’s wiki / PM tool of choice.

A few things to note here. The text of the page is all created in GPT-3, but it did take longer than 40 seconds. This is selling the idea of what the product could be, even though there’s nothing yet. That said, if your ad promises something that literally cannot happen, your test is not going to deliver you much value.

Once it’s written, I need to also draft up some ads. The main utility of your first few rounds of experiments is learning about messaging and language, so we want to make sure we create several different variant messages to work from.

Here’s the template we use in my team to create ads on various platforms.

I’m going to test 3 different theories I have for those who might be interested in a tool like the one I’ve outlined above:

1. Some people just like cool AI products & would love to try out something that simplifies GTM planning so much.
2. Tech founders who have a product but who have yet to figure out their go-to-market fit want an expert to help but often cannot afford it.
3. Early / first marketing hires at growing startups who need some help building & launching campaigns but who don’t have a budget for the ~$5k a month that it costs to bring in even a small agency to help.

So let’s write out a few ad headlines:

CrowdTamers AI ad copy

Design

This bit—for the moment—is hard to automate away with AI. But with a few minutes of work, you can build something from a high-performance ad template.

There are tonnes of resources out there, but I figured I’d share the ad templates I’ve been using recently to great effect for a number of clients.

Pick it up, completely free, here on Gumroad.

We use them in CrowdTamers because they’re simple enough that every team member can edit them—swap out the image, the highlight color, and change the text, and you’re done. They do pretty well, too, with winning variants doing >5% click-through rate for 3 different clients.

We also need a landing page to send traffic to. If you’ve never launched something before, I can honestly recommend that you just build a WordPress site and grab one of the free themes from WordPress.com or one of the relatively cheap ones from ThemeForest.

I built a landing page in 2 hours using the default Gutenberg builder in WordPress. It’s not beautiful, but it gets the job done:

The page itself is live at https://crowdtamers.com/ai, although the buttons don’t really do anything yet as of the time of writing this blog post.

With your ads designed and your landing page created, you’ve actually got your MVP for whatever product you want to test. No need to actually create the product. The CTA here on these landing pages should say something like “SIGN UP” or “SIGN UP FOR BETA ACCESS” and collect an email address. You’re building a list of interested people if you do launch.

Launch

We’ve already detailed how to create a top-of-funnel ad on Facebook here a few months ago, so I’ll refer you to that article for the mechanics of how to launch on Facebook.

Targeting your ad is an important part of your launch, and there are a few things you will want to consider that you won’t find in the Facebook ad targeting options.

1. If you block all mobile phone traffic, you’ll reduce your reach hugely, but also, in my experience, you massively reduce the amount of bot traffic your ads get. This means that a higher overall percentage of your visitors are real humans and more likely to convert. If you’re selling something that’s not mobile-centric, you may want to test desktop-only traffic.
2. The most powerful targeting comes from either combining targeting with AND statements or adding in exclusions to your targeting. When you fill out the detailed targeting information on Facebook, these are “advertise to people who fit ANY of the following”.

You have to click on “Narrow Audience” to be able to add in an “AND” statement (so they must fit any of the first categories AND ALSO fit any of the seconds), and you’ll get much better targeting. For example, the above targeting is 12 million people, which is really too broad. With narowed reach, it’s much better: 270,000.

Learn

I’ll be launching these ads on Monday, and by #buildinginpublic, with my product launch, I hope to learn and share what I’m learning with this product. I’m following the same steps I’ve used for many clients before, but I can’t share all of the details for that because it’s someone else’s project. CrowdTamers AI is our own project, so I’m looking forward to continuing to show you what’s going on behind the curtain.