Inside S1G: How we Saved $60k, Launched in 3 Weeks and Skipped Pre-Seed Funding

Anvil.works is a drag and drop Python editor that’s changing how founders bring products to market.

When I quit my job in September 2016, I was ready to stop getting a pay check. I thought I knew what I was getting into.

I set to work breathing life into algorithms I’d been mulling over for a year and a half. I’m a data scientist by training, so I’m used to a bit of coding.

But I’m not a full stack developer, so I started getting quotes to take the core of the code I’d already written and build a web applicationAnd they were anywhere from $60,000–$100,000. Yes, it’s a complicated product. But I’d already written 80% of the backend. Why am I paying someone else to try to replicate my vision?

Frustrated and very much feeling my lack of stable income, I tried to learn how to build my web app from scratch. That’s when I found Anvil.works. It’s a drag and drop python editor that’s easy enough for beginners to use and customizable enough for machine learning algorithms.

Instead of a three-month timeline using an agency, I launched in three WEEKS. I started using the product internally and we ran a closed alpha test. It wasn’t pretty, but it didn’t need to be.

We charge $5,000/month, and that’s a product built on Anvil (which starts at $49/month). We were able to skip pre-seed funding by building a functional product. Six months out, we are a profitable business. And it turns out that the product works so well, no one really cares that it’s not pretty yet.

Meredydd and Ian are changing the way that founders can build and launch products. I shutter to think of where we would be right now if we hadn’t stumbled into Anvil.

The support I’ve gotten from them, both as a customer learning how to code and as a fellow founder, is astounding. They’ve spent hours with me debugging code, working through logic problems and building my company. And they’re genuinely nice people. We struggle in failures and share in successes together. They gave me the confidence to be my own technical founder, even though my resume doesn’t have the word “engineer” anywhere.

Here’s the thing: building a product is hard. It’s always going to be hard. But creating barriers to entry around the level of technical knowledge that founders need in order to launch makes something very difficult almost impossible. Founders need to have products and people who can support them. For me, I found that through Anvil.

Introducing: Branch Deep Link Integration for Facebook Ads

I have a customer who makes an app for small business productivity. January is their best month, and I met them two weeks into the start of 2017. We needed to launch ads, fast.

But here’s the problem - they didn’t have a mobile marketing partner (MMP - like AppsFlyer or Adjust). Here’s the scary thing about mobile - attribution for ads is really, really hard. When I run ads, I send users to the App or Google Play stores. This means they move from a Facebook or Google ad, to a page where they install the app and then actually open up a mobile app. We know which users click on a link, but the hard part is verifying that the same person who clicked the link for the ad is the same person who later installs the app.

We found a problem, and so we fixed it.

Introducing a Branch Deep Link Integration.

Now, you don't need to have an MMP in order to run mobile app install ads on Facebook. You'll just need to sign up for a Branch account (it's free!).

We'll walk you through the steps -

1. Send us a message through chat or email with your Branch API key and secret key, which you can find under Settings in your Branch dashboard.

Screen Shot 2017-02-12 at 9.14.09 AM.png

2. We'll activate the Branch integration on our back-end. It's quick - only 2 minutes and you'll be up and running.

3. Now you just need to build new ads! Click on Ad Upload in your Step One Growth dashboard.

4. Fill out the form as normal and click "Build Ads". Once they're created, you'll see a new button that says "Add UTMs".

5. Click the button, and you'll have Branch deep links added to your ads. These will track your campaign ID, adset ID and ad ID. Now you'll get access to all of the data in your back-end database, instead of waiting for Facebook to show you the results of your ads. Enjoy!

P.S. Hoping to be able to track your down the funnel metrics from Facebook ads in your S1G dashboard? Upgrade to the Elite plan now and see all the data you want in one place!

Stop Wasting Time in Power Editor

I remember the first time I'd ever used Power Editor. I was coming from Google AdWords, so I was used to exporting my data into Excel, making changes and then re-importing.

So in Facebook, I tried to follow the same steps. Exported the raw data file. Changed ads and made new creative, then tried to re-import to Facebook. Error. Four hours later, more errors. It took me two weeks to learn how to work with Power Editor, by which time I was already wondering how anyone could possibly use the product.

If you're ever used Power Editor, you're probably familiar with a few things:

  1. Dozens of import errors.
  2. The spinning loading wheel that stays for wayyy too long.
  3. Facebook reps asking you if you remembered to restart before you log a bug.

We are in the age of big data and data science. Why are marketers still working in and out of Excel? Why are we saying still while every other industry is moving towards advanced programming?

Stop wasting your time building Excel spreadsheets and move into the future of advertising. Step One Growth will beat your efficiency every single time.

Go ahead and try me. I dare you.

4 Reasons You're Advertising to the Wrong Audience

Let's say that you're selling a toothbrush. Who are your target customers?

A marketer might tell you to find people interested in dentistry. Or maybe people with braces, gum disease, or dentures.

But here's the truth: everyone has teeth.

1. You're limiting your audience size before you even start advertising.

As a business owner, you need to think about the entirety of the addressable market for your product. If there are 10 different groups of people who could use your product, why would you only show ads to 1 group? You are limiting the potential of your marketing before you even start running campaigns.

2. You're only advertising to 1 group.

Facebook has more than 72 quadrillion targeting options. Google has even more. Limiting yourself to a handful of groups means that you're not only decreasing your potential audience size, you're decreasing your learnings. If you want to know who your true users are, you need to test.

3. You're not programmatically finding new groups.

Here's the thing about people: we don't know what we don't know. When I think about toothbrush audiences, I relate back to myself - when did I last buy a toothbrush? But that means I'm missing the other billions of people and their specific reasons that they might have purchased the toothbrush. If you're still trying to brainstorm for ideas yourself, you need to let go of your ego. Data science is here to help us find new targeting groups, and we need to let that in.

4. You're using black-box algorithms.

This is the difference between saying your target audience is a 2% lookalike Facebook audience and your audience is 30-35 year old women with 2-3 children under 7 years old with advanced degrees interested in photography. It's time to use data to test audiences and learn from those tests.

Who are your customers? You'll only find out one way: test.

CVRs up 50% and CPCs Down 80% Since Holidays

January is my favorite month of the year for advertising.

If you missed out on the post-Christmas advertising extravaganza last week, don't worry! You still have plenty of time in the haze of the first month of the year.

Whether you're a B2B product looking to capitalize on new budgets or a nutrition app taking advantage of New Years' resolutions, it's the perfect time for you to advertise.

Message us now and let's get started!

What will artificial intelligence mean for marketers?

There are a variety of different answers to this depending on what part of marketing you’re referring to. My background is in advertising, so I’ll answer it just for running ads. This is definitely a tough and thoughtful question, so I’ll try to separate it out into parts:

What’s Available Now:

  • Setting Bids: Computers are better at adjusting bids and reading in data than people. This is just a fact. Computers don’t make math errors, they don’t miss rows in a table, they don’t get distracted by a YouTube video. At Step One Growth, we’re able to hit our customers’ goal metrics and help them expand and grow their company by having the computer calculate and set bids.
  • Ad Optimization: Just like computers are better at setting bids, computers are better at determining winners of different tests. This means they can select winning and losing ads, separate them out and make sure you’re making as much money as possible from your ads. We do this now, too :) On Facebook, we’ve actually found that performance improves with fewer ads in a campaign.

What’s Almost Ready:

  • Language Processing: Computers are good at a lot of things, but understanding language is not quite there yet. We’ve found that it’s easy for computers to do things like read a definition or from a thesaurus, but not so easy to find contextually similar words. Learning car is like vehicle is easy, but that car is like driving is hard. This is at the forefront of new technology in advertising, because this is what we’re creating at Step One Growth. We take keywords and expand them to find new and similar targeting groups. It’s still a work in progress, but is helping customers find new and different audiences.

What’s Coming:

  • Imagine just entering in your website or mobile app and having the computer automatically do everything else. From finding keywords to setting both your perfect price and your perfect bid to writing the best sentence and creating a beautiful ad image. This is the future of automation in advertising. This is what we’re building towards. As machine learning advances and as data continues to accrue, artificial intelligence will begin to not just build upon the existing infrastructure, but also create anew.

It’s going to be amazing. And I can’t wait to be the one to introduce you to the future of advertising.

Hope that helps! Feel free to message me or reply here if you have any other questions.  

How much does it cost to get an app into the top 20 games?

Hi there! I’m Colette for Step One Growth.

It’s going to be ridiculously expensive. I work with companies who have budgets up to $250,000/day. If you can’t compete with that, you’re going to have a hard time reaching the top. But there are some games that are absolutely successful without spending that kind of money on ads.

At the end of the day, the cost to get into the top 20 will depend on a bunch of factors, namely:

  1. App store optimization: are people searching for terms and finding your game?
  2. Virality: are people who are currently using your app referring their friends? This will help make sure that you can increase the number of downloads (which will improve your ranking).
  3. Velocity of installs: how quickly are people downloading your app? If you make your app look like the next big thing with thousands of downloads an hour, you’ll have a better chance of increasing your rank in the store. But be careful with this! Many companies have started to buy incentivized installs (where people will download your app but not use it), and the app stores have been monitoring these efforts more closely to stop this behavior.

Hope that helps! Feel free to message me or reply here if you have any other questions. 

Are certain days of the week better than others for advertising?

There are definitely day of week and time of day trends that you’ll find in advertising. But the trends do change for businesses.

I work with a bunch of different types of companies, and my B2B SaaS companies tend to have high volume and low costs on Mondays, but no conversions on the weekends. Dating applications face the opposite problem - high conversions on the weekends and few on Monday/Tuesdays.

The best thing you can do is gather data. Try to learn more about your audience and when they’re interested in your product. Also think about the time it takes to go through your funnel. If it’s just an hour, then it’s easy to correlate advertising time with purchase time. But if your cycle could take days or weeks, be careful before turning off your ads on one day.

Hope that helps! Feel free to message me or reply here if you have any other questions.  

How do I get traffic from Facebook ads?

Facebook ads are an awesome way to generate awareness, get traffic to your product and increase your revenue. The first step is going to be just setting up your accounts.

After that, you’ll want to try to advertise to users who are the most likely to be interested in your product. This is the process that we help and automate for you at Step One. You can start by entering in a few keywords that you think your audience would be interested in (eg if you’re a musician, you might write in music) and then we’ll automatically find more groups for you to target.

Once you get started, the sky is the limit! I have customers who pay $10 a day on Facebook ads and customers who pay more than $250,000 a day. Once you find the right product-marketing fit, you can do anything. Facebook has billions of users, which means you can find billions of potential customers :)

Hope that helps! Feel free to message me or reply here if you have any other questions. 

How do I get laser focus audience in my FB ads?

Hi there! This is Colette from Step One Growth.

This is exactly the purpose behind my software (and I’m sorry for blatantly talking about it here…). S1G builds out audience groups by keywords and splits by age, gender and location.

For example, if you have a photography website, your keyword might be “photography.” S1G then finds related interests, such as camera, picture, picoftheday, etc. The next step is to divide each of those interests into groups by age, gender and location. So at the end you could have an adset that targets “camera”, women ages 20–25 in the UK. This is an awesome way to get really advanced and detailed targeting to find out who your customers are.

Will two or more Facebook campaigns compete against each other?

Hi there! This is Colette from Step One Growth.

In general, yes, your campaigns will compete against each other.

One way to potentially avoid this is to increase bids by 1 cent in your Android/iOS campaigns and keep the bid the same in your other campaign, but if the ads end up having different relevance scores (this is likely), then you’ll be back to potentially competing against yourself.

Why do my 2 identical Facebook ads have huge reach difference?

Hi there! This is Colette from Step One Growth.

There are a ton of reasons why this could happen. A lot of statistics go into determining your reach on your ad, even when the creative is the same - bid, budget, relevance score, audience, etc.

For example, if one of your ads only targets men ages 30–32 and the other ad targets men and women 30–40 years old, the second has a much larger potential audience and will have a larger reach.

Even when your audiences are the same (let’s say both ads target men ages 30–32), you can still see different metrics in Facebook. This typically happens as a result of relevance score, bid and/or daily budget. The relevance score is calculated by the percentage of people who interact with an ad - this can be likes/clicks, comments, or just spending time looking at the ad. The higher your relevance score, the lower your cost per view, and the larger your reach is going to be.

How do I stop Facebook ads results from declining?

Hi there! This is Colette from Step One Growth.

Welcome to the holiday season. Ads started running on November 1st, so it’s not at all surprising that just recently you’re seeing sharp declines in your performance metrics.

You have basically two options:

  1. You can stop advertising until December 26th when ads go back to normal.
  2. Double down on testing.

If you aim for #2, I highly recommend testing your audience groups. If you’re not already, then start breaking out the groups by age/gender. Layer on custom audience targeting on top of your other successful targeting. If you’re interested in learning more about this, or if you’re looking for a general FB structure and creative audit, feel free to message me on my site. You can build out 1k ad sets in 90 seconds through Step One Growth, which helps so much with testing calendars and schedules.

Can I use paid ads to get views on a YouTube video and earn ad revenue off the video?

Yes, you absolutely can! You should know in advance that this is how Google makes money, so you’ll end up spending more per view than you would earn per view.

BUT if you have a video that’s prime for virality and that people will share with their friends, then it’s completely possible that you could basically boost your YouTube video to get some traction and then end up getting a bunch more views as a result of the ads (and if you’re interested in trying this, let me know - I’d love help you for free if you’re willing to be a case study!)

How much could a mobile app in the health/food/diet space make off ads with 10k daily active users spending an average of 15 mins a day on the app?

Hi there! This is Colette from Step One Growth.

OK - time for some math here.

You have 10k daily active users who are each spending 15 minutes per day on the app. Let’s say you run banner ads that refresh every 30 seconds. That’s 2 ads per minute = 30 ads per session = 30 ads per user.

Now that means we have 30 ads per user * 10k daily active users = 300,000 ads per day.

Banner ads have very low revenue numbers, so you’re probably going to be somewhere between $0.50 – 1.50 CPMs (cost per 1,000 ad impressions). Let’s say you’re making $1 per 1,000 impressions. That puts you at $300/day.

Then per month you’re at $9k/day. Pretty good!

I don’t know your app or your product structure at all, but in general I would recommend that you also try to monetize through something other than ads. Your app is definitely providing value to your users if they’re spending 15 minutes per day on it, so you need to understand what that value is and where you can add a paywall. Make sure you test everything! The more you test, the more you’ll learn and the better your product will become.

How do I target a need / Problem with Facebook ads?

Hi there! This is Colette from Step One Growth.

You actually can’t target users in that way on Facebook right now, and that’s exactly the problem that my product tries to solve. We take words that people would actually say or read online and try to match those to Facebook interest groups (that you can target through Facebook ads).

For example, if you searched for “photography” through the software, you’ll find photography related interests, but you’ll also find weddings, puppies, kittens, memes and a whole bunch of other tangentially related interest groups. This helps you match people who you know are your consumers with Facebook’s targeting algorithm so you can find the best audience for your product.

In general, just make sure you test everything and anything. You never know what will or will not perform for you as an advertiser, so the best that you can do is test as many ideas as you can.

How can I write compelling ad copy for Facebook ads?

Hi there! This is Colette from Step One Growth.

First, you should know that there are no hard and fast rules for every single company about what works and what doesn’t in terms of ad creative. Not even marketing experts can predict 100% of the time what ads will perform better than others.

That means that the key to writing good copy isn’t necessarily in the writing itself; it’s in the testing. I have a math degree, but I’ve written ad copy that’s been incredibly successful for years for multiple companies. Copywriters and blog writers might not even consider it “good”, but it still performs well in terms of both CTRs and CVRs. Set up a brainstorming event with anyone who you’re working with and let Facebook’s algorithm pick the winner (here’s my blog post on how to test ads on FB). The best way for you to learn is just to try as many different variations as you can think of.

I don't notice display advertisements on major online publishers. Why is this?

Hi there! This is Colette from Step One Growth.

There are a few answers to this question:

  1. They’re showing ads and you don’t notice. Some sites are amazing at integrating advertising into their normal content, so you don’t even realize that what you’re reading or seeing is an ad and not part of the regular site. Conversely, the larger publishers might also try to hide the ads so that they don’t detract from the experience on the site.
  2. They’re making money in other ways. Many online publishers have subscription memberships. Others work through promoted content - other people will write an op-ed piece on the site that links back to their own company and will pay for that opportunity.
  3. Not every company wants ads in their product. Sometimes, it’s just up to the founders. Giant companies like PayPal could run ads on their site and make some extra money, but they don’t want to detract from their product.