How to spend your first $10K on paid ads

This was the question I had to answer when kickstarting the first paid acquisition campaigns at Postmates back in 2018. Long before we received our massive funding infusion, eventually totaling to $763 million through six rounds, I was responsible for creating a strategy that would provide us with valuable learnings and be as efficient as possible.

Today we have a plethora of channels where we can send our advertising dollars in hopes of landing loyal customers for our startups.

The questions then become:

• Which channels should a startup leverage?
• How should a budget be created?
• Which tests are the most important in the early stages?

I’m here to answer these and other questions while also creating a mock $10,000 budget for the start of your paid acquisition efforts.

Which channel to start with

Without knowing anything about your startup or what it seeks to sell, I can confidently say that either Google or Meta will make for excellent paid channels for initial testing. They represent the duopoly of paid advertising for a multitude of reasons, including that they both contain robust targeting capabilities, varied campaign types and the audiences to go along with them.

For the sake of not making any further assumptions, I’ve compiled a list of questions you should ask yourself when determining which channel to select for your startup:

1. Within which channel can my target audience be found?
2. Is my product or service discovery-based?

Let’s assume you own a marketplace startup, where people sell their luxury watches from brands like Rolex or Patek Philippe. The target audience for watch buyers can be found on Meta or Google, and this marketplace is not discovery-based.

An example of a discovery-based startup would be a new type of sock that contains pockets. It’s an extreme example, but it’s used to showcase the fact that someone will not be searching for this, and you must throw it to the front of their social media feed for them to “discover” it.

Now that you’ve got your channels selected, let’s walk through a mock example of spending your first $10,000.

How to allocate your mock $10K marketing budget

Allocating your first $10,000 in ad spend is not overly complicated, but there are a few variables you need to keep in mind to take this money further. It’s imperative to ensure the $10,000 includes the following:

• Two channels.
• Campaign types.
• Messaging tests.
• Creative tests.

In the early days of your startup, it’s important to be as efficient as possible, not only with this $10,000 budget, but also with your time.

With the setup above, you will be testing two channels, two campaign types, and two flavors of messaging at the same time. Imagine the potential learnings! I’m not going to write a comprehensive guide on messaging or campaign types for each channel in this column; instead, I want your mindset to be about diversifying your spending across the three most important paid ad pillars: channels, campaign types, and assets.

You need to imitate when lean

When you’re running on such a lean budget, in contrast to those advertisers spending millions per year, you need to simply imitate. There are powerful resources to help with ideation and inspiration such as Meta’s Ad Library and TikTok’s Creative Center.

Both of these ad libraries allow you to search for competitors and industry leaders to draw inspiration from, in both their messaging and ad concepts. It is imperative to not reinvent the wheel when your budget is so lean. Instead, it is better to successfully imitate what other companies have spent many thousands of dollars on testing.

At every major startup I’ve worked at, including Postmates and Coinbase, we did a fair amount of imitation, even after we had secured far more significant budgets.

Selecting winners fast

Make sure you are tracking all your paid acquisition tests on a simple Google Sheet so that you can look back and know what’s been succeeding when you are months down the road. Too many times I see early paid acquisition spend be wildly inefficient with no set outcomes to learn which setup will work best.

As a growth lead at Coinbase, I was always working to create a brief that outlined the hypothesis for every test that I ran. Getting into this habit with your first $10,000 in spend and beyond will make your subsequent testing much more analytical.

When looking at all the metrics on the ad dashboards for your early tests, the ones that matter most are the CTR (click-through rate), CVR (conversion rate), and CPA (cost per acquisition). Whether you’re testing messaging or campaign types, your eyes should constantly be on these three metrics, even though you will ultimately pick only one to determine your victor.

Typically, the most down-funnel metric such as CPA should be used as the determining factor to pick your winners. Leveraging the insights from CTR and CVR should help guide your assets and messaging even if they’re not the ultimate decider.

There is no preset delta needed to call a winner during tests, but it is imperative to clearly define how you will determine future winners so there’s consistency. I highly recommend using a stat-sig calculator (there are many available for free online) and using a 10% to 20% significance level.

Your next $10K

After spending your first $10,000 in ads, you should be armed with an abundance of new data if you followed the mock budget plan above. Every test should bring you closer to a better ROAS (return on ad spend) and should compound over time.

Throughout the testing, iteration to the budget can be made if the goal is to create greater knowledge. For example, if one of the channels is performing terribly and you become more bullish on the other, then a cut in spending can be made.

As you think about the next $10,000 and beyond, the same principles stand, including constantly testing to outperform your evergreen campaigns. If there are two primary takeaways from this entire column, it’s that you should constantly launch new tests and that you must document everything when running paid acquisition.