How Ads Platforms works

Many people think of ads as a "Put $1, get $2 back" machine. After all, you're paying for ads to get more sales, right?

Sure, that's YOUR goal, but it's not how it "works".

There are no guaranteed results. In fact, most of the time, you will likely lose money on it, at first. (And anyone pretending the opposite is trying to sell you a sketchy course).

Ads platforms don't print money magically: They only show your content to who they predict will convert.

The word predict is important, because modern platforms use AI algorithms to choose who will see your ads. Here is how it works (in a very simplified way):

0. You create a campaign with a specific optimization goal (ex: signups).
1. The platform show your ads to people likely to signup.
2. Some of those people signup, and the tracking sends this data back to the platform
3. The platform's algorithm tries to find more people like the ones who just signed up
4. It shows your ads to those new people, and optimizes further.

Of course, this is a very simplified version, there are many other factors, but the point is that usually the more data you have, the better the AI predictions will be, and the better your ads will perform.

This is why people usually say you need a high budget for your ads. But in reality, it's not about budget, it's about volume of data.

There are other ways to get more data without necessarily increasing the budget.

But first, we need to understand how do platforms chooses when to show an ad to their users?

Audiences & Targeting

Ad platforms have billions of users. In fact, you can reach almost anyone in the world if we combine all platforms together.

But obviously, we don't want to target everyone, we want to show our ads only to our target audiences. And luckily for us, ad platforms have billions of data points about their users.

Platforms need that data to select content that it predicts will be relevant to this person, based on all their interests, and then show it in their feed.

Because of the massive amount of data, each user has a unique feed tailored to what they like.

This is why social media are so addictive: they spend billions training AI models that find the perfect content for YOU

But this data isn't just for showing content: It's also used to let advertisers target precise audiences.

For example:

• On Meta, you could target men between 35 and 45 years old, who like fishing and survivalism, have kids, and read Tim Ferriss.
• On Google, you can target people searching "How to create a landing page for my website"
• On LinkedIn, you can target CEO of tech startups with 50+ employees.
• On Youtube, you can target people watching videos about entrepreneurship.
• On Reddit, you can target people who hang out in "Work from home" subreddits.

Every time one piece of content is shown to a user (whether it's organic content or ads), it's called an "Impression."

Now, since most people are not scrolling or searching non stop all day long, the number of impressions available on each platform are scarce.

So platforms need to decide how to prioritize one content over another.

And since you're probably not the only one who wants to target those people, platforms need to decide how to prioritize one ad over another.

So, how do platforms decide? By making advertisers bid against each other.

The auction system

If platforms were only filled with ads, people would just find an alternative and leave, meaning there are only so many Ad Impressions available on platforms to keep the user experience optimal (often called the Advertising Inventory).

For advertisers, it means it's a zero-sum game: Every time someone sees your ad, they don't see your competitor's.

It also means that the more competitors run ads for the same audience, the more expensive ads will usually become.

You're competing against other advertisers to reach each person in your target audience. The more demand there is for a "person", the more it will cost to target them.

Of course, you can't target a specific person, instead, you target people inside an audience. The cost is expressed as Cost Per 1000 impressions (CPM)

The winner is the one that can afford to bid the most in the auction. That's why maximizing your Lifetime Value is an absolute priority to run ads.

Wait so are you saying I need to spend thousands of dollar every day to be able to run ads?

Well, yes, you do need money, but there is a little plot twist that can allow you to run ads even with lower budgets:

The bidding system doesn't only take your budget in consideration, it's just one part of it. There is another part that is often even more important than your budget...

Quality Score

"The better your ads are, the cheaper it will be"

This is a rule you find in most advertising platforms, and it's what can allows you to compete with bigger advertisers.

But how do platforms determine what a "good ad" is?

By using a "Quality Score".

What is the Quality Score?

The Quality Score is a "black box" metric that determines how good the user experience your ad provides.

I specified "Black Box" because platforms don't share how it works (to avoid advertisers gaming the system). But they still give us a few hints. And often, it works similarly to how organic content is ranked.

It considers things within the platform, like:

• The engagement your ads generate (watch time, likes, comments, etc)
• Following the platform's compliance & rules
• Avoid deceptive marketing tactics like click or engagement baiting
• How "good" your ads are compared to other ads targetting the same audience (yes, quality score is relative)

But also in your funnel and product experience, like:

• Landing page conversion rate, bounce rate & dwell time
• Disruptive experience (ex: intrusive pop-ups, etc)
• Misleading products (ex: Product different than what you advertise on your landing page)
• Bad customer experience (ex: Not replying to emails, refunds, etc)

Platforms have a few different ways to track those things, but the most common ones are to use their built-in metrics (likes, comments, etc), the data they have about your website (using tracking pixels), as well as direct feedback from users (ex: Facebook Page Feedback)

Ok great, but what's the point of having all this, don't they just care about maximizing profit anyway?

They do! And that's exactly why they use the Quality Score.

Why do platforms use quality scores?

Platforms use those algorithms to balance between 3 priorities:

• Make the maximum amount of money for the platform: That's the "final" goal.
• Providing the best experience possible for users: More active users = More ads to sell = More money
• Give advertisers the best return on investment possible: ROI = Advertisers come back for more = More money

It's a very tight balance to maintain, because if you show too many ads, users will be annoyed and leave the platform. But if you don't show ads at all, you go bankrupt.

Here is a quick story to illustrate this. You don't have to read it but it gives you a good examples of why platforms uses that.

"Ok that's great but I don't care, just tell me how to get cheaper ads"

Coming to it!

Why you should care about quality score

So, how does quality scores impact us? Well, simple: It's often half of how the auction system works!

Let's see an example with Facebook's auction system:

As you can see in the image above, there are 3 components. Each focuses on a priority:

• The Advertiser's bid is here to maximize the amount of money Facebook makes.
• The Estimated Action Rate is here to make sure your ads are shown to people likely to convert. Platforms prefer to send the "good traffic" to landing pages and funnels that convert, because getting conversions means you'll probably come back and pay for more ads.
• The Consumer Experience" (Facebook's quality score) is here to make sure the users see content they like and stay on the platform. If users don't like your content, you ads will be more expensive to compensate, or simply not be shown.

I'm not making that up by the way, those examples directly comes from Meta's guidelines, and most platforms use something similar.

But as you can see, the "money" part is just a small part of the equation (Only 25% according to the image).

The rest is how good your funnel is (25%) and how enjoyable your ads are for users (50%).

This is how you can compete with other advertisers. Most of them are lazy and don't want to put in the work.

They just throw money and hope it sticks. But if you're smart about it, you can spend less and still get good results.

How to lower your ad costs

So, based on this, you can see that one "easy" way to lower your costs is to maximize your quality score.

We've mentioned a lot of the things above, but let's make a recap:

At the Ads level, you can:

• Make content that's more valuable (Example: Higher production value, funny, story, educational, relatable, relevant)
• Avoid creative fatigue
• Avoid clickbaiting or deceptive practices.
• Answering comments under your ads

At the Funnel level you can:

• Making sure you choose a conversion goal that you can actually get quickly (ex: Don't ask for purchases if users purchases 2 weeks later signing up, ask for signups instead).
• Congruency: Make sure your landing page matches your ads as much as possible (similar wording, branding colors, etc)
• Keep users engaged on the funnel (Making sure they don't bounce, for example by making an engaging onboarding, showing a video on the landing page, etc)

Now, I keep repeating it, but again, this article is a simplification.

Ads platforms are very complex and I suspect not even the platform's developers don't fully understand how they work (mainly due to the opaque nature of AI).

But hopefully this already gives you a better idea of how it works.

If you still aren't sure if ads are the right fit for you, you can check When to run ads for your SaaS