If you browse any sort of online forum about digital advertising, there's one piece of advice that routinely gets offered to new Facebook advertisers. It goes something like:

• When you're starting out, your pixel doesn't have any historical conversion data. As such, it doesn't make sense to run a conversion campaign; you should just run traffic campaigns.

This advice is often followed by some variant of the dreaded phrase: running traffic campaigns will help warm your pixel up.

Don't get me wrong, there are some small elements of truth in this advice. It is true that, as a new advertiser, you're at a disadvantage from the fact that you don't have any historical conversion data. As you accumulate more data, this disadvantage will evaporate.

Even giving credit for this, the above advice about running traffic campaigns is still wildly inaccurate. Let's understand why this is.

What actually is a traffic campaign?

A traffic campaign is simple. It's a campaign that tries to generate as much traffic as possible within certain constraints, be these budget or efficiency (cost per traffic). The word traffic itself can refer to either link clicks or landing page views; it doesn't matter which you choose to think of.

Now, what's so bad about generating as much traffic as possible?

Let's say your traffic campaign is using lowest cost bidding, and is therefore budget constrained. Facebook is essentially trying to deliver you as much traffic as possible within your budget, which is equivalent to saying that it's trying to get you the lowest cost per traffic possible.

Audience stratification

Within the total audience of people who use Facebook properties, there are different sorts of people. There are some people who tend to convert on ads (people who actually buy stuff from ads), people who just click on ads but don't convert, and others who don't interact with ads at all. This is a bit of a simplification, but it's broadly correct.

Now the people who tend to actually convert on ads (call them Converters) are the most valuable to advertisers who care about conversions. If you run a conversion campaign, these are the people you'll be targeting.

Because conversion campaigns make up for such a huge percentage of overall ad spend on Facebook, and because only a limited pool of people are actually Converters, the price to reach Converters can be significantly higher than the price to reach people who click on ads but don't convert (call them Clickers).

So, it's more expensive to reach Converters than it is Clickers. Given that traffic campaigns are really interested in getting the cheapest clicks possible, they're going to target Clickers.

It's not that the campaigns are intentionally going to avoid showing ads to people who are likely to convert (Converters), but rather that the campaigns are going to be priced out of those auctions. Because there are plenty more advertisers willing to bid far more to show an ad to a Converter, it's inefficient for a traffic campaign to show ads to Converters.

As a result, traffic campaigns tend to optimise for showing ads to people who are likely to click on your ads, but not likely to convert. Any users who have a tendency to both click on ads and convert on them will trigger more competitive auctions, which are likely to be won by advertisers running conversion campaigns.

Conversion propensity

You might be reading the above and thinking: how does Facebook know who is and isn't likely to convert, if I've only just set up my pixel?

This slightly misses the point. The point is that some people are fundamentally more likely to convert on an ad that they see than others, and that this fundamental propensity to convert can outweigh factors like whether or not they're likely to convert on your ad specifically.

Don't believe me? Ask your friends. You'll almost certainly have friends who either don't click on ads out of principle (so they're not even Clickers), or who may out of curiosity but will intentionally never convert (making them a Clicker). You might even have friends who acknowledge that they are happy to sign up to something, or make a purchase, off of the back of an ad (making them a Converter).

Each of us has our own relationship to ads that determines how we choose to interact with them. What many people don't realise is that this plays a crucial role in determining who ads get shown to on ad platforms.

You might wonder why, out of a target audience of 1 million users, your conversion campaign only ever shows to 100,000 of them. In truth, it may be that only 100,000 of your target audience have ever converted on an ad before, making the chance of getting the other 900,000 to convert incredibly low.

Tying this together

So, where does this leave us?

Well, as we've seen, running traffic campaigns optimises for bringing in the lowest cost per traffic possible. Because users who have a propensity to convert through ads are more expensive to reach, traffic campaigns don't tend to show to these users. Instead, they tend to show to users who are likely to click on ads, but not convert on them.

What this means is that traffic campaigns will almost always result in conversion rates far lower than conversion campaigns.

That the conversion rate on a traffic campaign is lower is hardly surprising. What is worth noting though is that the conversion rates are typically so much lower that they are nowhere near compensated for by the low cost per traffic that a traffic campaign will bring. As a result, traffic campaigns often have a sky-high cost per acquisition (CPA).

All of this might be forgiven, if there were such a thing as pixel warming. If the high CPAs on traffic campaigns in the short term were made up for by lower CPAs later on, then all would be well.

But this is hardly the case. The fact that traffic campaigns almost intentionally optimise away from users who are likely to convert means that they're ineffective at providing conversion data to your pixel. You'd be much better off running conversion campaigns from the get-go, regardless of how warm your pixel is.