Advertising Facebook Featured Social Media Marketing

3 Lessons I Learned After Spending $5 Million on Facebook Ads

Image courtesy of Pixabay

As a Meta partner agency, we spend about 3 million each year on Facebook ads. Last 2 years combined, we spent about 5 mil.

The more I run Facebook ads, the more I find that it is like investing in the stock market. You must control your emotions and stay calm when things go up and down.

1st Lesson – Do Not Be Reactive

Some marketers asked me- “hey Ted, I optimise my campaign every hour; why is it my results are still so bad?” Then I tell them this is precisely the reason why! One beginner mistake is touching the campaign too much and over-optimising. These advertisers launched their campaign, found that the results were not good, and then they made more changes to the campaign, and the results got even worse, and they made even more changes, and the vicious cycle continued. That was what happened to me early on. I thought that if I did more, then I would get more results. But it doesn’t work that way. Here’s the thing: once you launch your campaign or make changes to your ads, you must wait at least 48 hours before making any more changes. You have got to let the machine do its thing. Machine learning needs time to take effect. When you interfere, you are meddling with machine learning. And the machine learning is the most crucial part of Facebook ads.

Let’s say things are going well, and suddenly results drop. Once again, as advertisers, we need to keep our cool. We should not panic and have a knee-jerk reaction. There is something called fluctuation. In the stock market, there is fluctuation. In advertising, there is also fluctuation. Why? The common denominator between these two activities is human beings. The other side of the table is human beings, not robots. Human beings have emotions and do not behave in a fixed pattern.

It is only when you aggregate data then you will see patterns. Some marketers see the results drop, panic, and ask me, “Ted, this whole morning has no leads, I think the campaign is dying, or Facebook is dead?” I tell them, “no, this is fluctuation.” Some statistics 101, you can only draw conclusion after gathering enough data. One morning is not enough unless you spend a big budget, 10k per day or more. Rule of thumb, assuming a 2% average conversion rate, you need at least 100 clicks to achieve statistical significance and be able to conclude if it works.

2nd Lesson- Facebook Ads Have No Place for Ego

I can’t tell you the number of times I have been surprised by FB ads. There are many times I presumed something would work based on past data, but it doesn’t, and vice versa. E.g., We made a recent free giveaway offer for a renovation campaign, and it was supposed to work but didn’t. In another case, we ran a campaign to promote a defibrillator as a Father’s Day present, and it worked well even though I didn’t think much of it. The problem with FB ads is not the unpredictability; the problem is when we get emotionally attached to our ads or campaign.

This is the biggest mistake I made early on, which was to become too attached to my ads. There’s this thing that all marketers are prone to, which is the Ikea effect. When you build a piece of furniture with your hands, you tend to like it more. You are invested in it. You don’t want to throw it away, even if it is broken or ugly.

Likewise, we all had the experience whereby we favour a particular ad and swear by the advertisement to the end. Even when the results show that this ad is not working, we hold on to it. We don’t want to cut off the ad because we think it will work eventually. We become emotionally attached to the ad.

However, like investing in the stock market, you must be logical. As marketers, we have to be rational and make decisions based on numbers. If it doesn’t work and is statistically significant, we must cut off the ad. Advertising legend David Ogilvy said that a good marketer is a poet and a killer. On the one hand, you have to take pride in your work, but you also have to ruthlessly kill it and replace it if it is not working. Don’t let ego be your stumbling block.

Running Facebook ads is a humbling experience. You are proven wrong whenever you think you have become an expert. Personally, even though I have seen at least one thousand ads across more than 20 industries that we are helping, I cannot predict which ad works and which ad doesn’t work with more than 60% accuracy.

We cannot predict because many factors exist in a Facebook ad campaign. As marketers, when we create or analyse ads, we always look at them through our lens. We bring our notions, assumptions, and perspective, which is different from that of the target audience.

This is also what I like about modern marketing vs traditional marketing. In conventional marketing, everything is subjective. You have to assume and guess. However, in modern marketing, we have data for everything. So, we can tell objectively if an ad is effective or not. We use numbers to make decisions, not expert opinions.

The Final Lesson – More Technical Rather than Mindset Driven—The Importance of Messaging

Facebook is removing a lot of their interest targeting this year. Many advertisers are panicking and going nuts. But for us, we saw it coming. Facebook has been building up on its machine capabilities over the years. In several of our campaigns, we targeted broad, like 1 to 2 million people or even the whole of the country, and the campaign worked better than when we used specific targeting.

Why? This is because when you set targeting manually, you restrict the machine from doing its work. When you remove all targeting, you are giving the machine free play. Here’s the thing: the machine is smarter and is doing a better job finding the target audience than us, as long as it has enough data. So, I am not surprised Facebook is removing several targeting options. They are confident in their AI and machine learning targeting.

What I learnt is to focus on the messaging. When you get your messaging right, you are doing the targeting. The right group of people clicks on your ad, and Facebook collects data so they will find more people who are similar to the initial sample group. That is how you scale the campaign. This year and the following year, you should continue to get good at messaging.

So, there you go, the 3 lessons of Facebook advertising comes from our experience of spending a lot of money on Facebook ads. Don’t be reactive, don’t be egotistical, and focus on messaging. You will do very well if you follow these 3 lessons.

About the author


Ted Chong

Ted is the co-founder of Ice Cube Marketing, a digital marketing agency in Singapore that has been operating since 2015 and has helped more than 500 SMEs grow their business through Facebook and Google ads.