From Testing to Triumph: The Facebook Ad Strategies That Drove Five-Figure Sales

written by Tony Ruiz
On Jun 14, 2023
Hey Traffickers! You catch me in a motivational mood today, I can’t stop reading and watching people explaining how to make ads and I really don’t know why, but I do them in a totally different way than most people 😂.

According to what I see, people make it way too complicated. Let me tell you how I achieved this ROAS, exactly as we did it for some e-commerce clients who sell infoproducts. Here’s the lowdown!!!

  1. Take advantage of custom audiences. This is something a lot of people tend to overlook when it comes to retargeting, but it’s crystal clear that it was a key factor in generating a large number of sales for this brand.

Custom audiences are audiences you can create using Facebook pixel data to recapture web visitors, for example, ATC 30 days, Page views in the last 7 days, etc.

Given that this particular brand already had a pixel installed, we used their data and immediately retargeted those who had shown an interest in purchasing in the last 30, 14, 7 and 3 days, focusing solely on abandoned carts and checkouts. (We did it this way because there was more than 1.5k of traffic in this audience in the last week, it’s a considerable audience.)

If you have less traffic, the 3-day conversion will not be as useful because you will hardly reach anyone. You need to increase the traffic for it to be useful. If you don’t have traffic and a 3-day conversion window, Facebook will always try to get the cheapest conversion for you, so it may show the retargeting ad to someone who abandoned their checkout 27 days ago because it’s cheaper, and it will convert worse than if it was just 2 hours since they abandoned their cart or checkout.

That’s why it’s important to INCREASE TRAFFIC + 3 DAY RETARGETING)

For each audience, we created variations of the creatives using a different copy style and a different offer in each one, as I always say, 3 copies/creatives x 3 variations in titles, 3×3 = 9 ads, the more ads the better, as long as they make sense, I see you coming 😆.

For the audience of the category of abandoned shopping in the last 30 days, we offered a 12% discount on their purchase and the same was applied to the abandoned carts, but using more direct text to get them to come back and complete the purchase. (It’s a colder audience, we need to warm them up a bit!!)

This principle was applied to the audiences of abandoned carts and checkouts of 7 and 14 days, with the difference that we offered a 10% discount on the entire purchase to the 14-day audience, and a 5% discount to the 7-day audience with softer ad texts (they’re warmer, you know) but still creating a sense of urgency for the customer to return and buy the product.

You can imagine the temperature level for the 3-day ATC 🤣.

Just by using custom audiences in a simple and effective way, we were able to generate multiple figures, five digits to be exact, in sales for this brand.

  1. Test cold audiences TO THE MAX. (TO THE MAX IS TO THE MAX) The key element for TOFU (Top of the funnel) ads (those cool names that the gurus come up with 😂) is testing.

This includes testing different copies, images, image formats, and audiences. (AND AUDIENCES) To do this, we started by testing different audiences.

We created lookalike audiences that are in line with the custom audiences to capture new prospects for the brand that shared similarities with current customers. (you know, the custom audiences are large, with little segmentation, let FB do it for you, I won’t tire of repeating this, 1 darn interest/behavior per adsetNO MORE, DO NOT SEGMENT WITH A THOUSAND INTERESTS, this is done later by stacking interests that work, patience, always start the lookalike at 1% (don’t go crazy, slowly!!)

We also looked at the product reviews on the website and their backend to see which age range was buying and saw that most of the traffic was coming from people over 25 years old.

Therefore, we tailored our copy and targeting to an audience over 25 years old. (something quite obvious on the other hand, that’s why it’s important to let the algorithm do its job and tell you; HEY!! you are selling to people over 25 years old!!! with this info you can segment, but if I target people over 40 because I had a dream where a fairy godmother told me that was my audience, I will be missing out on people aged 25-40, don’t listen to the fairy godmother who speaks to you in dreams👸 listen to the algorithms 😂 leave your convictions/imagination out of the equation and you’ll do much better)

As for the ad copy, we tested short and long texts and found out that the long ones gave better results than the short ones, i.e., the metrics for the long ones were better than the short ones, BUT (there’s always a but) the short ads had a higher conversion %.

How did we realize this? With the UTM parameters of the URL, they’re important for scaling the ads and adsets that convert better, not those with better metrics. If the audience is large you’ll be able to scale it much better if you know which specific AD converts more, regardless of the CTR etc.

Then, we tested professional images and simple images taken in a more informal way (with an iPhone, basically).

Being an e-commerce, the professional images yielded better results, although the informal images still had a decent performance and were also able to convert.

In addition to the image styles, we also tested different formats. We tested individual images, videos, and a carousel. We noticed that the videos didn’t perform as well as the images and carousels, so we stopped using all video creatives and focused on images and carousels. (thanks fb for the data 🤪)

Individual images performed slightly better than carousels, but both were still generating incredible results and conversions.

Once we had this info, we started splitting into custom campaigns (cross-targeting, nothing out of the ordinary 😂) and re-tested by [formats-devices/time].

After all, testing is really the most important thing if you want to have good long-term results, until you see what’s working and why it’s working, let the algorithm work for you, that’s what you’re paying for!!! 😂 so make sure not to put a large budget at the beginning ($5/adset) and also don’t use CBO (DON’T EVEN THINK ABOUT IT) since you don’t have any info and you’re not sure what’s going to work, better let them learn with tested ads in normal campaigns (don’t try to run before you can walk, you’re rushing it 😉)

Once we had everything tested (7-10 days, updating campaigns daily), we put everything into a CBO, and here are the results.

ROAS example facebook ads campaign - Traffic Roosters

Always remember to keep things simple and stick to the fundamentals, because that’s what scales, don’t bother trying strange things, just keep it straightforward.

In summary;

  1. Lots of Ads.
  2. Many large unsegmented audiences (1 interest/behavior).
  3. Let the algorithm do its magic 🔮.
  4. Remind the people who have left how good your product is depending on how cold they are, and scale lookalike with the data you have!!!!
  5. Stick to the ads that Facebook is telling you are working, don’t use the ones you think are prettier (let the algorithm speak, that’s what we pay its salary for 😂)
  6. Start removing Ads and adsets that have a cost higher than the average purchase cost. (why would you leave them? Check the URLs with UTM to be sure!!!)
  7. Take all the data and split into cross-selling campaigns (time, device, blah blah blah, the usual).
  8. Retest everything.
  9. Take everything that works for you and put it in CBO campaigns.

I hope you guys were able to find some value in the post and that it helps you beautiful people!!! 🤗 I’m leaving you a screenshot of the last CBO campaign (tested ads and adsets).


Tony Ruiz

Mastered complex sauces in the kitchen, but found a natural talent for digital strategy. Former chef turned proud CEO of Traffic Roosters 🐓



Submit a Comment

Your email address will not be published. Required fields are marked *

Related Posts

No Results Found

The page you requested could not be found. Try refining your search, or use the navigation above to locate the post.