how to quickly test your way to a flood of high performing ads

Hey friend, Here's my "secret sauce" for FB ads - the simple strategy that allows you to: • Find winners and scale to the moon • Churn out never-ending creative to feed the FB beast I've used this strategy with companies from small shops all the way up to behemoths like Lacoste. It has consistently found control-smashing ads. It's allowed companies to take one good ad, tweak it, and run it for MONTHS at scale. 1. The success of a Facebook ad rises and falls on what is seen while scrolling. This is really just self evident when you think about it. Stopping the scroll and getting the right people to start reading in the right frame - that's the name of the game. 2. With that in mind, we should test what people see while scrolling (and test the most important parts of that). I've seen time and time again, what really matters when testing are images, video thumbnails (and first few seconds), and the copy "intro" (whatever is seen before the See More click.) Images are important to stop the scroll, and they're super important in ecomm for easily understandable products. But it's the intro copy that is generally the most important element in an image ad (particularly with products that are hard to explain with a simple image). So this strategy is based on testing copy intros and images. 3. Write a really awesome ad and write a few different variants of the intro to test. 4. Test the 10 different ads in the same ad set. So I'll publish 10 ads - with the ONLY difference being the intros. Same body copy after the intro. Same headline, same image, same URL, same everything. I typically will run them in the same ad set. I'll choose a good audience like a 1% LAL of purchasers. I will choose just the FB news feed as the placement, and I'll exclude all warm traffic (page engagers, buyer lists, site visitors, etc.) One important point: You'll want your image to be a generic image that doesn't necessarily "correlate" to one copy variant over the others. 5. Have automated rules turn off the ads at a certain number of impressions. I've seen 1500 impressions be the sweet spot where you're getting a good feel on the data with a minimum of ad spend. So what I'll do is run each ad to 750 impressions and then have automated rules turn them off when they hit 750 impressions each. Then I'll turn them back on and run them a second round until 1500 impressions. (All that turning on and off may be overkill but that's just how I do it.) For lower budget clients, you can run to 1000 impressions but I don't like to go that low. I'd cut back on the number of variants if budget was a concern. Then again, for advertisers with generous budgets, running to 3000 impressions on each ad is no big deal. 6. When all the ads have the required number of impressions, evaluate based on CPC. 7. (Optional) If the jury is out on some ads and you want some more data, feel free to turn on just those ads and run them longer. That's the beauty of this approach. You are in control of where the impressions go (that's one reason I prefer this method to testing with dynamic creative). 8. Choose a winning copy variant. Repeat steps 4-7... but THIS TIME test images. I'll normally test 10 different images, and I'll use the exact same protocol (automated rules and all). I will test out different "types" of images - I may test out a few memes, a few personality shots, different words on the images, etc. 9. Roll out the copy winners and image winners. Profit. I typically will take the best 2 or 3 copy variants and combine them with the best 2 or 3 image variants to create 4-6 ads. Then I'll roll these out into different audiences to test / scale. (At this point, I wouldn't say hands-down that your copy variant with the lowest CPC is absolutely the #1 best. There is an 80/20 aspect and yeah, there's still some statistical uncertainty. However, I would bet money that the top 2 or 3 variants will still do super super well.) 10. Learn, iterate, repeat. THIS is where I geek out. Because we've got the same number of impressions on each ad, we're comparing apples to apples. So we can see what wins - but we can also see what loses and what does meh. Then we can go back and try to find out WHY. Was it a certain combination of benefits that really resonated? Did one pain resonate but another didn't? Now you can take all that learning and roll out into another and another and another round of testing. One client had decent success after our first round of testing - but we took the learning and did another round of testing. On that second round, we created a control ad that months later has never been eclipsed. (Super secret bonus geeky tip... The learning from these tests can also help with your landing pages, emails, and other platforms.) 11. (Optional) Use the learning from your copy and image tests to help you create videos. If creating video is a colossal pain for you, then you'll be able to learn about what works and why it works - for both copy and images. That way you can create a video that has a much better chance of success. 12. Consider intelligently testing copy and image variants. Ok so technically this is an earlier step... but now that you see how everything is working, what if we created copy variants and image variants that deliberately tested certain things? Like specifically combine certain benefits / pain points / avatar groups / objections - intelligently write the intros to give you learning for the future. 13. Tweak just the intros and the images to have a flood of new and fresh ads. So this is a corollary to the strategy above. But just by changing our images and our intro copy, we can create a flood of different ads. Not only will those changes have a huge impact, but they look like completely different ads to people as they scroll through their feed. I worked with a major brand spending thousands a day, and we took basically the same 3 ads but used dozens of different images and intros - and we ran that (quite profitably) for months. Any questions? Just reply back! - Kevin Urrutia To make sure you keep getting these emails, please add kevin@voymedia.com to your address book or whitelist us. Want out of the loop? [Unsubscribe](. Our postal address: 333 W 39th Street New York, NY 10018