I’ve been quietly running tests in my “laboratory” for a few weeks now, and I’m here to talk about the things I’ve learned and results I’ve had (or didn’t have).

Why put so much effort into reporting things that didn’t work?

I think it is important to the learning process.

By analyzing the steps someone else took, ideally you will be able to conduct your own experiments and cut the learning curve or reduce errors.

So,without further ado, let’s dive into some of my experiments.

My Tests

Ok, so I started these experiments after learning more about funnels and how they are built. I built a couple of super rudimentary ones to start everything out.

Test 1: An ad promising a lead magnet that lead to a Clickfunnel with an upsell to my product.

This ran over several split tests to find the right ad and audience combination which included seven campaigns over the course of almost two months.

I paid freelancers to create lead magnets (digital PDF downloads) and I ran ads to the free download. The download was delivered via a squeeze page in Clickfunnels and at the conclusion of the email submission the next step in the funnel was a sales page for one of my products.

On both the squeeze page and the sales page I was super authentic, telling a story about my family and why we started our brand. Despite this, I didn’t make a single sale. However, after roughly $300 spent on advertising, I managed to gain about 165 email subscribers.

Unfortunately, subsequent attempts to activate those subscribers with promotions or other activity did nothing. Open rates stayed below 10% and click through below 2%.

It was a good learning experience but ultimately an utter failure.

The question is….why?

What did I do wrong?

Be more nurturing. Be more engaging. Create curiosity.

I have a couple of ideas. For starters, I think it is hard to activate someone from a free download to a get-your-wallet-out-and-buy scenario. Not that it is impossible…just difficult. Also, my funnel was absolutely terrible looking. Now, I know some of the most successful funnels are ugly from a design perspective, but this was pretty Amateur level.

Finally, I should have had a regular email sequence running from the beginning. I didn’t send out any emails for days after subscribing, and I could have either gauged how responsive the list was sooner or activated them more readily if I had communicated more frequently.

Notes for the future: Better design and more interaction.

Test 2: Discount through Messenger ad.

This one definitely has seen success with many others, so I was sure it would work. I put a great ad up, using all the winning elements from my test, plus an awesome video. I paid about $0.94 per message (not great, but not horrible) but still didn’t get a single sale. Not even with a coupon for $4 off a $12 product.

What went wrong here?

Well, for starters, I used $4 off. Turns out anything less than $100 should use a PERCENTAGE off instead of a dollar amount off (learned that from the book “Predictably Irrational” by Dan Ariely).

But that shouldn’t have deterred people that showed interest and interacted with my bot. Well, upon closer inspection of my chatbot flow I see that it just wasn’t very engaging. It was extremely cold and to the point: Want a discount? Here!

Not that there is anything wrong with cold and to the point, however, the ad didn’t mention the price. So, most of the clickers were likely shoppers looking to see what kind of deal I had to offer. At the end, I had done nothing to make them see the value of the product or the value of the deal.

Notes for the future: Find a way to better convey the value of the product.

Test 3: I offered a lead magnet again, but this time through Messenger and added a sequence to do the selling for me.

This time, I used Messenger (being frustrated with the terrible open and click metrics on email) to deliver the lead magnet (pdf download) promised in the ad. Everyone who engaged with my bot got the lead magnet, but also were subscribed to a messenger sequence. In that sequence I placed both helpful information and a sales pitch to my product.

This resulted in no sales.

Ultimately the winning ad cost only $0.71 per message. Even with a 93% open rate and a 30% click through, the moment I asked if they were interested in learning about our products, clicks reduced to almost nothing.

The few that did make it through simply did not buy. So, this was also a complete failure.

But why? What went wrong here?

Again, I think getting the people onto my list with a freebie and then trying to sell them is a harder way to go. Also, my sequence was only four days long. I didn’t nurture them enough to try and sell them.

And I guess I didn’t find compelling enough language to use to get them interested. So all in all I’d say this was mostly a copy failure.

Notes for the future: Be more nurturing. Be more engaging. Create curiosity. (All easier said than done).

Test 4: Free plus shipping offer.

So, I found some really inexpensive but complementary products on a dropship website and decided to do some dropshipping. NOT because I am changing business models, but because it gives me a few more concepts to test with and I can do it without devaluing my products and losing too much money.

I set up the ads with the winning elements from my tests plus a video. The ad went straight to the order page on my Shopify site. I set it up as landing page views because I wanted to teach the pixel what success looks like. I ended up paying an average, across five ads, of $0.14 per landing page view. What’s more, I spent $50 and ended up with 15 add to carts and 13 initiate checkouts ($3.85 per checkout…which seems standard).

The problem is, only three people actually completed the checkout, so I had 10 abandoned carts. Overall, I was paying around $17 per purchase, which does NOT work with a free plus shipping model. Even still, I don’t see this as a failure.

Why? Because the ads KILLED it. The offer also KILLED it. My cart was the problem (and my abandoned cart sequence, apparently). This means I finally cracked the code (somewhat) on making sales with Facebook ads.

Notes for the future: Definitely have to optimize the site and checkout process. Also, the biggest takeaway here is that it is all about the offer! The offer is the most important element. Even with crappy homemade “meme videos” shot with a mobile phone (which I used) a good offer stands out.

Test 5: Rebate offer ad to Messenger flow.

Ok, when I said all of my ad tests failed, I kind of fibbed. This one, I believe, was a huge success. I ran an ad that offered a free product. It was a Messenger ad that lead the customer through a flow which ended at a rebate offer.

I spent $17 and ended up paying $0.25 per message (yay…those are all subscribers). I also got 34 purchases, so each rebate cost me about $0.50!

I targeted a pretty low volume keyword, but even still it put me to number one, and after two weeks the listing is still number three for that keyword. Furthermore, the listing is number five for a related key phrase that encompasses that keyword. But that’s just the ranking effects.

The other benefits are that I have subscribers, all tagged for participating in the rebate. This means I can reach out to them and ask for feedback (**cough cough..ahem…feedback) on my product. It also means they will likely be interested in future deals and promotions.

And, in fact, they are. I already invited all of them to a special group for my brand, and they ALL joined. Talk about laser targeting your market.

Test 6: 30% discount to Messenger flow.

I created a discount deal which is being delivered through Messenger. So far, I’m paying a little over $2 per message (bleh) and have only seen three coupon redemptions (at roughly $17 each again).

This ad isn’t performing nearly as well but I think I have an idea why. First, the offer isn’t as amazing. There are a LOT of offers on Facebook for physical products these days. Real or not (many are scams) they are offering lots of free products or rebates or deep discounts. So much that I think the general Facebook user is getting used to that.

When the offer isn’t as crazy, it’s all up to whether you’ve done a good enough job showing the value of your product. Now, for novel products or products that get amazing results, this is easier. I, however, sell commodity products. Yes, they are different (slightly) and they are improved upon, but they are still in the commodity market. Doesn’t matter how many bells and whistles you add to a garlic press, in the end, you are still selling a garlic press.

Now, that’s MY situation, and possibly many of yours as well. It just goes to show that everyone will have their own, unique challenges to overcome in their niche. How you do that is by….drumroll…..TESTING.

By analyzing the steps someone else took, ideally you will be able to conduct your own experiments and cut the learning curve or reduce errors.

What Have I Learned From This?

Thankfully, two and a half months of non-stop testing has yielded some consistent results we might be able to draw conclusions from. First of all, in all tests, it appears that video ads perform better than images. One thing to note, however, it is good to create a “meme” style video, with text on the bottom and top, but this will get flagged and suppressed, so you need an image as the thumbnail.

Next, Lookalike audiences built from Bridge audiences always outperform interest-based audiences. And the longer you use an audience, the smarter Facebook gets and the more awesome your results.

In all cases, it is important to split test. There are three elements you need to split test:

  1. Media (images/video)
  2. Text (the main text of the ad)
  3. Title (the headline that appears at the bottom near the button call to action)

After you dial those in, you’ll have a winning ad formula. Then, you can test new audiences or refine your current audience if you like, to get your targeting even more laser focused.

Finally, as I mentioned before, the most important thing is the offer. It has to either be a stellar offer, or be perceived as one (because you found a way to convey the value of your product properly).

Those are my Facebook ad test results, the beautiful and abysmal. If you enjoyed us learning together, hopefully I’ll see you again on the journey.

Tools of the Trade

A quick note on what tools I used to facilitate these tests.

First, I use Manychat as my chatbot of choice. I deliver coupons and deals (and take advantage of the awesome Heatseeker URL) by using Code Courier by SixLeaf.

Next I use Bridge to create my seed audiences. From a highly targeted and filtered audience of past customers, I create a Lookalike audience.