Tag Archive for: facebook ad creative

Person getting ready for a jog.

Kicking Dynamic Creative Ads to the Curb: How we Decreased CPA by 66% For Joyn

Person getting ready for a jog.

Joyn represents everything positive about the future of movement. True, we’re biased, but one of the best parts of what we do is choosing who we work with. And the truth is, our jobs are much, much easier (and more fun!) when we believe in our partners’ business. So, it goes without saying: we’re big fans of Joyn. 

Simply, Joyn is a movement app for every body. Built on the conviction that feeling the joy and freedom of movement shouldn’t be exclusive to muscular influencers in size 00, Joyn’s online library includes a wide range of videos led by instructors that are positive, warm, and inclusive—truly. At the beginning of each class, the instructor introduces themselves, shares their pronouns, and takes a moment to talk through their recommended modifications to the movement they’re about to facilitate. That way, people that might need to be seated can still have fun and participate. 

When Joyn reached out to us in late 2020, they had a well-established brand, product market fit, and a growing (and super excited) audience. But what they were looking to accomplish was replicating their positive growth across multiple channels and supercharging it with a growth marketing agency like Tuff. 

The Backstory

When we jumped in and got access to Joyn’s Facebook Business Manager, there was already quite a bit of historical data accumulated from past campaigns they had been running.

Similarly to many new Tuff clients, Joyn knew that Facebook Ads were a key tactic for scaling their subscription user base, so their in-house team jumped in, whipped up some creative, and launched ads to start getting a finger on the pulse of which combination of targeting and creative would drive the most conversions on the site. 

Truthfully, for any startup seeking product-market fit, this is the perfect approach. Get scrappy, launch some ads, glean some learnings, and when you’re established and ready to scale, call in more resources. 

When we stepped in, we did it with a pointed goal: drive down CPS. We paired with a clear game plan:

  1. Dig into the historical data 
  2. Test Non-Dynamic Creative 
  3. Get UGC Influencer Style Creative 
  4. Optimize what performs, ditch the rest

Joyn’s Facebook Dynamic Creative Ads: Were they Working?

After pulling and organizing the historical data, we realized that Joyn was relying heavily on dynamic creative ads, giving us an excellent place to jump in and uncover more insights. 

Dynamic ads require the Facebook strategist (or whomever is executing the ads strategy) to jump into the platform and upload several different types of creative along with several different headlines and body copy. Then, when the ad is published, Facebook—using its algorithm—automatically tests different combinations, eventually prioritizing the combinations that are most effective (“effectiveness” is measured based on whether you’ve chosen to run a conversion, traffic, video views, reach, brand awareness, or app install campaign). While generally, marketers can see some positive results using this approach, there are some significant drawbacks:

  • It’s a challenge to drop in creative that’s going to be cohesive no matter what combination Facebook serves
  • It’s relatively challenging to optimize on the fly with dynamic creative campaigns

So, we decided to take matters into our own hands and launch non-dynamic ads. The results speak for themselves. 

Launching Non-Dynamic Ads: a 66% Decrease in CPS

Our main goal when we launched non-dynamic ads was to first optimize spend toward the best performing asset. Although typically when we pivot to test new Facebook strategies (whether it’s a new audience, new bid type, or new ad creative) results are far from immediate—especially given Facebook’s seven-day attribution window.

But, after just $545 of spend, we saw a sharp drop of 62% and within the very first week we saw a 66% decrease in CPS. 

  • Dynamic Spend: $6,583.44 | 153 Start Trials | CPS: $43.03
  • Non-Dynamic Spend: $1,744.40 | 121 Start Trials | CPS: $14.41

faceboook cps decrease chat

True, non-dynamic ads aren’t for everyone. They necessitate a much closer eye and the oversight of someone that can spend time reallocating budget to best performing assets and manually testing creative combinations frequently. When done right, though, the results speak for themselves. 

Fresh Creative: Tapping Into Influencers

Next up on our agenda for Joyn was to deep dive into their creative assets and emerge with…

  1. A full assessment of the creative that’s historically been performing at the top of the pack
  2. Clear ideas for new types of creative we’d like to test

Here’s a peek at what that looked like.

We noted that the strongest-performing creative tended to be shorter videos that open with high energy and/or full-screen movement. Both of our best-performers featured a modified way to access the movement, and bright colors with quick, varying shots. Finally, we were immediately able to see recognizable Joyn branding. 

Strength Training Video  |   Yoga Video

So, the next steps: recommending fresh creative. Joyn’s library of body-positive movement classes is populated by a cohort of inclusive, positive, extremely personable coaches. So when we recommended testing influencer content, we were able to create fun, big-energy new creative on a few day turnaround without having to source or negotiate with influencers. 

We were able to get two raw videos back from two of Joyn’s most memorable coaches, Kanoa Greene and Anna Chapman, use Joyn’s internal team for some extra editing and text overlay, and deploy them without a hitch. 

While there are a few extra steps to take (both on the Joyn Instagram page and on the influencer’s personal page) once we began promoting the two videos, the results were extremely interesting.

  • Kanoa Greene Influencer Campaign: $643.18 | 14 Start Trials | CPS: $45
  • Best-Performing Audience: “Female Leaders” — CPS: $14.81

The name of the game for Joyn: drive a CPS of less than $20. And while our results for Anna’s influencer campaign were not quite as tight, serving Kanoa’s ad to fresh audiences, rotating in new copy, and testing new placements have us seeing results that are closer to what we’re aiming for. 

Ready to See What Works Best For You on Social?

While we have been able to deep dive into both creative strategy and rigorous optimizations with Joyn, no two brands are built the same. Have a suite of creative you think we could supercharge? 

Let us take a deep dive into your brand and develop a strategy built for your business. 

Download a Sample Growth Marketing Proposal

Facebook Historical Analysis 101: The How & The Why

A laptop, cordless mouse, yellow glass, pen, and camera on a wooden table.

So you’ve just signed on a growth marketing team to help you run your Facebook Ads. Now what? Before your social ads strategist can jump in and create a new strategy for your Facebook ads, there’s a crucial first step they’ll need to take to ensure that you’re set up for success.

Keep on reading for a sneak peek into how and why we conduct a historical analysis on all of our new partners’ Facebook Ad Accounts. 👇

What is a Facebook Ad Account Historical Analysis?

If you’ve ever run any social ads, it’s crucial to take a look at previous campaigns and analyze how they’ve performed. Even if you found your previous Facebook campaigns unsuccessful, there is always something we can learn from the data. 

At Tuff, a historical ad account analysis is when our channel experts sift through all of the ad campaigns that you’ve ever run on your social media accounts as well as the data from campaigns in your Google Analytics account (given that you were tracking website behavior from your social ads). This historical data allows us to quickly identify a baseline of what success has looked like in the past with social ads for your brand. During this process we will also identify the low-hanging fruit that we can quickly implement to optimize your Facebook results and see improvements right away!

Conducting a Historical Facebook Ad Analysis 101

At Tuff, we use two tools to analyze the results from your previous Facebook and Instagram ad campaigns: Facebook Ads Manager and Google Analytics. 

First things first, we’ll create a custom report in Facebook Ads manager analyzing the lifetime data from your campaigns. To do so, you’ll…

  1. Go to Facebook Ads Manager and click “Create Custom Report” (find it on the right under the drop down menu “Reports”) 
  2. Toggle the time frame to include all lifetime data. 
  3. View your custom report which should look something like this. 👇

A screenshot of a custom report in Facebook Ads Manager

1. Facebook & Google Analytics Results

When it comes to identifying key metrics, there’s no one size fits all. You might be an ecommerce brand focused on bottom line revenue, an app or service focused on application installs/subscriptions or a B2B company focused on leads—every brand has their own way of measuring the success of Facebook ads.

However, we do have our own process at Tuff, so in addition to your main business KPIs, we’ll recommend keeping an eye on the metrics below to determine what’s performing well and what’s not.

In Facebook Ads manager, we’ll look at: 

  • CPA/CAC – Cost Per Acquisition (purchases, leads, subscriptions, app installs, etc.)
  • Link Clicks & Landing Page Views
  • CTR (Click-Through Rate)
  • CPC (Cost Per Click) 
  • CPM (Cost Per Thousand Impressions)
  • ROAS (Return on Ad Spend)
  • Video Engagement 

Then, we’ll compare the Facebook data to the data in Google Ads Manager, where we’ll look at your website behavior. We’ll look at: 

  • Last-Click Goals/Conversions (purchases, leads, subscriptions, app installs, etc.)
  • Conversion Rates
  • Average Session Duration 
  • Bounce Rate
  • Pages/Session
  • Users and New Users

2. Audience Analysis

Once we have our campaign results, it’s time to get into the meat of the historical Facebook Ad analysis—drawing insights from all the variants that have been tested to figure out what we’d like to test again and what works best moving forward. 

First, we’ll take a deep dive into the audiences you’ve tested and see which audiences are most engaged with your Facebook ads. 

When looking at the data from a historical audience analysis of our new partner Dumpling’s Facebook traffic campaigns, and saw that their top performing audiences in all of their key markets were Facebook users who fall in the top 50% household Income of their area. 

A screenshot of a historical audience analysis.

This learning gave us insight into who their audience was on Facebook, which allowed us to create other audiences targeting similar behaviors and interests as the users who fall in the top 50% household income of their area. 

3. Creative Analysis

Creative and copy play a huge role in Facebook ad performance, so it’s extremely important to know what type of images/videos and messaging resonates best with your audience. 

We recently conducted a creative analysis for Joyn and learned that videos with high-energy and big movements (like this one) outperformed their videos that were more calm, relaxed and had little movements (like this one). 

We also learned that their audience was more engaged in video ads that were created in an actual workout environment rather than video ads created in a studio with a backdrop. Since we’ve rotated in new videos that implemented the data-driven insights from our creative analysis, we’ve seen a significant decrease in Joyn’s cost per subscription from Facebook ads.  

4. Placement Analysis

There are a lot of different placements you can choose when building out Facebook Ads and oftentimes, we’ll see our partners choose automatic placements instead of analyzing which placements yield the best results with their audience.

That’s why we always analyze placements in our historical Facebook Ad analysis. For example, when we pulled the data below for our partner VIDA, we discovered that there was a big opportunity to test three new placements (highlighted in the table below) since there was an impression share available that they weren’t currently capitalizing on.

A screenshot of a Facebook placement analysis report

Using Data to Inform Facebook Ad Strategies

Our one final thought we’ll leave you with: 

  • Paid social strategies without historical data = 🥴 
  • Paid social strategies with historical data = 🤓

After we’ve created the historical Facebook Ad Account Analysis, we’ll use the data and insights to guide our strategy. The historical analysis saves us and our partners a lot of time on testing things that have already been tested in the past and allows us to focus on what’s working and how we can better optimize what’s working for growth. 

Interested in learning more about our process for social ads? Give us a shout!

Got a Crappy iPhone Video? You’ve Got the Perfect Facebook Ad

A selfie of three smiling friends on the beach

There’s no one-size-fits-all recipe for successful social ad creative. After working with 50+ brands in all kinds of industries and executing countless A/B tests, we’ve learned a thing or two about what good paid social creative looks like. 

Disclaimer: what works for one account might not work for another. Make sure you’re testing different kinds of creative to find what works best for your brand. 

Yes, every Facebook ad account has its creative differences. But when it comes to the Facebook ad creative that works the best in 2021, there are definitely some common themes. Check out our tips for developing Facebook ad creative that yields the best results.

Low Fidelity Video > Everything Else

Say “hello” to iPhone videos, and “see ya later” to hella expensive video production. On most all of our accounts, low-fidelity iPhone quality video typically outperforms the polished high-end video we’re testing against. 

Typically, brands think that developing video assets = 💸💸💸. But that isn’t the case. Shoot the video on your iPhone, edit it together in TikTok, and export it to use on other social channels. Mimic the latest video editing trends on social, and you can get scrappy with creating your own video content — and see great results.

We tested this for Felt — an app that lets you send handwritten cards right from your phone.

Note: our Facebook campaigns are optimized for app installs, and one of our primary KPIs for measuring success is cost per install (CPI). 

The CPI for ads that featured low-fidelity video assets was 20% lower than its high-fidelity video counterparts and 50% lower than image ads. 



The data above shows the average CPI for $36K in Facebook spend. 


Curious? Check out some of our ad creative for Felt. This will help give you a better idea of what we’re talking about when we say low fidelity and high fidelity video content. 

Low Fidelity Example | High Fidelity Example

Over time, we’ve gleaned a few other tidbits for editing video for social ads. Selfie-style video of someone talking about your brand works really well (this could be the founder, an employee, or an influencer you’ve partnered with) – just make sure you add subtitles! We’ve also noticed that videos with subtitles and graphics that match the in-app design features are also some of our top performing video assets. Ask yourself: What are the kids makin’ on TikTok these days? And how can I copy and paste that style for my brand in a way that makes sense? 

What are Non-Dynamic Ads?

Dynamic ads allow you to upload up to 10 images, 7 videos, 5 captions and 5 headlines, and the algorithm will pair different combinations of creative together to make what it thinks is mostly like to perform the best. 

In theory, Facebook’s dynamic ad creative option sounds too good to be true. And my grandma always told me that if something sounds too good to be true, it probably is. 

For some brands, Facebook dynamic ads can work. But if you’re currently using dynamic ad creative and seeing less-than-stellar results, try switching back to the regular non-dynamic ad creative.

Non-dynamic ads give you more control to make optimizations, and it’s much easier to translate the data into actionable insights. 

When we onboard new clients at Tuff, we often see that they’ve been running dynamic ad creative on Facebook – especially if someone in-house has been running their Facebook ad campaigns because it’s such an easy ad format to execute. 

We experienced this with two Tuff clients: Joyn and Team Boocamp. One of the first optimizations we made was to switch their Facebook campaigns from dynamic to non-dynamic ads. 

And by making that simple switch, our cost per signup for Joyn and Team Bootcamp decreased 42% and 47% respectively. 

a sharp drop in Joyn's CPS after launching non-dynamic ads

Images with Text > Lifestyle Images

You would have never, ever caught me saying this two years ago, but well, here I am. Social media advertising is ever evolving, and so is our approach to developing creative for Facebook ads. 

Creating images with text used to be one of the biggest Facebook ad creative faux pas because of the 20% text rule – AKA the bane of my existence. 

When Facebook semi-recently did away with the 20% text rule (praise be), it changed the type of image creative that works best for ads. Now, images with graphics and text overlay tend to perform better than the typical lifestyle image. 

When you add a bit of text to an image, you give yourself more real estate to get your message across. Think about a text overlay on your image as your headline, and the native text elements in Facebook ads manager as supporting copy. 

We put this to the test for Offline – a restaurant subscription service in North Carolina. We found that graphics yielded an 48% lower cost per signup (CPS) compared to regular images with no text overlay. 

Aside from CPS, we saw that graphics had a stronger CTR, lower CPC, and drove healthier traffic to the site – doubling the time on site (TOS) compared to regular image ads. 

a chart that shows the decrease in price and increase in efficiency between images with graphics and images without


Interested in learning more about our A/B testing methods for social ads? Give us a shout