$10M+ paid media managed over 8 years
Founder of Connected Brands, a performance marketing agency
Author at The Conversion Lift
If you’re a performance marketer, you know how much impact good ad creative can have on your campaign’s success.
Some would argue it’s the most crucial element in any campaign.
However, you have also likely faced Facebook’s not-so-intuitive interface, which can often make it hard to understand which creatives are working and which ones aren’t. If you’re working with dynamic ad creatives, that can make it even harder.
So, how do we efficiently find our top-performing assets?
Most often, we’ll end up with something like the image below.
If this looks familiar to you, then let me show you a better and more intuitive alternative.
Let’s get to it.
3 Steps to Get Started With Polymer Search
Think of Polymer as a super-powered pivot table that leverages AI for automated insights.
You can upload datasets from Google Sheets, Excel, Airtable, and others, and Polymer will automatically build an interactive web app that you can use to discover top-performing segments in your data set.
In short, Polymer lets you analyze which creatives perform well within different segments and gather meaningful insights for our campaigns quickly and easily.
1. Export Your Facebook Ads Data into a Spreadsheet
To start, you’ll need first to download your data from Facebook Ads into a spreadsheet.
Make sure to include the most important metrics you want to analyze and add as many breakdowns as possible: placements, demographics, headlines, etc.
2. Edit Your Columns for More Granularity
In the next step, add some new columns to your spreadsheets (using Excel, Google Sheets, or whatever you prefer) to have more segments for Polymer to analyze.
I typically tend to keep solid naming conventions across my accounts. For instance, I’ll label campaigns by “Stage” and label ads by funnel (i.e., “Top of Funnel”). With this information, you can edit your spreadsheet and add new columns for Polymer to read.
Other ideas of custom columns you can add to your spreadsheet include:
- Creative Type: Images, Videos, Carousels, etc.
- Audience Type: Interests, Lookalikes, Remarketing
- Video Length: 15s, 30s
- Landing Page: variant A, variant B
- Creative Type: UGC, Testimonial, Endorsements
- Any other segment you like
Organizing your data this way may take a bit of time, but it’s totally worth it. Solid naming conventions are a must (i.e., naming ads like “remarketing-image-lifestyle-testimonial”).
If you follow consistent naming conventions, you can simply split your ad names into columns and analyze your data much faster.
3. Import Your Spreadsheet to Polymer Search
Then import your spreadsheet into Polymer Search.
Once you’ve created your account, click on the “Add New” button in the left sidebar. Choose your spreadsheet and then click on “Launch Site”.
How I Use Polymer Search to Find Top-Performing Creative Assets
There are many different use-case scenarios for performance marketers using Polymer Search, but I’ll share a quick example of how I found some interesting insights in one campaign.
Identifying Low-Hanging Fruit with the Auto-Explainer
Since I added some different custom columns to my spreadsheet, I now have many segments to work within Polymer Search. To start, I’ll add “Purchase Conversion Value” and “Amount Spent” as metrics to calculate ROAS.
I’ll then add two segments: “Stage” and “Media Type”.
In this scenario, I was looking to hit a 2.3x ROAS in this account. To achieve this goal, I had to cut some losses before anything else.
As you can see below, not only were images on top of the funnel campaigns performing poorly, but they were also responsible for +60% of the accounts spend.
Let’s take a closer look and see if we can find out more.
Add More Segments to Break Down the Data Further
To look into it further, I’ll click on the “Images” and “Top of Funnel” labels so I can filter these results. I’ll then add another segment: “Creative Type”.
Now, we can see that not all images are performing poorly. Some of the most recent “Lifestyle” and “Features” images I had recently added were off to a good start!
So, what else can we do with this data?
Once again, I’ll click on the “Product” label to filter for these results and add the “Ad Name” segment.
Curiously enough, among all available images, Facebook seems to have invested +85% in one single (underperforming) image.
I can now head back to Facebook and pause this ad so that the other images can get some delivery, too. And that’s it.
In under 5 minutes, I was able to identify one poor performer that had a significant investment behind it. You can do the same to determine which assets are performing well and under which segments and use this information to design new creatives.
This is just the tip of the iceberg regarding fantastic insights you can quickly uncover with Polymer. If you have any questions or suggestions, feel free to email us at firstname.lastname@example.org.