Facebook Campaign Structure Best Practices

Facebook ads have rapidly evolved. Privacy restrictions and advances in machine learning have made the ‘go-to’ strategies of old, now completely obsolete.

This is written alongside Robbie Miller, I trust him more than anyone else in Facebook ads. He is our head media buying coach at Daily Mentor. If you are doing over 7 figures, apply to join and come work with us.Most agencies you’d hire haven’t caught up. The truth is - managing Facebook has never been as easy as it is today.Any founder can save on agency fees and effectively manage their ad account in no more than 2-3 hours a week - as long as they’re running the stable, scalable account structure industry leaders rely on in 2023.Before we get into the structure - I am launching a 100% free Shopify course. It will rival your favourite gurus $10,000 pdf.

Click below to sign up:

Join Free Course

‍

‍

‍Alright... here is the structures:

‍

Key Account Principles

‍

1. Control and Test Variables

‍

In order to improve results over time on Facebook, we must have a structured testing protocol. Every structured test contains a winning “Control Variable” and a new “Test Variable.” In most cases, this is done in the context of creative testing. We must establish a “Control Variable” of our current best performing ads, then launch “Test Variables” (new ads) against this Control. This constant testing process gradually improves upon our “Control Variable” overtime and creates incremental improvement test after test.

‍

2. Consolidation & Simplification

‍

‍

Facebook’s “learning phase” and the bulk of its optimization happens on the adset level. The fewer adsets running at a time, the more data each adset has to optimize with. This leads to significantly smarter targeting and better performance versus segmented accounts.

‍

‍

However, we must segment campaigns per product/offer and per country in most cases. Every product will have varying Cost of Goods Sold, Average Order Values, and Customer Lifetime Values. Typically, there will also be significant differences in these metrics between countries as well. Because of the variance in total profit margin per sale/per customer between each product and each country, our target Acquisition Cost (cost per purchase) must also vary between products and countries as well. Therefore, we must have each product/country isolated within their own campaigns in order to maintain manual control of budget allocation and achieve each respective Acquisition Cost target.

‍

Common Mistakes

‍

1. TOF/MOF/BOF Segmentation

‍

In years prior, it was common practice to segment out campaigns and assets by part of the funnel. Today, this practice now delivers significantly worse results than running ‘All of Funnel’ because of numerous privacy/tracking crackdowns and advances in Facebook’s machine learning. Now, we should run ‘All of Funnel’ in our adsets, meaning no exclusions to mid or bottom funnel audiences.

‍

‍

When running ‘All of Funnel,’ Facebook will retarget automatically for us to deliver the best performance. We can visually see this retargeting occurring through the ‘Frequency’ column. Frequency is the average number of times someone has seen an ad from our brand within your selected time period. Lower frequencies mean that a certain ad/adset/campaign is driving a greater portion of prospecting impressions. Higher frequencies mean a greater portion of retargeting impressions.‍Note: Frequencies can be skewed based on the volume of spend. Campaigns/ads with greater amounts of spend will have higher frequencies.

‍

2. Detailed Targeting

‍

‍

Contrary to common practice, we should not run interest groups or lookalike audiences in any of our adsets (in the vast majority of cases). The only targeting we should be utilizing is “Broad Targeting.” Broad is not an adjective describing how large your audience size is, Broad is a noun. “Broad” means no interest groups, no lookalikes, and no retargeting audiences. We are leaving the targeting completely open for Facebook to control.

By not running multiple audiences, we are consolidating our adsets (making each adset ‘smarter’), we avoid bidding against ourselves as audiences compete, and we avoid audience fatigue - keeping our ad account significantly more stable. Today on Facebook, the ads themselves do the targeting. Each ad creates its own functional “lookalike audience” who it will appear to. Adset targeting now only acts as a restriction to Facebook’s targeting ability and hinders us from reaching ideal customers outside of that target audience.

‍

‍

{newsletter-cta}

‍

We should consolidate our account by running ‘Broad’ ‘All of Funnel’ in all of our adsets for optimal performance.

‍

3. No Structured Creative Testing

‍

‍

It is a common mistake for brands to take a "shotgun" approach by creating new ads, then launching them into new scaling campaigns repetitively. This is a very large mistake that is detrimental to performance. Every piece of new ad creative should be tested through a structured process. Ads that “win” in this testing process should then be moved into a winning “Control Variable” to scale. These campaigns are evergreen and should never be turned off.

‍

‍

Copy/Paste Campaign Structure

‍

Structure 1: ideal for most brands

‍

‍

• 1 CBO campaign (aka Advantage Campaign Budget)
• Run your best winning ads inside of the ‘Tier 1 Control adset.’
• Launch new creative test adsets inside the same campaign.
• If a new test earns spend and is more efficient than a current winning ad after 10-50 purchases, extract the best winning ad from that test and move it into the Tier 1 Control adset.
• Kill the testing adset once the ad starts performing inside the Tier 1 Control adset, then launch a new test.

‍For high scale accounts:
• Once extremely dominant winning ads are found that are absorbing so much spend that no new tests are able to earn budget, create an additional ‘Control campaign.’
• This campaign holds the ‘best of the best’ winning ads that dominate spend inside of our Hero Campaign.

Step 1: Test creatives in ‘Testing Adsets’Step 2: Move winning test ads into the ‘Tier 1 Control adset.’Step 3: Move winning Tier 1 Control ads into the ‘Tier 2 Control adset’ (for high scale accounts only)

Control Campaign can be either an Advantage+ campaign or a Standard campaign - needs to be tested per account

‍Adset settings:
• Website or Website+Shop destination (needs to be tested per account)• Purchase conversion event• Maximize number of conversions or maximize value of conversions (needs to be tested per account)• Dynamic creative (only in testing adsets)• Broad targeting• Exclude previous purchasers• Automatic placementsStructure 2: ideal for very high-creative output brands who want manual control over new creative tests (currently used by The Oodie)

‍

‍

• 2 Campaigns (adset budgets)
• Run your best winning ads inside the ‘Tier 1 Control Campaign.’
• Launch new creative tests inside the ‘Testing Campaign.’
• If a new test wins, extract the best winning ad from that test and move it into the ‘Tier 1 Control Campaign.’
• Kill the testing adset once the ad starts performing inside the ‘Tier 1 Control Campaign’ (optional).

‍For high scale accounts:
• Once extremely dominant winning ads are found inside ‘Tier 1 Control Campaign’ that are absorbing so much spend that no new ads are able to earn budget, create an additional ‘Tier 2 Control campaign.’
• This ‘Tier 2 Control Campaign’ holds the ‘best of the best’ winning ads that dominate spend inside of our Tier 1 Campaign.

‍

‍

Step 1: Test creatives in ‘Testing Campaign’Step 2: Move winning test ads into the ‘Tier 1 Control Campaign.’Step 3: Move winning Tier 1 Control ads into the ‘Tier 2 Control Campaign’ (for high scale accounts only)

‍Control Campaigns can be either Advantage+ campaigns or Standard campaigns - needs to be tested per account‍‍Adset settings:• Website or Website+Shop destination (needs to be tested per account)

• Purchase conversion event• Maximize number of conversions or Maximize value of conversions (needs to be tested per account)• Dynamic creative (Only in testing adsets)• Broad targeting• Exclude previous purchasers• Automatic placements

‍

Considerations

‍

1. Group Creative Tests

‍

Each creative testing adset should hold 2-4 variations of creative. These creatives are grouped by “Concept.” A concept is how we group and categorize creatives. Make multiple variations of creative for each test and group them by how they are similar.

‍

2. Dynamic Creative

‍

‍

Dynamic creative is a tool that allows us to test multiple creatives, primary texts (copies), and headlines together within a single test. Facebook will test all possible combinations of these elements to find the optimal ad. Do not exceed 2-4 creatives, 2-3 primary texts, and 2-3 headlines per dynamic creative test.

‍

‍

3. Post IDs

‍

‍

Each ad that is launched in Facebook has it’s own ‘Post ID.’ Facebook stores an ad’s engagement and optimization on the ‘Post ID’ of that ad. It is essential that when moving an ad from Testing to Control, we are using the Post ID of the ad itself and not rebuilding that winning ad from scratch.

‍

4. Attribution Windows

‍

‍

An attribution window is how Facebook credits and optimizes around sales.
‍
• 7 day click, 1 day view (7dc1dv) - Facebook credits and optimizes around a sale if that sale occurred within 7 days of clicking your ad, or 1 day of viewing your ad.
• 7 day click (7dc) - Facebook credits and optimizes around a sale if that sale occurred within 7 days of clicking your ad.
• 1 day click, 1 day view (1dc1dv) - Facebook credits and optimizes around a sale if that sale occurred within 1 days of clicking your ad, or 1 day of viewing your ad.
• 1 day click (1dc) - Facebook credits and optimizes around a sale if that sale occurred within 1 day of clicking your ad.

The larger the attribution window, the more data Facebook will credit and optimize around - however this data is ‘lower quality.’ The smaller the attribution window, the less data Facebook credits and optimizes with - but this data is ‘higher quality.’

‍

‍

In order to exit Facebook’s learning phase we must have 50 sales within that attribution window per adset. Because smaller attribution windows contain significantly less data and are a much shorter time frame, only high scale brands have the volume of spend to use these effectively. For most brands, select a 7 day click 1 day view attribution window (default). If noticing large inaccuracies due to view-through attribution, select only a 7 day click attribution window. For high scale brands, select a 1 day click attribution window.
‍

5. Custom Conversion Events

‍

‍

When running multiple products/offers, we should create a custom conversion event in our pixel. A customer that purchases a jacket is not the same customer that purchases socks. To prevent Facebook’s data from getting muddied, we should create a custom conversion event for the specific purchase of a ‘Jacket’ and another for the specific purchase of ‘Socks.’ Therefore, if running a ‘Jacket Campaign,’ instead of selecting a general purchase event as the objective on the adset level, we should select the ‘Jacket purchase event’ to optimize off of. In our ‘Socks’ campaign, we would select ‘Socks purchase event.’ This will ensure we can run multiple offers without degrading Facebook’s optimization ability.

‍

{webinar-cta}