UpriseUp - Up
UpriseUp - Rise
UpriseUp - Up
Back to EventsBack to Blog

Facebook Ads: How to reduce CPAs

Reducing CPAs for Facebook Ads - Blog Lead Image

How can charities improve Facebook Ads performance

In our last blog we discussed our predictions for paid media in 2023, which included a bleak outlook of the Facebook Ads landscape, with CPAs continuing to rise. However, it’s not all doom and gloom.

The most significant reason for rising CPAs is due to a general decline in third-party cookies, especially following the iOS 14 changes, whereby loss of data has resulted in less precise targeting and worse machine learning.

We’ve therefore shared some of our best tips and tricks you can start employing in your own campaigns, with the aim of reducing the cost of your Facebook Ads whilst actually improving results. We’ll touch on both settings and strategies to utilise on the platform which largely aim to help improve Meta’s own AI, by playing to its strengths.

Below are some questions for you to consider about your current Facebook Ad campaigns.

Are you maximising opportunities with automation?

As third-party cookies have declined, so has performance on Facebook Ads, meaning advertisers have reduced their reliance on the platform. Meta have since invested heavily into improving their AI, and developing their automation offerings – as seen by the frequent new automation features (labelled advantage+) in 2022

The general theory is, the more you resist the way Meta wants you to run your campaigns, the more your ads will be penalised. So these features have to be considered. They should be actively considered in your strategy, and not simply pushed aside. 

That being said, you don’t need to adopt every single feature. Take the time to understand what each release can do, and whether it really is going to limit your campaign overall. Meta push the importance of the Power 5, so we would recommend starting here: Graphic on how to reduce CPAs

This suite of tools includes:

  • Auto advanced matching (this can be toggled in Events Manager)
  • Campaign budget optimization
  • Automatic placements
  • Dynamic ads
  • Simplified account structure

While the last point isn’t exactly a feature, Facebook advise against splitting out into multiple campaigns and ad sets, where the objective is the same. This approach allows for greater data pools per campaign/ad set, which improves machine learning, by reducing the ‘learning phase’. Exactly how broad or granular can be dependent on the situation and is something we highly recommend still testing. We will discuss this in more detail later though.

Can you pivot the campaign to utilise lead-gen forms?

The loss of data following the iOS14 changes has impacted Meta’s ability to effectively optimise. At least for click or conversion-optimised campaigns that rely on the pixel. Using a campaign format like lead-generation, can help minimise this impact.

By keeping your users and data on the platform, their data will not be subjected to cookie-policies. This greater quality data pool can improve the machine learning as well as build better first-party audiences. 

Lead focussed generation strategies Facebook ads

Lead-generation focussed strategies do still need to be relevant to the objective, and rely on a good email strategy. For charities, we have found gift in will or legacy campaigns to work particularly well using this format.

Are you leveraging first party data where possible?

Remarketing or re-engagement is an extremely salient component of paid-social advertising. It’s vital to ensure these are well set-up and appropriately used. This includes audiences built on the platform, via the pixel or conversion API, as well as customer upload files from a CRM. This could be website visitors, Facebook engagers or previous sign-ups for example.

Using high quality, first-party data to create lookalikes is still a great place to start when prospecting and scaling ad campaigns.

Whilst ensuring your pixel is set-up to capture first-party data, it’s worth mentioning Aggregated Event Measurement. This is Meta’s protocol designed to allow for the measurement of events from (and improve delivery and reporting of) users on iOS14.5+ devices. If you haven’t already, it’s highly recommended you verify your domain and prioritise your AEM Events.

Are your audiences too granular?

We have historically adopted a highly granular and insight focussed approach, however, this is becoming increasingly challenging. Given the trend over the last year or so, we expect audience interests to be further decimated, and trying to target granularly, will start to seriously limit your reach and ability to effectively optimise, and may not even be possible. Allowing machine learning to tap into a wider pool of users speeds up the learning phase and improves results faster.

Sure, we’re moving away from singular interest ad sets and multiple lookalike segments. But we still like to keep some level of granularity within our ad sets, where creative can be tailored (e.g. for remarketing).

This is not a fixed rule though. We highly recommend testing and will continue to do so ourselves. Using entirely broad or run of network audiences will be more applicable in certain cases. One example could be if you are simply trying to reach a niche audience but relevant interests aren’t available. Another case could be you have large budgets available and need to scale. Or, you may have limited first-party data to create lookalikes.

Are you utilising all creative formats?

Different users need different messages, and may engage with content very differently depending on their format and placement. It’s key to ensure you’re offering Meta the greatest chance to test, learn and optimise messages for the right user.

This extends to all standard formats (single images, carousels and video). You should also ensure you have placement-specific sizes, in particular for stories and reels. This is essential for optimal user engagement, which can ultimately impact how competitive you can be in an auction environment. If your CTR suffers, so does your CPC/CPM, and thus raising your costs.

Examples of multiple creatives for Facebook Ads

Are you reusing the same creative? 

We can’t stress enough the importance of ensuring the creative concept and assets designed are high quality. Your technical set-up and Facebook strategy is only as good as the ad being run.  

The best creatives are bespoke, stand-out and work in harmony with the ad as a whole and landing page. Poor quality and over-used creatives will receive poor engagement and increase overall costs. Similarly, if your fundraising message simply doesn’t resonate with your audience, you will be fighting a losing battle. We’re happy to recommend additional creative agencies who can help, so please do reach out to us.

Have you tried alternative attribution settings?

In response to limited data after the iOS14 changes, Meta also updated how their reporting worked. This included changes to attribution and conversion windows, which determine whether a conversion is recorded depending on how soon after a user clicked or viewed your ad.

Attribution windows were reduced across the board, from 28-day click +1-day view as the default to 7-day click + 1-day view attribution. This means a conversion will be counted (and used for optimisation) if it falls either within 7-day after the ad was clicked, or 1 day after the ad was seen. However, due to reporting delays, the inclusion of the 1-day view makes use of statistical-modelling, which doesn’t always benefit Meta’s optimisation due to lower quality data being fed back.

Testing click attribution for Facebook Ads

There is a growing argument to suggest moving to a 7-day click only attribution setting, as this can drive higher volumes of quality traffic, and benefit performance in the long-run. This is especially effective when the consideration period for the service being advertised e.g. a donation, can be short, and the post-click performance (checking GA data) is fairly strong.

Our above recommendations focus on in-platform changes. Yet, it’s worth noting that Facebook Ad performance is dependent on other factors, such as landing pages and general website. Considering your SEO, CRO and UX is extremely important (and are all services we’re keen to support on!) although that’s a post for another day!

Want to optimise your Facebook Ads?

We’d recommend testing our above suggestions to see what works best for your specific charity. Although, it might not be long until Meta make many of the automation-based suggestions compulsory anyway…

If you’d like support to make the most out of your Paid Media budget for Facebook Ads, then we’d love to help you! You can contact us or send us an email at hello@upriseup.co.uk. We can’t wait to hear from you.

    Did you enjoy this blog post?

    Share this article:

    We use Mailchimp as our marketing platform. By clicking below to subscribe, you acknowledge that your information will be transferred to Mailchimp for processing. Learn more about Mailchimp's privacy practices here.

    Contact us