Making use of machine learning to get more value from your Facebook campaigns

If there’ s one thing the last few months have got taught everyone, it’ s that will Facebook has a vast amount of information about its users.

Obviously, the recent scandals have solid this revelation in a distinctly detrimental light — some users really feel uncomfortable with their private data becoming shared.   A handful of large manufacturers such as Mozilla and Commerzbank took a stand against the platform more than security fears.

Yet amidst all the doom and gloom, we marketers shouldn’ t drop sight of the fact that data, when utilized appropriately and securely, is a present. It allows us to run campaigns which are both highly effective for advertisers plus relevant to users, thereby minimizing the amount of annoying or useless ads they’ re bombarded with on a daily basis.

But that rich information isn’ t solely used for concentrating on purposes. When running Facebook advertisement campaigns, you’ re presented with a number of automated optimization options which, utilized correctly, can help you to run more efficient plus effective ad campaigns.

Let’ s look at some of the methods for you to use Facebook’ s highly advanced machine learning to improve the performance of the campaigns, with proven examples of the success of the claims.

Campaign objectives

Having worked in other channels prior to social, one of the first things I noticed plus loved about the Facebook Ads Supervisor is that it forces you to select a target for your campaign.

I’ ve seen countless examples of lookup or display campaigns that are wanting to be a jack-of-all-trades, and they inevitably turn out master of none. “ I want this campaign to achieve great achieve, a high volume of clicks, and of course an optimistic return on advertising spending (ROAS). ”

Sound familiar?

All too often, advertisers can overlook that if you want to achieve multiple goals, you need multiple campaigns. But Fb doesn’ t let you forget:

When each brand new campaign is created, think carefully about what you’ re trying to achieve, and select properly from the 13 options at your disposal. In the event that you’ re trying to achieve 2 of those objectives, create two strategies.

That campaign goal isn’ t just a vanity feature. It tells the Facebook criteria who it should be putting your advertisements in front of. If you choose Video Sights, the algorithm will put your own ads in front of users within your customers that it knows are more likely to complete movie views.

At Merkle (my employer) we’ ve observed dramatic results from ensuring campaign goals accurately reflect a client’ ersus needs. We once took over working Facebook activity for a new utilities-sector client and noticed that, despite it being generously clear their main objective has been account switches, all campaigns had been using the Traffic objective.

It’ s not possible to change the campaign objective post-launch, so we chose to rebuild from scratch and switch to Conversion rate. By making that one simple change, the particular team was able to achieve a 40 % increase in account switches and remain well below target cost for each action (CPA).

Top tip: Customers Facebook deems “ clicky” are usually very expensive, and you often find that strategies with a Traffic objective can have a good unexpectedly high cost per click (CPC).

As an alternative, consider using a Conversion rates objective, but select a custom transformation higher up the funnel than the last purchase or sign-up —   e. g., a page view or even getting a quote.

This way, you’ re not entering the particular fray with all the other advertisers competing for those clicky users, but the criteria still has enough data to get people likely to come to your site.


Much more when Facebook was just Fb. They now own four from the top 10 most downloaded apps on the planet: Facebook, Messenger, Instagram and WhatsApp.

And while they’ ve yet to monetize WhatsApp being an ads platform, Messenger and Instagram are available as placements in your Fb campaigns.

They sit down alongside the Facebook Audience System (FAN) and In-Stream as substitute inventory to be taken advantage of.

It’ s clear why Fb expanded placements in this way: The more advertisements they can serve in more places, the greater money they can make. But there’ s a big advantage for marketers, too. By selecting all positions by default, Placement Optimization can occur.

This is where the Facebook formula can decide where to serve your own ad based on where it can attain the lowest cost per thousand thoughts (CPM). For a large entertainment customer, we found that using Positioning Optimization achieved a CPM twenty percent lower than when just the Fb newsfeed and right-hand column had been manually selected.

Moving outside of the Facebook newsfeed does have some risks, though.

FAN, Instant Articles and In-Stream all carry brand safety dangers. While Facebook has provided tools in order to mitigate this risk, such as blocklists and category exclusions, there may be some instances where brand safety is such a worry that a slightly higher CPM might be a small price to pay to just show on the newsfeed.

All of us also see varying degrees of achievement with different placements. Where the FAN can perform reductions in CPM with some customers, for others we’ ve seen this perform poorly and decided to take it off. I would recommend starting any new marketing campaign with all placements selected, and then frequently evaluating performance and adjusting your own strategy as needed.

Ad delivery

Along with using campaign objectives to help Fb decide to whom to serve your own ads, at the ad set degree, you can go one step more by selecting how you want to “ Optimize for ad delivery. ”

The options at your disposal is determined by the campaign objective you chosen and will allow you to further home within on exactly the right users to obtain your goals.

Lately, Value has been added to this drop down menu where you select a Conversions goal. So , instead of just optimizing your advertisements toward users who are likely to develop a conversion, Facebook can deliver advertisements to people it deems likely to create high-value purchases, thereby increasing the particular ROAS of your campaign.

We’ ve recently been testing this and have noticed some great success. Of a selection of customers testing Value against Conversions, all of us found that, on average, optimizing towards Value improved ROAS by 110 percent!

However , there was clearly an exception: one large retail customer saw that their ROAS has been 40 percent lower optimizing towards Value, and we decided to switch back to Conversions.

But all that teaches us is that there are few hard and fast rules with regards to running Facebook campaigns. Testing is totally vital, and Facebook’ s split-testing feature allows you to precisely test one optimization option towards another to decide which is best for your own campaign.

Final thoughts

Machine learning and software are hot topics in the industry, yet as with many hot topics, it could be hard to see whether there’ h substance behind the hype. Could it be actually useful, or just a parole?

Facebook’ s combination of groundbreaking ad products and wealthy user data means that it’ ersus actually worth trusting them to generate decisions. They know a huge amount concerning the people you’ re trying to focus on, more than you could ever hope to find out with all the market research in the world.

So , test out as many of these choices as you can, making sure to remember that what realy works for one campaign may not work for another which you should be constantly evaluating performance.

Opinions portrayed in this article are those of the guest writer and not necessarily Marketing Land. Staff members authors are listed here .

About The Author

Laura Collins is Paid Social Movie director at UK-based paid media company, Merkle|Periscopix. In her five yrs in digital marketing, she has obtained in-depth knowledge of Facebook, Twitter, Ppc, and several other platforms. She has handled accounts across a range of sectors using a specialization in finance & store. Laura is a regular contributor in order to Marketing Land and recently talked at PPC Hero Conf Philadelphia and SMX London.

If you liked Making use of machine learning to get more value from your Facebook campaigns by Laura Collins Then you'll love Marketing Services Miami

Leave a Reply

Your email address will not be published. Required fields are marked *