English Español Deutsch Français Italiano Português (Brasil) Русский 中文 日本語

Advanced PPC #6: Level Up your Facebook Advertising Strategy




Joel: I'm here with Aaron, and also Karl. Aaron's going to be speaking about some exciting new Facebook advertising techniques, which I think will help us all perform better. Really, Facebook advertising, there are so many questions about how to do things. You could read all the documentation that you see about it, you can study up on it.

At the end of the day, what makes you perform properly, isn't just knowing what the rules are. It's about understanding how it behaves. It's about knowing if you flip this on or turn that off, what the outcome is going to be.

Facebook ads; they just make it more and more confusing. To me, it's not as complex to use as Google in the aspect that you don't have as many ways to drill down, as many options, as many retargeting techniques. You can segment reports in so many levels. 

You can't bid adjust on all these different elements like you can with Google. But for some reason, it's just like the rules always seem to change on it. You tend to do one thing one day and it doesn't work the next and you just don't know why. It'll be really good to hear what this group of experts we have has to say regarding that.

Aaron is an Indiana University graduate from Indianapolis, which is where he's currently based. Now, he spends his time managing his PPC campaigns for Hanapin Marketing. It's a very large marketing agency, I think one of the most well-known also based in Indiana. 

Joining us we have Karl. He is a digital marketing specialist for Booster Box. Booster Box is a very excellent well performing online marketing agency based in Tuscany, Italy. He's a Facebook advertising specialist there and he's got experience managing campaigns with various degrees of spend. 

I think at this point, I'll hand over the keys to you, Aaron, you can go ahead. If you're ready, share your presentation, and we can begin.

Aaron: All right. Like we said, we're going to be talking about some ways that we can level up our Facebook advertising today. I'm just going to be talking about how all these tools work together to really build on utilizing Facebook's automation and helping you save time in your day and just get better results from the platform overall.

Getting the Most From the Facebook Pixel

Starting with the Pixel, starting real simple, you guys probably know what the Pixel is. But just in case we have some viewers that don't, I just want to touch on it real quick. But this is going to be our key to unlocking Facebook's powerful tools. We're going to be able to use the Pixel for quite a bit of different things.


It's really going to help us track our site traffic, and how those users are interacting with our site. It's going to allow us to set up conversion events. Per conversion events, I always give a recommendation that if you have an important action on your website, no matter what it is, turn it into a conversion event.

If it's an add to cart, if it's a purchase, it's a form fill, a subscription, whatever matters to you, turn it into a conversion event, and that's going to matter a little bit later on the presentation when we start talking about conversion campaigns.

Continuing on, the Pixel is going to help us measure those conversions across devices to allow you to see the customer journey a bit more clearly. The Pixel is going to help learn who's converting on those events that you've set up so you can find more relevant users on Facebook like those users. Maybe most importantly, it's going to help us create some really powerful custom audiences.

Two of the tools on the Pixel that I see a lot of advertisers, they don't use are 1st Party Cookies and Advanced Matching. Both of these tools are available under your Pixel settings, and they're easily toggled on and off. 

1st Party Cookies and Advanced Matching both together, they're really just going to help give Facebook, the Facebook Pixel, a little bit more in-depth insight into who your users are. Learn who those users are, what they're interested in, and then they're going to help you find more users on Facebook like those that are already on your website.

But maybe more importantly, especially with Advanced Matching is, it’s really going to help you get around the issue of ad blockers or some of those web browsers that are becoming a bit more restrictive with how long they're allowing to keep data. 

Advanced Matching is going to allow us to see if someone's on your website, and they're filling out a form, if that user has an ad blocker on, the ad blocker might be blocking a Pixel event so it can't necessarily see that normally. 

With Advanced Matching on, the Pixel is going to be able to see the data that people input into your text fields and similar things on your website. It's going to be able to take that data and hash it, make it private, and send it back to Facebook to try to match that back up with a user who's on Facebook. Again, just two easy tools you can turn on that's really going to help you push the Pixels to the next level.

Conversions vs Traffic Objectives on Facebook Ads

Let's move right along into conversion campaigns. I see a lot of advertisers still using traffic campaigns. The one thing I always like to tell them is we shouldn't be just going after site traffic, we should be going after people who are going to be taking very specific action on our website.


Historically, Facebook campaigns or Facebook traffic campaigns, we're basically just telling Facebook to go after clicking users. These are people who may click on a lot of stuff on Facebook, but they're not necessarily taking any relevant action on our website. 

There's also a lot of studies Facebook has published that say that clicking users can tend to be more expensive than other users likely to take an action. Whereas with conversion campaigns, after we set up relevant events through the pixel, we can tell Facebook to not just go after someone who's clicking, but someone who's going to click, let the website completely load and are most likely to take that certain action that you've told Facebook is important to you.

Again, with the Pixel and the conversion campaigns, we're going to be able to tie conversion events back to Facebook and measure how effective they are. As people start completing those important events on your website, Facebook's going to learn what attributes those users have, and it's going to start helping you find users on Facebook who have similar attributes.

The Pixel and the algorithms are really going to learn exactly who they should be serving ads to on Facebook. That's something you're not necessarily going to get with other campaigns like traffic campaigns.

One thing to keep in mind, if you are using conversion campaigns, is that what Facebook tells us is the algorithms need about 50 conversion events per week to truly optimize. That's just them saying they need enough data to really learn who the users are that it should be going after.

If you're struggling to get 50 per week, then I would recommend that you optimize for a higher funnel event. You don't have to get 50 a week to necessarily get the results that you want. 

But if you're finding that you're struggling to get the results you want, if you're finding that you're struggling to spend your entire budget with these conversion campaigns, it may very well be that you're just not getting enough conversions on a weekly basis. 

That's where, instead of maybe optimizing for a purchase, we push it up and we optimize for an add to cart and get more users in the funnel. Or instead of optimizing for a form fill, we optimize for a landing page view just to get more people interacting with the ads and on the site.

Improve Your Facebook Ads Targeting   

Moving along to exactly who we should be targeting, making sure we're just reaching the right users with our audiences. The most important audiences for me come from using the Pixel and using custom audiences. You may be familiar with creating custom audiences.

But the most powerful ones to me are going to be either creating web users who have been on your website, your web traffic, uploading custom files... This one I see less used is creating engagement audiences off of people who have engaged with your Facebook page or your Facebook posts or your Facebook ads, and retargeting those users. All these users are really, really great for remarketing.


But whatever you do with a custom audience, always try to turn it into a lookalike audience, especially if those users are completing actions that are really important to you. Because lookalike audiences are going to be where you want to spend the majority of your budget in terms of going after an upper funnel audience or a cold audience.

Like I have there on the slide, if there's an action that matters to you on your website, turn it into an event, and then track it and create a lookalike off of it. Create a lookalike of your purchases and target those users. If you want to get more subscriptions, more email registrations, whatever matters, whatever you want more of, just make it a conversion event and turn it into a lookalike. 

If you really want to take your lookalikes to a higher level, I would definitely recommend using a customer file and adding lifetime value metrics to that of your highest value customers. If we're thinking about this in terms of purchasers, we can upload a customer file of Facebook and say, “here are all of my purchasers in the last 90 days, find other users on Facebook like them”. 

In terms of our other audience options, I'm sure a lot of people are already familiar with the core audiences that are available on Facebook. So, your demographics, your interests, your behaviors. 

If you have persona targeting, if you're a client, if you're a business, if you've got persona targeting built out of who your ideal customer is, what they do, what they look like; use that to help craft your targeting. If that doesn't exist, or that's not working and you want something else, I always recommend a really great place to start is Google Analytics. Head over to Google Analytics and start getting a better picture of who your best users are.

Look at what ages are providing the most revenue, what genders are converting the best. Take a look under the interest categories to see things like affinity categories and in-market segments. Take notes on all that stuff, all the important stuff, all the high traffic stuff, and start plugging that into Facebook's tool Audience Insights.

What is Facebook Campaign Budget Optimization?

Moving towards some of the tools that Facebook has put out semi-recently. One is campaign budget optimization. The goal of campaign budget optimization is going to help us get the most results for our current budget.


It's basically just an automated way to budget for our ad sets. Instead of budgeting at the ad set level, we are moving those budgets up to the campaign level, and we're allowing Facebook's algorithms to allocate that one single budget across ad sets to the ad sets that are getting us the most results most efficiently. 

In the example, before campaign budget optimization, we're spending $10 a day on three ad sets and getting 10 conversions. Maybe we just don't have enough time to go in there and shift around budgets to really looking at results and where we should be spending. But with campaign budget optimization on, it does that for us. It sees that ad set two we're underspending there and overspending on other ad sets.

It's going to see that and shift budget in real time. What I've seen how this works really, really well other than just getting us better results because of automated budgeting is it opens up so much time for you to focus on other things. 

If you have a really big Facebook account that you're advertising in, if you have to manage several platforms, Google Ads, Bing Ads, Yahoo, Facebook, you've got a lot of stuff to do in your day. A lot of times, it's great when you can automate these little things to open up more time to focus on strategy, bigger picture, and maybe testing some creative or other things on Facebook that you just don't have time to.

The other important part of this is that Facebook is actually getting rid of ad set budgets this year in September. If you haven't been acquainted with this tool, or haven't gotten a good process in place for it yet, I heavily recommend you start getting used to it now so you're not rushing at the last minute trying to figure out what the best setup for yourself is.

Using campaign budget optimization, I've come across some best practices that I think work really well to help you get the most out of this tool. Every ad set in your campaign needs to be optimized for this same event. If that's a purchase, if that's a form fill, no matter how many ad sets you have in that campaign, that should be all optimized for the same event.

Always segment prospecting and remarketing audiences into separate campaigns. Make sure that your ads, ad types, landing pages are similar across ad sets.

An example of why this really matters is if we go to that third point and think about prospecting remarketing. Eventually, if we have both these audiences in a campaign and they're optimized for a purchase, Facebook isn't able to differentiate necessarily between audiences or what your goals are for those audiences. It just sees events happening. 

It's going to see the remarketing campaign is converting at a better rate, at a cheaper rate, and it's going to start funneling all your budget into that remarketing audience, which is going to be okay right out of the gate. 

But if you keep doing that, and you're not spending on your prospecting audiences anymore, you're just killing your funnel, and you're not getting users to being pushed down to that remarketing audience. Your remarketing audience; it'll eventually dry up and you'll stop getting the results.

Facebook Ads Bidding Tips

Moving on to bidding. You've probably seen this window within your ad sets. A lot of times I see it heavily underused or misused.

I always tell people to bid based on what's important to you. What's your business's goal? If your business's goal is you have a very distinct ROAS that you have to get, you need to be doing value building and setting a ROAS goal. If you have a cost per acquisition goal, that's really important to a very specific cost. You do conversion bidding, and you set it up with a target cost.


By default, Facebook will always do the lowest cost bidding. Basically, just helping us get the most results for our current budget. But if you have a goal that you want to hit, I definitely recommend using a bit more of a specific bidding method. 

These bidding methods are really going to help us control our costs as we scale out accounts are as you scale out across several advertising platforms. It's just one less thing that you have to worry about making sure you're hitting goals. It's going to be really helpful when you use them in tandem with automated rules.

Like I said, automated rules are only available at the ad set level for now. But they will be available at the campaign level before September. Automated rules are even, I find much less utilized than bidding. The way I see this is it’s another means to take advantage of algorithms and machine learning and take the manual work out of our day. 

The automated rules got a lot of options. I use them to schedule ads a lot. I use them a lot to control frequencies, to raise and lower my budgets based on customer results or based on return on ad spend. Again, it's just another really good way to help you scale out accounts and make sure that we're always staying within the goal.

The Breakdown Effect

Into the final portion of the slide deck here. This is a bit more of a media section. We're going to be talking about what Facebook likes to call the breakdown effect. The goal of this section here is to get people to understand that we should stop over-segmenting our ad sets within our campaigns. 

To understand what the breakdown effect is, we need to understand how Facebook bids on a real-time basis. What Facebook uses is called discount bidding. What that means is that it will always, no matter what your audience is, go after the lowest cost results first, and then once those results are achieved, Facebook has to go after what's left. By logic, that's the more expensive results.


What you'll see when you run an ad set a lot is that you'll get a lot of low-cost results right out of the gate. But over time, the more time we run the ad set, the more budget we put into it, we see that cost per result rise.

If we continue spending on Facebook, even if it was great at first, our CPA is going to be actually worse than if we were to redistribute that spend. If we're doing this, if we're opting into several placements on an ad set level, the algorithms are going to be able to see where that inflection point is and automatically start shifting our budget to the next best placement in terms of efficiency.

This is just another way to look at that where, at the start, the algorithm starts testing both placements. Over time, our cost per acquisition is going to rise. But when we opt into more placements, there's an inflection point that Facebook can see. It's going to automatically shift budget to the next best placement.

In terms of thinking about this, of how I want people to stop segmenting by placement is to always remember how the average cost per result rises. If we're always segmenting by placement... if we're not opting into multiple placements, we're never going to be able to experience that inflection point automatically. We'll never be able to take advantage of Facebook's algorithms and machine learning to shift budget in real time. 

That's why as a best practice, I recommend a healthy mix of placements. Either automatic or at least four plus in order to help keep your cost per result down. That's really going to help Facebook learn what placements are the best and where to distribute the budget most efficiently. 

With automation, not even just with placements, but with automation in general, campaign budget optimization, for example, I always recommend that you evaluate results at the aggregate level.

Making sure that we're not over-segmenting or making sure we're spending on the right audiences or the right placements. That is it from my slide deck. I believe that we can go ahead and transition into some panel discussion.

Joel: The breakdown effect. You're saying that basically Facebook experiments, like A/B tests between different placements. It will start off with the one first that it thinks will have the best cost for the result. And then once it starts to exhaust it, it'll jump over to the next best?

Aaron: Yep, that's exactly how it works.

Joel: That's super-interesting. 

Karl: I think the point on placements is a really valid one, which a lot of people overlook, and for other reasons. Apart from the reasons which Aaron quite rightly put forward, we also have to bear in mind the different audience types. 

When we're splitting up Instagram, and we're splitting up Facebook, we also need to imagine and understand that sometimes it's a double audience. We have an overlap of the people that are in the Instagram ad sets and an overlap of people that are in the Facebook ad sets. Separating them can either mean that you're double serving or if you try to remove the audiences from the other, means that your audiences are really too small to function. 

Aaron: Another reason I've seen people split out by placements is that they always want to make sure that their ads are really tailored to the specific placement. With that in mind, you can still opt into multiple placements and just make sure that we're using asset customization. We're able to be in multiple placements at once but make sure that we can actually manually tell Facebook to serve a very specific ad on those placements.

Facebook Ad Conversion Measuring and Optimization

Joel: Yuri is asking, “what is the story with Facebook attribution? Is this view through conversions by default? How would you recommend measuring conversions better beside what we can see in Facebook reporting?”

Aaron: I always tell people who want to advertise on Facebook, I would say without a better term, it's a very high funnel, much higher funnel than any kind of last click attribution that you might see in Google. A lot of your results, you'll probably see our view-throughs. Again, you just got to understand how those view-throughs impact your advertising.

Maybe you don't trust the view-throughs, I think you should, if you set up your Pixel correctly, and track everything correctly, then whatever Facebook is telling you as a conversion that happened because of your ad, you have that conversion. 

You don't have to worry about it over reporting necessarily. If you really want a deeper understanding of how people move from Facebook to Google Ads to organic traffic, there's a tool you can use within Facebook actually called Attribution, where you can connect all those different platforms together. 

I believe there are just some URL appendages that you can put into the other platforms and Facebook will be able to see if someone's started on Facebook and ended on Google ads. You can see that customer journey a bit better.

Joel: Thank you very much for that. I have a question related to conversion optimization. Now, you said you need 50 conversions per week for it to be effective. Now, let's say your account is segmented, let's say you're trying many different ad sets. Does it share data across the ad sets if that was working, if you're optimizing for the same conversion?

Aaron: Yeah, I would say it's the event that matters and it's not necessarily per ad set that necessarily matters. It is the one single event that you need to get that conversion data to come through on.

Joel: If you create a new ad set, a fresh ad set, is it going to use data from other ad sets or previous ad sets, or is it starting from fresh?

Aaron: No, if you're doing a conversion campaign, it's going to be able to use that historical data of, if people are completing a purchase on your site, and you set up an ad set, optimize for a purchase, it already has that purchase data that the Pixel has used to learn. It doesn't necessarily have to learn over from scratch.

Joel: Okay, that's a very, very good to know. Karl, can you give any recommendations on how to better use the Pixel for conversion optimization?

Karl: Yeah, I think one thing about the Pixel, which people don't think about enough, is the size and the integrity of the Pixel that you want to use, especially to create lookalike audiences. The most obvious one would be to say, okay, these people bought stuff. I would really like people to buy more stuff. Go out and find more people like that.

However, if you've got a really diverse, let's say, eCommerce site, and you have shoes, to trainers, to T-shirts to caps, it would be really valuable for you to make sure that you're segmenting that data based upon the categories that you're actually providing for so that you can create lookalikes on the specific things.

I really want to create an audience so I can show them the caps. I'm not just going to create a lookalike audience on purchases, what I'm actually going to do is create a lookalike audience on everybody that bought caps specifically.

I can tell Facebook, “hey, it's these guys I'm interested in, not the people that bought the T-shirts from the last week.” Be as specific as you can there, and because the most similar the custom audience is, then the easier it is for Facebook to go and find the relevant people.

Aaron: Yeah, that's a good point. I would always to go along with that definitely, once you've segmented out is make sure the ad copy is relevant to those users. Don't just serve them a general ad if you're going to segment out like that. Make sure your ads get relevant between T-shirts and shoes as well.

What to Do When Campaigns Become Too Expensive

Joel: What do you do when it gets too expensive and the campaign stalls? Do you then recycle the campaign? Do you start targeting something new?

Aaron: Generally, what you're going to want to look at is something that you can change maybe a little bit quicker than necessarily launching a new audience. Look at your placements. Do you only have one placement or two placements? That might be why you're stalling out. 

Your reach is just not as high as it would be if you were opting into other placements, giving more to Facebook to test. Look at what you're bidding. Maybe you have a cost per acquisition bid in place, and you've given a goal of $50. If you're struggling to get results on that, that's either telling you that you need to maybe possibly raise that cost per acquisition a bit, that goal, or if you're not willing to do that, then start taking a look at how your ads are resonating with your audience. 

Are they even clicking on them? Are they engaging with them? What's your relevance score? Look at that kind of stuff too. Because you may just need to change your ad copy, change the imaging, make sure it resonates better with the user.

Karl: When you totally reach the maximum cheap CPA and now it starts to get a lot more expensive, I think campaign budget optimization is really going to help with that in the future, just automatically anyhow. One thing that you guys can do to help combat that is make sure that you separate up your lookalike audiences and the audiences that you're using correctly.

Aaron: To go along with that, I'd definitely say, with any lookalike audience, but especially as we're segmenting out by that and getting more, is that if you're doing any kind of additional targeting via interest, or demographic or behavioral, make sure we're excluding our lookalikes from those ad sets, because that's really going to become an issue with audience overlap. 

Where a lot of those users that exist in our lookalikes exist in our behavioral targeting as well. We don't want to be double serving ads, because then that skews results.

Karl: I think on that note, Aaron, you also mentioned something which is quite simple but often overlooked, and that is the Facebook engagement. Retargeting people that actually engaged with your ads. 

This is something which not enough people are doing. That's another exclusion that if you are going to use that, remember, that's something you need to exclude as well. Excluding people who already engaged with an ad.

One of the reasons we started separating things like that as well, is audience burnout is not necessarily determined entirely by audience size. It's also determined by the relevancy of the ads you're serving. 

Tools for Managing Campaigns

Joel: I have a question from Maxine River. Aaron, which tools do you use on the everyday basis for managing campaigns? 

Aaron: Like I said it in my slides, campaign budget optimization for sure. Unless you're wanting to use automated rules. I have a couple of clients of mine who we've got some automated rules set up at the ad set level. They have just given us great results. No reason to rock the boat just yet until they're compatible with campaign budget optimization.

In terms of other tools, I think Ads Managers, it's just Ads Reporting in terms of metrics, there's a lot you can look at there. I rarely use their default ones. I'll always set up custom views for different clients because different goals for different clients, different goals for your business, you can really look at what matters to you on a bird's eye view.

I think another good one that's a bit outside of Ads Manager that was recently launched was the Facebook Ads Library that Facebook just launched. It's kind of the new way to look at what ads competitors are running that used to exist on a business's Facebook page, but now it exists on its own unique URL. That's going to give you a lot of insight into what your competitors are running.

Joel: One last question. Someone is asking, "Is there a learning period for campaign budget optimization?"

Aaron: I think with anything you do on Facebook, there's always a learning period. I always say that Facebook can detect results pretty rapidly with campaign budget optimization or even with ad types. 

We may launch at the beginning, and we could probably start seeing Facebook start favoring a certain audience or certain ads within 24 hours. A lot of that has to do with your budget too. The more budget we spend there, the quicker the algorithms are going to be able to learn where to spend that budget.

Joel: Okay. That's good to know.

Karl: I was just going to say, that point is a really good point because a lot of people don't realize how much Facebook is learning about you, your company and your activity as you actually go. 

For example, if we were to take the same ad, the same text and the same headline on two separate accounts, then we're bound to get two separate CPMs. Because Facebook understands already from your past activity, how different people are reacting to your accounts. It's really important to remember that Facebook has data, and it will use it on your behalf or against you.

Joel: That's one of the reasons why sometimes you might want to open up new accounts, right? If you're trying that totally different conversion funnel or sales funnel in your website, and conversions don't mean the same thing necessarily. 

Takeaways on Leveling Up Your Facebook Advertising Strategy

We have about one minute left. How about we finish it off with a takeaway. 

Aaron: I would say, like Karl already mentioned, is that Facebook's really heavily pushing for automation on all levels. They're really wanting to force advertisers to trust their machine learning and their algorithm. 

The more you try to fight against that by segmenting out and controlling things manually, eventually, it's just going to become a headache for you. You need to learn how to use these automation tools now and use them effectively to be prepared for the future.

Karl: I think along with the same note, working with the machines. One of the things we didn't mention yet, it's quite new, and it's really simple for you guys to use if you haven't already is dynamic delivery on creatives. 

Now, you can add four, five, six different images, four, five, six different headlines, four, five, six different texts. Give the information to Facebook and let them help you find the best combination of creative text and headlines.

Joel: By the way, this is also why it's very important also to stay up to date with the latest because things are always changing. If you're not following the newsletters or whatnot, you might miss out and you might fall behind the competition because they're paying attention. Actually, even more important than that are these webinars here, where you're able to actually hear from experts themselves that are experiencing them.

It's really been a pleasure. I look forward to hosting you guys again in my next webinar with SEMrush.

Aaron: Awesome. Thanks for having us.

Joel: Thanks a lot, guys. Bye.


Check out other webinars from this series