FB Ads Manager can be confusing. This confusion can be made even worse when performance is pointing down and you aren’t entirely sure what changes will have a positive impact. Below, we will look at my top 5 favourite indicators for evaluating poor performance and how to optimize them.
A quick note: This is a repost of one of the very first The Marketer's Playbook articles! Due to so many hosting migrations, many articles have been lost, but there are a few that have been found. I will continue to re-upload any I can find, but will only send the ones I believe to be most helpful via the newsletter. For the rest, you can visit the website. Enjoy!
ROAS, or Return On Ad Spend, is a ratio of your revenue vs. spend that shows how profitable your ads are. We will be using this as our benchmark to optimize towards in this example because ROAS is the guiding light for most eCommerce brands. If you are in the SaaS/B2B space, this might look like Cost Per Lead, but the fundamental principle remains: how much did I get back for my money spent?
Of course, there is an elephant in the room – this metric (ROAS) is likely wrong. Yes, you heard me correctly. Unfortunately, with all of the issues surrounding attribution, it’s tough to understand what your ROAS actually is. The metric shown in Facebook is almost certainly wrong because Facebook no longer has the ability (not that it ever truly did) to correctly attribute sales to specific campaigns, ad sets, or ads.
For your own performance measurement, you could look at your MER (Marketing Efficiency Ratio) to determine whether or not advertising is sustainable. However, that is separate from today’s exercise.
Another method you can use for the time being is creating a new ROAS column in your dashboard that will consider FB’s modelled data. This method won’t be a fix-all forever but should help in the meantime.
WHAT AM I LOOKING FOR?
With that out of the way, let’s look at ROAS. Something to keep in mind is that ROAS at different stages of the sales funnel SHOULD look different. For example, in a typical eCommerce funnel, it’s usually okay for ROAS in prospecting to be somewhere around 1x. In contrast, ROAS in the bottom of the funnel retargeting could be around 3x at a minimum. ROAS goals depend entirely on each company’s goals, margins, etc., but this is an excellent place to start.
The reason for this is purchase intent. Those at the top of the funnel are cold and have lower purchase intent, meaning that they will likely cost more to convert right away. In contrast, those at the bottom of the funnel are higher in purchase intent, it makes it cheaper to convert them as they’ve already been exposed to the brand/product a few times.
It’s also usually okay to take a slight loss on prospecting (ROAS of around 0.8x, for example) as the initial acquisition is the most expensive part. If your lifetime value is substantial, you can make up that money over time. This won’t be true of single SKU impulse purchase shops; however, if you have a strong retention strategy, this may be a tactic you can use. That said, the closer you get to breaking even or above with a 1.0x ROAS, the better off you’ll be in prospecting.
HOW DO I OPTIMIZE THIS?
This is a broad question, unfortunately. Usually, ROAS is the first indicator that I look at in determining if a campaign/ad set/ad is working, and if it’s not meeting my goal benchmarks, I will dig into the ‘why’ further by looking at other metrics.
To figure out how to improve ROAS, we have to look at other metrics and see what issues we can uncover and resolve.
So, without further adieu, here are the top 5 steps to figure out why ROAS is low and what I do to resolve it.
STEP 1: SPENDING ON CAPPED AD SETS
While I don’t use Cost Caps/Bid Caps on all campaigns, I will use them depending on the situation. For this reason, if the campaign is not using Lowest Cost as its bidding optimization, I will look at the cap’s performance.
There are a few things to look at, but what I use as my primary indicator for cap health is how it is spending vs. ROAS. Look at how ad spend looked yesterday compared to the daily budget. Did it spend through its budget? Did it spend under its daily budget? This can mean one of two things, assuming performance is below your goal ROAS:
- Spent through budget: This means that the cap is too high, and there is room to lower it. Lowering your cap will tell the system to spend less on acquiring a customer, and if you are spending through your daily budget, this means that it’s not restrictive enough, and the cap can probably be scaled down further. Be careful not to decrease it too much and cause:
- Did not spend through budget: This means that your cap is too low, and you should increase it. If your cap is too low, it restricts the system from having enough room to optimize for purchases. When your cap is too restrictive, you aren’t giving your ad sets enough room to decipher who your ideal customer is and optimize towards that freely. In this case, you want to increase your cap to a point where it starts spending through your daily budget but not flying through it.
The key here is to find the sweet spot. Start with a cap at 200% of your Average Order Value, and slowly decrease it by 10% every day until you see it stop spending through your daily budget. Once you find that point, increase it slightly so that it spends through its daily budget while being exactly as low as it can get to spend efficiently.
It’s worth noting that caps are not a guarantee for success. As with all things in FB Ads Manager, there is no saying that it will or won’t work. Each case is unique, but it’s worth testing, especially if Lowest Cost hasn’t worked efficiently for you.
STEP 2: CPMs
CPMs are a tough one. CPMs sit right at the top of the ‘waterfall,‘ where a high CPM will cause all other metrics to be more expensive. Typically, if you see a high Cost Per Purchase, you see a high Cost Per Add To Cart, Per Click, etc.
For that reason, monitoring your CPMs is essential. The hard part here is that CPMs are mostly outside of your control, with a few exceptions.
Quick edit: There is some debate these days surrounding whether there is a real correlation between CPM and CPA. Some people believe that an increase in CPM means the algo has locked into a higher-intent segment of your audience (and therefore more expensive) that, due to their higher intent, are more likely to purchase and offset the potential CPA increase. I am personally not in this camp, but it's worth mentioning. Continue!
First, take a look at your CPMs. Are they higher than average? If so, here are some things to consider:
- Relevance Score: This is determined by three metrics you can track in your dashboard. Quality Ranking, Engagement Rate Ranking, and Conversion Rate Ranking. Once your ad achieves enough impressions, Facebook will give it a rating on each of these metrics, determining your overall Relevance Score. The lower your relevance score, the higher your CPMs will be.
- Audience overlap: If you go to your Audiences page, then select a few audiences and click the three dots at the top, you can click 'See Audience Overlap.’ This will show how many people appear in your selected audiences. Typically, the more people that overlap, the higher your CPMs will be if you run them simultaneously in your account. To reduce your CPMs, try limiting overlap as much as you can by either going with larger audiences or segmenting your audiences to be unique from one another.
- Size of your audience: Another good way to judge CPMs is by the size of your audience. Typically, the more segmented your audience, the higher the CPMs will be. While audiences at the bottom of the funnel will have higher CPMs because they are highly segmented, this is usually balanced by purchase intent being higher. For this reason, many people recommend using more broad audiences in your top-of-the-funnel campaigns to help reduce costs.
- Frequency: This is two-part. While frequency doesn’t directly affect CPM, it directly affects the cost in some ways. First, if your frequency is high, it could be because the engagement rate is relatively low, which would be an indicator for the above metric, Relevance Score. Second, if the above is true, you are likely overspending on audiences who will never engage with your ads, causing you to spend money on CPMs with no value.
STEP 3: CLICK-THROUGH-RATE
Click-Through-Rate is pretty straightforward. CTR is how often your ad was clicked vs. how many people saw it. Naturally, if your CTR is low, your Cost Per Click is higher, and so is every ‘cost per’ metric below that in the funnel. Typically, what best defines performance is how aligned your ad creative is with the audience seeing it, so making sure that your ad and audience are synced is very important.
Every account is different, so take a look at performance and figure out your benchmark for this. Typically, if your CTR is below 1% at the top of the funnel, there is room for improvement. Unfortunately, this can be a big game of reiterating upon creative until you find something that hits. You can cycle audiences all day long, but in my experience, the brunt of the responsibility towards performance lies in creative.
If your creative has a less-than-desirable CTR without any meaningful conversion performance, turn it off. Did it gain some traction, but the results weren’t exactly what you hoped for? Iterate upon it with different variations until you find something that moves the needle closer to success. Find something with a good CTR? Make more variations of that and scale them.
I would be remiss if I did not mention that CTR does not directly correlate with ROAS. In many cases, I’ve found my best ROAS ads to have average CTRs, and some of my highest CTR ads have had poor ROAS. While this is not an exact measurement of how to tell if an ad will be successful, it can help determine if an ad can be successful when it’s early, and data is still trickling in.
STEP 4: SPEND BALANCE
Balancing spending is a big one in many accounts, in my experience. Typically, when performance is pointing down on an ad set, the system has determined that one ad is the winner among the rest, and it focuses all budget there, despite performance being worse on that ad than the others. Facebook uses a method for deciding where spend goes that predicts and relies on long-tail performance, but most businesses don’t have the time or money to wait for a ‘maybe.’ So, what can you do to be more proactive? Well, there are a few things:
- Recreate big spender: If one ad is hogging all of the budget but bringing down performance overall, recreate it and drop it in again. This doesn’t mean duplicating it. Recreate it from scratch (don’t forget to change the file name and reupload the creative) so that system sees it as an entirely new piece of creative and starts from scratch. If the ad converts, it should have no problem finding its footing again.
- Turn it off: Typically, it just doesn’t make sense to keep spending up on a poor performer. If the results were lukewarm, it could make sense to give it another shot. Still, in most cases, the system hasn’t allocated an efficient level of spend to the other creatives in the ad set, stunting their ability to find success. Turning off the ad hogging the budget can force the system to divide the daily budget among the remaining ads to discover new winners.
- Use fewer ads per ad set: The concept is simple enough – if the system doesn’t have the option to balance the budget as it sees fit across multiple ads, the problem no longer exists. When testing creative, you can try duplicating your ad sets and leaving only 1-2 ads in each ad set. That way, you can tell the system exactly where to spend without worrying about an imbalance when finding new winning creative. I would recommend only using this method if you’ve exhausted all others, though. Generally, I will reserve splitting creative into batches between similar ad sets when testing different angles and/or formats.
STEP 5: CONVERSION RATES
When you think of a conversion rate, you typically think about the purchase as the key indicator, but to get a better idea of the whole picture, we can bring that concept to other metrics further up the funnel. For example, how many people saw the ad viewed a product page (View Content)? How many people who Viewed Content also Added To Cart? Monitoring fluctuations in these metrics can be telling about both the health of your website and the alignment between your target audience’s expectations vs. reality. To help better understand, you can create a few custom metrics in your dashboard that show your conversion rate (represented as a percentage, by the way) between different metrics, such as:
- Click To Purchase: (Purchases / Link Clicks)
- Click To Add To Cart: (Add To Carts / Link Clicks)
- View Content To Add To Cart: (Add To Carts / View Contents)
- Impression To Purchase: (Purchases / Impressions)
- Initiate Checkout To Purchase: (Purchases / Initiate Checkouts)
Monitoring these metrics and paying attention to how they change can tell you how buyer behaviour changes, if new website updates are positively or negatively affecting performance, etc.
BONUS: COST PER X
Don’t forget that you can add a Cost Per column for each metric to your dashboard. I use this a LOT. If you put Add To Cart in your dashboard, including Cost Per Add To Cart, and monitor that over time. Like Cost Per Purchase, this can indicate that something in the funnel isn’t working and requires action.