No matter how adept you are at managing performance marketing campaigns, your account will at some point experience less-than-impressive results. There are numerous variables that contribute to crappy performance. Some of them will be outside of your control, but fortunately, some will be very much within your control!
I’ve put together this guide to help you troubleshoot several key elements of this situation:
This may be obvious to some. For most, “poor performance” means cost per acquisition (CPA) is up, or return on ad spend (ROAS) is below goal.
In short, conversions are not happening, and you (or your client) are potentially losing money.
Fluctuating results and occasionally erratic algorithm behavior are part and parcel of being a performance marketer. However, outside of seasonal trends, we’ve listed metrics that will affect businesses small and large, and how to handle anomalies when they creep up.
We’re going to walk through how to identify the root cause, piece the story together, and implement next steps. If you’re not hitting your goals, fear not. There is a reason, and you can find out what’s causing it.
There are a few different metrics you can look at to determine why performance has dipped. CPA and ROAS are usually the red flags, but these metrics are the result of a problem, so these numbers won’t tell you much except that a problem exists.
The first place to start is by asking: What recent major changes have I made that could cause this? Did we launch a new audience? Pause out certain creative? Remove a promotion? Launch a new website theme?
Next, it's important to look at your data week over week (and month over month) so you can identify trends that you may not have seen otherwise. Then you can drill down to specific metrics, i.e., CVR, CPM, CTR, and AOV. Let’s explore these metrics in detail.
The first and most complex metric to explore is conversion rate (CVR). If your Facebook conversion rate drops off, you know that it came from one of two places. First, the problem could be on the website itself. For instance, some element of your website is causing friction, even though you’re driving traffic to it. Or, the problem could stem from the platform itself (read: the traffic you’re sending).
To figure out where the drop in CVR is coming from (website or platform), compare Facebook traffic CVR to direct/organic traffic CVR. If you see a steep drop on all traffic sources, something is probably up on the website.
See the below gif to understand how, with all other metrics held constant, changes in conversion rate will ultimately affect CPA and ROAS.
CVR Fluctuations Due to Website Changes
One of the biggest problems our smaller ecommerce clients experience is inventory issues. If a brand is out of their bestselling items, conversion rate may drop off. Especially small to mid-size clients, who might have a week when performance is awesome, but then see a lull. Oftentimes we see that one particular product took off, sold tons of units, but then went out of stock.
Another trigger for fluctuating performance is changing the user interface. It might be the homepage banner, a new pop-up, a change in the free shipping threshold, or the basic layout of the home, product, collection or cart pages. Changes to your “path to purchase” that cause unnecessary friction and make it harder for customers to complete their orders will immediately impact conversion rate.
These triggers may seem obvious, but people don’t always think about them. And here’s why: Brands assume that these points of friction are things people are smart enough to get around.
One eCommerce brand launched a new seasonal collection and wanted to promote it with paid ads. However, the entire collection was on pre-order for a month. As we anticipated, performance was far from awesome, but our client was surprised. What I've learned from being both an advertiser and consumer, is that people want "it" now. They want the path to purchase to be simple, and they want their products immediately.
Brands mistakenly assume that when they have a product customers like, users will fight to find it and purchase it. But if the website has points of friction, i.e., pre-orders, having to scroll through eighteen lines of out of stock products, potential customers may get frustrated and bounce off the site.
To avoid potential disruption when making website changes, you might try running both versions of the site as an a/b test, so you can measure the potential impact and compare conversion rates. Running an A/B test will allow you to gauge how your changes will affect overall website performance before you make the final switch. Remember, the best way to prevent volatility is to monitor and analyze week over week KPI’s, taking note of seemingly insignificant fluctuations before they become significant.
Now that we’ve talked about the website changes that can affect conversion rate (CVR), let’s talk about the platform side.
CVR Fluctuations Due to Platform Changes
Let’s say your inventory is perfect and you’ve changed nothing on your website, but performance has fallen. Look to changes you've made recently on Facebook.
For instance, you may have launched a new audience based on website visitors or post-engagers, or used an audience from a brand collaboration network like Wove. Using a less qualified or shared seed audience could result in a lower CVR (but not always). If you notice that your click through rate (CTR) is strong, but conversion rate is down, there could be an audience issue. The users may not be qualified, i.e., the products are out of their price range, or they simply don’t match your customer persona.
It’s also possible for poor conversion rates to originate at the creative level.
Perhaps you recently launched a new ad that’s vague and lacks transparency. That can often lead to poor performance if potential customers are surprised or feel misled once they get to your site. All the more reason to be as transparent as possible in your ads. It weeds out unqualified users who are less likely to convert.
Take an expensive apparel brand. They would ideally describe their products in their ads as “elevated” or “luxury” to set expectations. If the copy is clear about what the product is, users will be pre-qualified before they even click.
The ultimate goal is to have more qualified people visiting your website.
If a luxury brand advertises their products as “affordable” and users are directed to a page advertising $300 pants, there will be negative consequences. For one, you’ll have a lot of angry users (who usually end up leaving negative comments on ads) and an awful conversion rate.
In conclusion, if your CVR is the culprit, look back at website changes, look at recent changes to targeting/audiences, and look at newly launched creative.
Because Facebook charges per impression with conversion objective campaigns, a decline in click-through rate (CTR) will ultimately affect CPA and ROAS. Even if all other metrics are consistent, a drop in CTR means that you're paying for the same number of impressions but driving less clicks. A decline in click-through rate usually occurs either from the creative itself, or from ineffective targeting/audiences.
See the below gif to understand how, with all other metrics held constant, changes in CTR will ultimately affect CPA and ROAS.
When Creative Drives A Low CTR
It’s extremely important to continuously test new ads. Whether that means new imagery, new copy or different calls to action, brands need to keep creative fresh and exciting. The most obvious cause of a decline in CTR is that a new batch of ads is just not resonating with the audience.
The tougher question to answer is : Why not? My approach to writing copy + selecting imagery is to be as transparent as possible about the product, and provide imagery that’s eye-catching and will stop a user mid-scroll.
If your CTR is below average for new creative, you’ll want to swap out copy and imagery and test another batch. If you’re lacking inspiration, try looking through our Creative Playbook for guiding your creative process.
When Targeting Drives A Low CTR
To identify if your targeting is the reason behind low CTR’s, compare the click-through rate of ads that are running in two separate audiences, and note significant disparities.
One of our clients wanted to test a Wove audience (an audience shared by another, similar brand), to see if we could use another brand’s customer lookalikes to break into different pools of users.
We took our top-performing ads and tested them in the new audience. The result? CTR was 67.5% lower in the new audience. This tells us we either need different messaging, or this audience doesn’t fit our target customer persona and isn’t interested in our ads.
Metric #3: Cost Per 1,000 Impressions (CPM)
Unfortunately, CPM increases are one of the variables we don’t have control over. CPM increases are a direct result of increased competition in the Facebook Ads auction. Even if you can't control it, you can still track CPM trends to identify if this is the reason performance is fluctuating.
See the below gif to understand how, with all other metrics (including spend) held constant, changes in CPM will ultimately affect CPA and ROAS.
You can’t control how much competition is in the auction. But, you can try to mitigate the increase in CPM by launching ads with higher CTR’s, or adding incentives to increase CVR.
It’s also beneficial to anticipate exactly when these cost increases are likely to happen. For example, the cost of advertising on Facebook increases during Q4. During this time, many advertisers are dumping their marketing budgets into the platform to capitalize on increased shopping activity & hit their yearly revenue goals. Facebook has limited ad inventory, so if more advertisers are active, the platform is going to prioritize accounts with the highest bids.
Below is a real example of how CPM may fluctuate MoM - CPM for this client was 27% higher in November than in September.
Another important trend to note during this time period is conversion rate increase. Specifically for e-commerce, conversion rates increase mid-way through November and into December due to holiday shopping. The best-case scenario is that higher CPMs are offset by higher CVRs.
See below for an example of how a CPM increase can be offset by a seasonal increase in conversion rate:
Metric #4: Average Order Value (AOV)
One final metric that’s often overlooked is average order value. This is an important metric for eCommerce brands that have variable pricing across different product types. If you’ve recently promoted a collection with a lower AOV, even if all other metric stay consistent, overall ROAS will be affected.
One of our clients launched a ton of new colors for their lower-priced product in September. Compared to July (one of our best months) conversion rate increased 68% and volume of sales were up. However, AOV was ~$100 lower, resulting in a lower ROAS. Had AOV been the same, we’d be looking at a much higher 335% ROAS (other differences in CTR and CPC are also at play).
Pushing Through To Drive A Better Outcome
This framework for evaluating Facebook performance and identifying issues is definitely not all-encompassing, but should serve as a starting point for getting your campaigns back on track. No blog post can map out a solution for every issue, because one size never fits all. In the world of performance marketing, many problems and issues will creep in unnoticed, some of which you can control, some of which you can’t.
Ultimately, the goal here is to approach the problem in the most strategic way possible, engaging both your grit and curiosity to push through to the other side. Without that mindset, you will never get results to where they belong.