This is part two of a response to the Tweet below, which took off last week in the DTC Twitter world. Part one was about facing the situation with the right mindset. This post is going to focus on tactical steps you can take to improve your Facebook Ads Performance after iOS 14, an Apple update that drastically reduced the amount of data in the Facebook Ads ecosystem.
The advice in this post is layered. If you find your problem in the first subsection, you can stop reading and get to work on implementing a fix. But for this advice to work, you’ll need to run the audit in the order described here, not skip to where you think the problem is. Let’s dive in.
An unfortunate habit we have in eCommerce and marketing is solving the wrong problem. Something will change in our business performance, or we’ll miss expectations. We then enter a reactive state, grab hold of the first explanation that resonates, and don’t let go.
But if you solve the wrong problem, you typically wind off worse than when you started. Most of you are here because you want to improve your Facebook Ads Performance after iOS 14. But it’s worthwhile to pause, breathe, and take a beat to understand if that is really your problem. Here’s how to do it.
You are going to need the following data points by week for the past five years, or for as far back as you can go:
From this data, calculate the following ratios for each week:
I get it, this is a big ask. But you should start tracking these metrics weekly anyway. No better time than now!
Shopify offers a new vs returning customer report as one of their out of the box analytics dashboards. If you use another eCommerce vendor you may have to calculate this yourself. These views are critical to the understanding of your business beyond a single channel’s performance. So it’s worth hiring someone to run a one-off analysis or investing in software to create corresponding dashboards.
Create a line graph of your weekly ratios, with each metric in its own graph:
Create a trailing 12 month average calculation of your raw data. This is a formula that averages the last 12 months of data leading up to the reporting month. So your trailing 12 month average new customer revenue for December 2020 would be an average of all new customer revenue for Jan – December 2020.
Raw data vs the 12 month trailing average for that same date
This is an example of charting raw data vs the 12 month trailing average for that same date. The T12M view smooths out seasonal trends and other outliers, giving you a clearer picture of what’s actually going on.
Before you create the T12M average, sum the weekly data into months. As a reminder, these are your raw data points by week:
This doesn’t mean “jump straight to a hypothesis!” nor does it mean “fit the data to your existing hypothesis!”.
For each of the graphs you created in step two, you want to flag events that are foundational to your business. This will enable you to easily view trends in context of macro factors. Some examples of events you will want to flag:
Try to explain the trends on the graphs within the context of the events you just flagged. The rolling 12 month average charts will show you inflection points–moments when revenue or customer counts increased or decreased considerably. See if these align with any of the events you flagged. If they don’t, think about what else was happening in your business, or the world, at that time.
I’ll be honest: executing steps one through three above requires some time, and prior experience helps. If you’re left feeling overwhelmed, here are some common patterns to look out for:
The “Covid Bump” – Covid-related store closures were a macro event that resulted in more people spending more time online, and buying more stuff there. This, plus some risk-averse brands pulling out of Facebook during the same period, resulted in above-average efficiency from appx April 2020 – Jan 2021. Learn more here.
Over-reliance On Customer Retention – The number of returning customers surpassed the number of new customers at some point in the past 2-3 years and it has stayed that way. This will eventually impact overall media efficiency.
Acquisition Plateau – read more about this here.
Changes To Product Assortment – if you made any changes to the formulation, design or pricing of your product line in the last 2 years, this may have caused a ripple effect on media efficiency.
There are even more patterns that could be impacting your Facebook ads performance. Sometimes media performance is symptomatic of other things going on in your business. If you identify one of these issues, you need to solve that before you can improve your Facebook Ads Performance after iOS 14.
But if your performance declines are truly isolated to Facebook, read on.
I wasn’t going to include this one. You know how I abhor “best practices” (see: the name of this website). But a few folks on Twitter brought up a good point: some agencies still don’t quite know what they’re doing regarding Facebook.
I am going to frame this section as if your intention is to use Facebook as a way to build focused, distinct campaigns for acquisition and retention. A lot of media buyers, and a lot of brands, don’t think that way about their businesses.
But if you’re relatively young or looking to grow more than 10% per year, you’re probably using Facebook Ads for customer acquisition. So here is how to audit your ad account in that context.
Your highest budget Facebook Ads campaigns should serve one of three purposes:
If you’re trying to grow consistently and the business is relatively young, acquisition should comprise 50-80% of your total spend.
Within Facebook’s ad architecture, campaigns should be limited to one of these three objectives. Each campaign should contain a maximum of 1-3 ad sets. If there is more than one, you need a strong reason why, because multiple ad sets result in slower learning. Each ad set should contain a maximum of 1-4 ad creatives when it’s out of the learning and evaluation phase.
Facebook provides a few options for acquiring net new customers:
If you’re a relatively young business without a deep customer file, broad targeting is typically your best bet. If your product has an obvious use case or niche audience, targeting with interests can also work well. Audience-based lookalikes can perform well once you have 3-4 years of transaction data and some true “best customers”.
If you want to improve your Facebook Ads Performance after iOS 14, broad targeting with lots of creative iteration is probably your best bet. But all four strategies are worth testing. However, pixel-based audience data has become less robust since the update.
When you’re running a pure acquisition campaign, you should implement an exclusion list of all your prior customers. In some cases, you should also exclude prior site visitors. If you don’t do this, you don’t truly have an acquisition campaign.
Why does this matter? Because warm leads (like existing customers) are less expensive to convert than cold leads (net new customers). If you mix them into a single campaign, the Facebook algorithm will optimize towards existing customers because they will generally deliver a higher return on ad spend. So you won’t acquire many new customers.
By the way, this is a dirty trick some less-than-honest media buyers will pull on you. They’ll report great ROAS metrics for an “acquisition” campaign with retention audiences secretly mixed in.
There are certain creative tropes that tend to perform better on Facebook and Instagram. You can’t take a photoshoot that was designed to run as a print ad, crop it into a Facebook ad, and expect it to perform.
Creative testing has become the most crucial component of success on the platform. To get the best performance you should be generating and testing a minimum of 5-7 new creatives per week. If you have been running the same creative for more than a month, you may have exhausted its performance potential.
Here are some general guidelines for winning Facebook creative:
If you haven’t tested into all of these guidelines yet, it’s worth a try. And if you have mostly run dynamic product feed ads in the past, it’s worth testing out traditional creative formats focusing on a single product or offer.
Double-check Facebook’s official list of actions you can take to counteract the data loss resulting from iOS 14. See here and “Actions You Can Take” here.
You can also use the media and Facebook Ads efficiency data you pulled in the first section to determine if your true problem is the accuracy of Facebook’s built in reporting. Sometimes your campaign performance is only 25-50% of the problem, but that performance loss is amplified by data loss in Facebook analytics.
Generally, the longer your typical time to conversion, the less accurate Facebook’s attribution reporting has become. If you were managing the account based on seven day click (or longer), you’re less likely to have accurate data now.
Some media buyers on Twitter have been saying that Facebook Business Manager is no longer a reliable same-day source of performance data. If you find yourself in this situation, you can use traffic quality metrics in GA to get a read on ad performance before the conversions roll in. High time on site and low bounce rate are good indicators that the ad has found its audience.
This section might ruffle some feathers. If you don’t have a lot of knowledge of the Facebook platform and you rely on an agency to run everything, getting a straightforward audit of the points above may be challenging. And full transparency is critical to improving your Facebook Ads Performance after iOS 14.
Unfortunately there are some bad people out there looking to make a quick buck. Others don’t understand how to use Facebook as a prospecting channel and will claim their own way is the best way…or at least good enough.
A few reasons to be on elevated alert regarding your agency’s Facebook Ads expertise:
If your agency checks any of these boxes (especially the last one!), you may be best off having one or more third parties do a covert audit of your account setup and performance. This is especially true if your agency can’t provide a compelling narrative for the downturn in performance and hasn’t proactively tested new approaches towards account structure and creative.
A big caveat to this advice: it really only applies if you have been working with the same agency for at least a year. It takes time for the agency/client relationship to ramp up and yield results. If you have been cycling through new agencies every 2-4 months, the agency is not the problem.
If you want to arm yourself with the knowledge to better navigate agency relationships, sign up for the No Best Practices newsletter.
At this point, let’s assume that your Facebook Advertising best practices are locked down, your agency is trying its best, and your performance is still struggling. This is the point where you can truly and legitimately say: “these iOS 14 changes have disrupted my business”.
The result of iOS 14 is that Facebook has less data to inform their advertising algorithm. Specifically, they have less visibility into what types of products users are browsing and buying across the web. This has two outcomes for advertisers: it’s harder for ads to find the right people, and it’s harder to track when viewers of your ads converted.
When you leverage any sales channel–digital or IRL–there is a concept called product/market/channel fit. When you achieve this fit, it means that your product is appealing to the audience in the channel, and its price and characteristics lend themselves to being showcased in the channel.
Facebook and Instagram advertising example: ads featuring a bright, floral print dress for $250 are always going to convert better than ads featuring a taupe oversized blazer for $750. This is because bright clothing is naturally more scroll-stopping. It’s also because the upper-middle-class women between 35-55 who are most likely to buy things from Facebook ads would find the dress more appealing and its price more reasonable.
With more conversion data, it was easier for Facebook to find micro-audiences that aligned with your product, even if that product was less mainstream. After iOS 14, that’s harder.
Here are a few reasons specific to product/market/channel fit why your performance may be suffering:
It’s impossible to know the true upper limit of your total addressable audience. But if that number is only 100,000 – 200,000 people, it may now be impossible for Facebook to find those people efficiently. Products addressing a niche medical concern, products serving a niche hobby, and hyper-luxury products could all fall into this bucket.
To succeed here you’ll need to expand beyond your niche, find a place where your niche hangs out in more concentrated numbers (an online forum?), or build a community for that niche yourself.
The further your sales cycle extends beyond a one day click conversion tracking window, the less accurate Facebook’s in-platform analytics will become. A higher AOV also means that you’ll need to spend more to get a campaign to leave “learning mode” and fully optimize.
You’ll have to get creative about tracking the return on investment for high ticket items. It may be possible to use Facebook ads as a lead capture tool, and then track the quality of those leads within your own systems. You can also use some of the metrics from the first part of this post to gage the performance of your ad spend outside of Facebook.
This one goes out to both multi-brand retailers and drop-shippers. I believe that drop-shipping was an arbitrage opportunity that has, at this point, been fully competed away. If you are selling a product that anyone can purchase from Alibaba using aggressive copywriting tactics and taking a big markup, stop right now. Find a new business model.
For everyone pursuing more legitimate and meaningful avenues of wholesaling…there is nothing inherent in the mechanics of iOS 14 that makes selling someone else’s product harder. What is happening: the data loss is making everyone’s spend less efficient, and you’re already working with a smaller than average margin for error.
To create some breathing room, take a second look at your value proposition. Some ideas: exclusive access to desirable brands, subject matter expertise, selection, curation. Try speaking to some of these aspects of your business in your advertising. “Here’s all our stuff!” isn’t going to cut it anymore.
Maybe the most your business model can handle is a CAC that is one third of your AOV, achieved within seven days. You used to hit that effortlessly, and now you can’t. You need to buy yourself some breathing room. Try the best practices audit in the second section of this post. But don’t overlook other options you may have to improve your Facebook Ads Performance after iOS 14.
Maybe there is a way to bring the cost of goods down. Maybe you can get away with raising prices or removing costly benefits like “free shipping on all orders”. Maybe you can cut some opex costs by negotiating with vendors, going remote, or finding ways to work more efficiently. Maybe you can expand your assortment to bring a greater percent of your new customers back again, sooner.
This happens when you identify something that seems to work well and zoom in on it for too long, at the expense of exploring other options. The local maximum effect can manifest in several ways within your Facebook advertising strategy.
Maybe you have a five product offering, and one of those products works really well in ads, so you’ve neglected testing others for a while. Maybe you’ve landed on a bundled offering and landing page format that consistently beats your targets, so most of your testing has been iterating around that format. Maybe you’ve seen a ton of success with dynamic ads for broad audiences, so you’ve stopped testing other ad formats.
You are never going to know what riches lie beyond the local maximum unless you go out exploring. Make a list of everything you’ve learned about what “works” for Facebook advertising for your business. Then make a list of “worst practices”–what would it mean to do the exact opposite? You don’t have to execute all of these ideas, but you can use them as a jumping off point for identifying new opportunities.
This is also an area where getting an outsider’s perspective can be incredibly valuable. Consider hiring a consultant, joining an industry mastermind group to get perspectives from other brands, or engaging an agency in an account audit.
I am confident that by now, almost 3,000 words into this piece, you have found something to help improve your Facebook Ads Performance after iOS 14. But if I’m wrong–if you have truly tried all of this stuff and none of it worked–then you’ve reached the end of the line. You need to find a way to diversify your advertising strategy away from Facebook.
The truth is: any brand who relies heavily on Facebook advertising for growth should start thinking about diversifying their advertising mix. Facebook Advertising is no longer the bulletproof go to market strategy that it once was. And frankly, it adds a layer of abstraction to what is actually happening when you sell a product. You create a product to serve an audience, not hit a ROAS goal on Facebook.
If you’re looking for short term, low(ish)-lift solutions, there are options. But the best option for your specific business is highly dependent on the type of product you sell.
These are products that solve a problem. Examples: a cream that clears up acne, a spray that dissolves soap scum from your shower, a UV-blocking swimsuit. These are probably the easiest products to advertise, because the audience is very straightforward–it’s people suffering from the problem. And chances are, these people are Googling for answers.
SEO and SEM are a slam dunk for products like these. Youtube can work as a compliment because it operates in the same ecosystem (if you have the resources to produce video ads). Content marketing, with SEO attached, can be a valuable long term investment.
These are products that solve a less pressing, less tangible problem. But they have an obvious audience. Examples: The Nugget Play Couch–no one needs a configurable play fort, but it’s obvious who would absolutely love one. Fielder’s Choice Goods–no one needs a wallet made of retired baseball gloves, but it’s obvious who would think this was much cooler than a regular wallet.
Selling these products outside of Facebook requires you to find ways to connect with your target audience. You need to figure out which social networks have a core, engaged audience that aligns with your demographic. For example, moms and DIY-ers love Pinterest. Influencer marketing can also work well here.
These are the hardest products to sell, because their audience is taste-based. Anything fashion-related is a great example. Price and sizing put some tangible limits on your potential audience, but you need to find a sub-audience who likes your aesthetic and/or values.
To score near-term wins in channel diversification, you need to reverse-engineer your core customer group. You may not be targeting a specific demographic profile, but what profile has your brand attracted in the form of best customers?
Lilly Pulitzer is a great example. If you’re familiar with the brand, you know who the Lilly “woman” is. You can find her younger iteration on TikTok (#rushtok) and in sororities. You can find her older iteration in preppy resort towns like Palm Beach and Newport Beach.
You can also try to identify tangible needs that your brand serves. La Mer sells some of the most expensive skincare on the market. But for a long time they drove awareness and customer acquisition by framing their brand as the best, most effective anti-aging solution on the market.
If you’re looking to get things back to the way they were ahead of Holiday 2021, I have bad news: that’s going to be hard to pull off in six weeks. The worst case scenario is a situation where you planned to comp your Holiday 2020 growth, purchased inventory against that plan, and now find yourself in a situation where it will be challenging to move that inventory profitably.
There are solutions here, but they are highly specific to each business. As you work through all the other steps outlined here, start to think of creative ways you can offload some of that inventory or tap into new sources of demand.
A few ideas:
Side note–industry folks have been talking about and analyzing the “Covid bump” since late 2020. If you made the mistake of attempting to comp Covid demand this holiday, you can charge it to the game, aka the school of hard knocks. To avoid similar costly mistakes in the future, ask for (and pay for) help from those who have been there. Good help is hard to find, but a $10K investment today could save your business tomorrow.
The common theme in all of the advice I shared here is creativity and flexibility. That’s why I started part one of this series with mindset advice. I hope what I’ve shared has been helpful, not overwhelming. Take this turmoil as a positive sign: your brand is evolving from a simple tactical loop into a full-fledged, complex business. So dig in and get to work.
Supercharge your growth with a purpose-built ecomOS for brands and agencies.