ArticleLet’s Talk Meta Advertising
I’ve been spending a lot of time auditing Meta Ad Accounts – so we’re talking about the 5 Big Things I’d like to see in more ad accounts:
1. More Emphasis On Foundation & Strategic Stuff
One of the most frustrating things I see in accounts is an over-attention to things that – quite candidly – don’t matter. I’ll provide specific examples below, but this runs the gamut from trying to tell Meta where to serve ads (you’re not as good as Meta), to trying to micro-target hacky audiences, to testing 500 different variations of a given ad. You get the idea.
What I’d like to see more marketers & brands doing is spending an inordinate amount of time iterating on big-picture, strategic components that can 10x their account performance. By that, I mean:
- Competitive Landscape Assessment – It’s honestly shocking to me how little time most brands spend understanding their competition. One recent example: a home services brand that was spending thousands a day promoting a “big sale”, when their largest TWO competitors were running objectively *better* promotions, to the same audience. The Meta Ads Library is an INCREDIBLE tool that is criminally under-utilized by marketers. Seriously. Who wouldn’t want to see exactly what creatives your competition is running, to what landers?
- Audience Understanding – Fundamentally, Meta Ads is a tool to connect brands to people. That’s it. Given that your audience is 50% of that equation, it stands to reason that investing time regularly to understand the people you’re trying to reach might be a good life choice. There are a ton of ways to do this – from focus groups to surveys (Pollfish is great) to SparkToro (an audience insights platform we subscribe to and I swear by). The number of brands that proudly proclaim they conducted audience research in 2018 (no joke, happened TWICE recently) and don’t need more is baffling. People evolve. Trends change. If you want to succeed, you’ve got to understand people.
- Behavioral Economics – This is a much bigger topic than a bullet point in a (relatively) long newsletter – but it’s one that I wish more marketers thought about. How you present options is almost always more important than the options you present. Some of my most successful ideas/campaigns/initiatives have come from designing campaigns + experiences that lean heavily into consumer psychology, cognitive biases and choice architecture, coupled with a deep understanding of both my competitive landscape and audience.
- Pricing, Financials, Etc. – More brands die from success than from failure. Rather than try to pull tiny levers in a massive ad account, spend time understanding the financials of your organization. How does money flow through the business? Where are there opportunities for margin expansion – either by reducing expenses, increasing AOV or enabling follow-on sales / repurchases? One of the easiest ways to make an ad account work better is by making a business work better first. This is one of the reasons I ask for unit economics and financials up-front – because that’s what enables me to (a) set proper targets in-platform and (b) have intelligent, action-oriented conversations with clients. It’s really that simple.
2. Streamlined Account Setup & Strategic Management
The Issue: Most accounts I see have way too many campaigns running on outdated strategies – whether that’s way too many campaigns, or 15 ad sets, each with a different Lookalike Percent (%LAL) or slightly different interest stack, or 100 creatives per campaign, or daily changes to creative/audiences/targeting/bidding, or weird targeting where certain cities are excluded and weird places are included to lower CPMs, or all of the above.
What You Should Be Doing: Focus on a “minimum viable” campaign structure. Practically, this almost always means consolidation – I typically run one campaign per offer/angle/audience, with one LAL (5% or 10%) or interest stack (current site visitors excluded), one “maximally broad” audience (defined below + w/ LAL + site visitors excluded) and one retargeting campaign. I typically exclude existing clients / active customers. That’s it. A maximum of three ad sets in a campaign; if that’s too many for you (either because you don’t have the budget or you don’t have a sufficiently large website audience), lose the %LAL or retargeting audience.
Maximally Broad, defined:
An audience with only the restrictions that are absolutely necessary, such that anyone with the excluded trait/characteristic would be a non-viable prospect/customer. I use this strategy to give Meta a “head start” on optimization, but exclude the people who aren’t going to be viable based on traits/characteristics I’m relatively confident are accurate within Meta’s dataset.
Example #1: I conducted an audit for a DTC fertility brand. In this case, a “maximally broad” audience is women over the age of 22 located in the lower 48. That’s it.
Example #2: I’ve reviewed an ad account for a home services provider that serves four specific metro areas. They work exclusively on owner-occupied residences (no renters, no landlords, no contractors). Meta’s homeowner data is very good (thanks, credit bureaus). The maximally broad audience here are people who live in those areas and are homeowners. Yes, this could exclude some weird exceptions where a renter (like an adult child) may be helping a parent or grandparent owner-occupant, or where a mortgage is held in one spouse’s name only. But candidly, those aren’t our customers, and we’re OK missing out on a few golden nuggets if it means avoiding slogging through a mountain of dung.
Strategic Management: there’s *always* a temptation for media buyers/advertisers to tinker with the ad account, especially if results aren’t immediately exceptional. You’ll see ad sets and creatives not performing at the level you’d like. You’ll see campaigns not hitting your CPA/ROAS targets and want to turn them off. Resist those temptations.
Give your account at least 1 week (ideally 2) following major changes to stabilize. Then, once you’ve given it a chance to reach an equilibrium, consolidate any must-make changes into a single update, made in a single session. This does three things: (1) it forces you to stop micro-managing (which you shouldn’t do, since you’ll never out-trade a machine) and start macro-strategizing (which you absolutely should do, because Roombas are stupid); (2) it minimizes time spent in the learning phase; and (3) it makes you more efficient and avoids duplicative or conflicting work.
3. Proper Settings & Structure
The Issue: Far too many ad accounts have improper objectives (leads/purchases), don’t use an optimal automated strategy (most run lowest cost instead of cost caps), have staggeringly long attribution windows and don’t have a sufficient budget to reach Meta’s 50 optimization events per week per ad set threshold. The reality is those mistakes compound, dramatically dragging down performance.
What You Should Be Doing: Honestly, I think this has a relatively simple solve:
- Optimize your account on the thing that matters to your organization (purchases, leads, etc.), not something else. I understand that some brands have strategic uses for Awareness & Engagement campaign objectives; I tend to avoid them unless absolutely necessary. Focus the machine on achieving the goal you care about, not some intermediate thing that may (or may not) correlate to the thing you care about. I can’t stress enough how sensitive the Meta Algo is to initial conditions.
- Tighten your attribution windows. I use 1 Day Click (1DC) almost exclusively, because (a) most people have terribly short attention spans and (b) there are plenty of studies that show most people don’t remember the ad they clicked on a week ago. Remember that Meta Ads will delay reporting by up to ~72 hours anyway, so your 1DC is, in effect, up to 96 hours delayed.
- Configure the Conversions API & Event Prioritization – this is one of those things I can’t believe I still see, but here we are. If you’re running ads on Meta, CAPI & Aggregated Event Measurement (AEM) are table stakes. It’s non-negotiable. Data is the primary optimization layer for Meta Ads, and these two components are integral to its functioning – so please, please, please take the time to configure it.
- Set Appropriate Cost Per Result Goals (CPRG). Get clarity on your acceptable new customer acquisition cost (nCAC) and your Meta Ads Coefficient (not sure what this is? Check out last week’s edition here), then use it to set a proper Cost Per Result target.
- Don’t Be Constrained by Budget. At a minimum, your campaign’s daily budget should AT LEAST equal to the following: (50/7)*(nCAC * Coefficient)*(# of Ad Sets). In an ideal world, the budget should be uncapped and your CPRG should be the limiting reagent in your money-making experiment. Why? Because each day, there’s a different number of auctions that satisfy your targeting criteria, which means each day you’ll need a different total budget in order to participate in each auction. If you’re able to drive more conversions at or below your target cost, why not spend more?
That’s it. Simple.
Now, if you’re a Cost Cap (err, Cost Per Result Goal) skeptic, I understand. I used to be one. But after seeing this work fantastically well in many ad accounts AND doing the math myself, I’ve been converted. If it’s of interest, I’ll do a more detailed breakdown on why you should convert, too.
4. Structured, Pyramid Style Testing
The Issue: Candidly, there’s too much micro-testing and way too little macro-testing in many ad accounts. There’s little reason, post 14.5 and 15.0, to be testing a 1% LAL vs. a 2% LAL vs. a 6% LAL. Ditto for testing IG Feed vs. FB Feed vs. Audience Network placements. Ditto for testing traffic vs. conversions (purchases/leads) optimization strategies. Ditto (controversially) for over-testing creative (more on this later).
What You Should Be Doing: My preferred framework for thinking about testing is a pyramid:
Your testing approach should start from the bottom up. Don’t move up until you’re reasonably confident that you’re in a good-enough place. Why? Simple: this structure prioritizes the highest-impact levers.
Case in point: for ~75% of Meta Advertisers, offer/angle is the most impactful lever for improving performance within your account. That’s doubly true in newer/start-up accounts – which is (ironically) where I almost never see said testing.
Once you have a solid offer/angle, that’s when I typically experiment (at least a bit) with owned audiences – LAL vs. Interest Stack vs. Maximally-Viable Broad – along with partner/3rd party audiences (via whitelisting). If you’re not familiar – whitelisting is a fantastically under-utilized tactic to activate other audiences, simply by running your ads through a partner’s (an influencer, a larger brand, a media brand, etc.) page. Especially for smaller brands with minimal credibility + brand recognition, whitelisting lets you reach your audience via a more credible page.
Once you have the offer/angle and audience in a solid place, then it’s time for Landing Page testing. Suffice it to say, this is an area where most ad accounts are flat-out terrible, which is why I’ve dedicated an entire section to it below.
And finally, at the very top, is creative testing. And let’s be clear: great creative is a catalyst for an ad account – one which can take performance to previously-unreachable levels. But it can’t be the only component of your testing strategy (and for far too many brands, it is).
There are multiple reasons for this, but here are my two primary ones: (1) your creative plays a significant but limited role in the overall success of the ad account; it’s job is to intrigue your audience enough to click. There’s a LOT more that needs to happen for a successful lead/sale and (2) at some point, you’ve made the best ad you’ll ever make – and at a point well before that (usually once an ad is performing in the top ~85% – 90% percentile), the marginal utility of making a better ad does not justify the marginal cost. Instead, spend that time, money and energy improving other areas of the pyramid OR opening up other funnels.
That all being said, creative testing is important. My philosophy on it follows the above structure – test big ideas/concepts first, then narrative arc/components, then the rest (CTA text, static ad layout, primary text, etc.). At a minimum, I want to see 5-7 different creative types tested in a given campaign, ranging from CGC (“Customer Generated Content”) to NUGC (Narrated User Generated Content) to Before/After, social proof/awards/credibility, testimonials, unboxing, benefit carousels & general statics.
My typical process for creative testing involves testing big ideas/concepts first (i.e. CGC vs. NUGC), then iterating on winners (i.e. testing different narrative arcs & hooks) while introducing new concepts. Honestly, I rarely (if ever) get around to different button colors or layouts. If you’re hoping that re-arranging components in your ads is going to magically unlock your ad account, Hermione Granger has a message for you:
My definition for winning is simple: sustained spend at or below CPRG. That’s it. I know some people disagree with this philosophy, but (candidly) I don’t have the volume of data Meta does, both at the user and advertiser level. I’m not the Oracle at Delphi, and I certainly don’t have a crystal ball. Meta, on the other hand, has a platform that can pick up on patterns that would otherwise be impossible for me to see, and over a sufficiently long period of time, that results in better outcomes.
5. Landing Page (“Lander”) Optimization
The Issue: Of the 9 accounts I reviewed over the past month, all of them had a majority of (if not all) traffic going to a PDP or Category Page (CP) or Home Page (HP). While there’s nothing inherently “wrong” with this, it certainly isn’t optimal – these pages aren’t speaking directly to the audience/offer/angle that’s highlighted in your ad (immediate disconnect), and they certainly aren’t tailored to the singular purpose of achieving a singular goal: converting a specific audience segment (as identified by your creative) into customers/leads/whatever.
What You Should Be Doing: This is a pretty simple one: build and test landers regularly. The major reason I’m so insistent (and bullish) on this is because landers provide incremental value to your other marketing channels (i.e. Paid Search, Organic Search, etc.), along with your Meta Ad Account, all while increasing conversion rates, AOV and customer loyalty, while lowering return/non-qualified rates. Good landers are the proverbial tide that lifts all boats.
A well-constructed Lander should be a natural part of the customer journey, seamlessly connecting the in-platform ad experience (your offer/angle and creative) to your brand in a way that meets the user where s/he is and guides them to where you’d like them to be. The less the ad destination feels like a generic experience or a cold, unfeeling sales pitch, the more likely it is to convert AND the more likely it is to convert the right kind of customers for your brand.
Fundamentally, this is no different than a high-end sales call or in-store experience at a boutique: upon arrival, there’s a warm greeting and introduction, some rapport + credibility building, pain point/challenge identification, presentation of viable solutions, decision-making, then the transaction. If we’re doing each of these things on both physical and 1:1 synchronous interactions (calls/Zooms), then why is it OK to be so fantastically generic in our paid media? Spoiler alert: it’s not.
Landers are the ideal tool to create an experience that empowers the customer AND moves them to action. The question then becomes: how do we go about this process?
To keep this (relatively) short, I think Lander testing should mimic the creative testing strategy described above: start conceptual, and move into the tactical/granular.
That means starting with big-picture landing page concepts/types:
- Offer-focused – Product/Category Specific Lander – this is the primary lander type I see in the wild, and for good reason: the content of the lander is specifically tailored to the creative in the ad account – creating a seamless experience. The downside is that it (by design) limits product or offer selection to what’s included – which can cap AOV for ecommerce brands or result in a increased competition for B2B/B2C lead gen (i.e. the customer isn’t aware that Len the Plumber also does HVAC, so s/he doesn’t contact them for that service).
- Quiz / Assessment – I love a good quiz – it’s an oldie, but a goodie. There’s no better way to trigger the Ikea Effect than giving your potential customer the ability to customize their product bundle, all while gaining insanely-valuable zero-party data that you can leverage in future interactions/communications. It just works.
- Listicles – there’s a reason advertorials are insanely popular in traditional media: they provide a side-door into the prospect’s mind that builds credibility and avoids the typical “defensive” response to being marketed to / sold, while providing ample opportunity for storytelling and objection-diffusion. You can achieve the same result with a listicle that feels like a third-party article, but is really just a delightful lander. Here’s an example from Earthling that I think is fantastic.
- Influencer-Specific Page – there’s no denying that influencer marketing can be effective, especially if there’s an audience overlap. Continuing with one of the themes of this article, I think continuity is critical – so why not feature the influencer throughout the experience? I came across this example from Molekule, but I’m sure there are others.
- Article/Write-Up – similar to the listicle above, any chance you have to highlight your product on a third-party or publisher website (I’ve seen these everywhere from the Business Journal or Trade Publication to Conferences (“Sponsored Sessions”) to actual businesses that focus on this. It’s basically the weird love-child of PR & Advertising – and if it’s done well (and tastefully), it can work wonders. Here’s an example from “SweetKick”, produced by Premium Purveyor.
- Shadow Brand Lander (“Homepage-Style Lander”) – an under-utilized variant of the “offer-focused lander”, these are pages that look like a home page, but are actually landers tailored to the audience/angle targeted by ads. Each one has all of the components you’d look for in a lander – compelling headlines, clear + targeted benefits, social proof, scarcity, objections + responses and clear next steps, all weaved into a seamless story. Here’s a great example from Caraway Pans.
Once you’ve found a core concept that works, follow the same process as you would with creative testing: create variants of the winning concept that have different page structures, flow, components, etc. And once you’ve found a good structure/flow, then think about copy, CTA, color, orientation tests.
The good news about landers is that they tend to have longer shelf lives than ad creative – so your investment in them will (likely) pay larger dividends.
