Facebook Advertising Is Capturing Demand You Already Had

Facebook advertising is not broken. It is just doing something different from what most marketers think it is doing. For the majority of brands running paid social on Meta, the platform is not growing their customer base. It is collecting people who were already going to buy, wrapping attribution around them, and charging for the privilege.

That is a problem worth taking seriously, especially if you are making budget decisions based on what the platform tells you about its own performance.

The Attribution Problem Nobody Wants to Say Out Loud

Early in my career, I was deep in performance marketing. I believed the numbers. I watched ROAS figures come in, felt good about them, and built the case for more budget. It took me longer than I would like to admit to start questioning what those numbers were actually measuring.

The shift came when I started running incrementality tests on accounts I had inherited. The results were uncomfortable. Campaigns that reported strong ROAS were, in many cases, reaching people who were already in market. People who had visited the site, searched the brand name, or were close enough to purchase that the ad was essentially a receipt, not a reason to buy.

This is not a fringe view. It is a well-documented pattern in performance marketing, and it is one of the reasons I now believe that much of what performance channels get credited for was going to happen anyway. The channel captures the conversion. It does not always cause it.

Facebook’s attribution model has historically been generous to Facebook. View-through attribution, multi-touch windows, and the platform’s own pixel create conditions where the same conversion can be claimed by multiple channels simultaneously. Meta reports what Meta sees. It does not know what Google saw, what the email saw, or what the brand’s outdoor campaign contributed six weeks earlier.

This is worth reading alongside broader thinking on why go-to-market feels harder than it used to, because part of the answer is that marketers have been optimising for metrics that flatter the channel rather than metrics that reflect business reality.

What Facebook Is Actually Good At

To be clear about where I stand: Facebook is not a bad platform. It has genuine strengths. The audience scale is real. The targeting capabilities, even post-iOS 14, are meaningful. The creative formats have evolved. For certain objectives, in certain categories, with certain audience dynamics, it can work well.

Where it works best is in retargeting warm audiences, supporting consideration for high-involvement purchases, and driving short-term activation for brands that already have strong awareness. In those contexts, the platform is doing something useful. It is not creating demand, but it is closing it efficiently.

Where it struggles is in the job most growth-focused marketers actually need it to do: reach people who do not yet know they want what you sell, and change their minds. That is a fundamentally different task. And it is one that Facebook’s algorithmic optimisation, which is built around finding people most likely to convert, actively works against.

Think about it this way. If you optimise for purchase conversions, the algorithm finds people who are already close to buying. That is exactly what you asked it to do. But if your business problem is that not enough people know you exist, or that the category is underdeveloped, or that you need to shift consideration among a new demographic, optimising for conversions is solving the wrong problem.

Growth requires reaching new audiences, not just capturing existing intent. Those are different strategies, and conflating them is one of the most common and costly mistakes I see in marketing plans.

If you are thinking through your broader approach to growth and channel strategy, the Go-To-Market and Growth Strategy hub covers the commercial frameworks I find most useful for making these decisions with clarity rather than habit.

The Audience Saturation Problem

There is a ceiling most Facebook-heavy advertisers hit, and it is not a budget ceiling. It is an audience ceiling.

When a brand builds its growth model primarily around Facebook, it tends to cycle through the same pool of people. Lookalike audiences help, but they are still modelled on existing customers. Retargeting pools are, by definition, people who already know you. Interest-based targeting overlaps heavily with in-market audiences. Over time, you are spending more to reach the same people with diminishing returns.

I have seen this pattern play out across multiple accounts and multiple industries. A brand scales quickly on Facebook, the numbers look great, then growth plateaus and the instinct is to increase budget or test new creatives. Sometimes that helps temporarily. But the underlying problem is structural: the addressable audience on the platform, for that brand’s category, has been largely exhausted.

This connects to something I think about often in the context of brand building versus demand capture. A brand that only advertises to people already in market is not building a brand. It is running a very expensive closing operation. The next wave of customers, the people who will be in market in six months or two years, are not being reached. They are not being primed. When they eventually come to buy, they may well choose a competitor who invested in awareness earlier.

BCG’s work on brand strategy and go-to-market alignment makes a related point about the long-term cost of under-investing in brand relative to activation. It is a structural risk that shows up slowly, then all at once.

The iOS 14 Hangover

Apple’s App Tracking Transparency framework, introduced in 2021, changed the data landscape for Facebook advertising in ways the industry is still adjusting to. Opt-in rates for tracking were low. Signal loss was significant. Conversion modelling filled some of the gap, but modelled data is not measured data, and the two should not be treated as equivalent.

What this means practically is that the confidence intervals around Facebook’s reported performance widened considerably after iOS 14. The platform shifted toward modelled attribution, which is better than nothing, but it introduces uncertainty that many advertisers have not fully accounted for in how they evaluate results.

I have spoken with performance teams who were still using pre-iOS 14 benchmarks to evaluate post-iOS 14 performance. The numbers looked worse, so they concluded the platform was declining, when the more accurate conclusion was that measurement had changed. That is a meaningful distinction. One implies a platform problem. The other implies a measurement problem. They require different responses.

The honest position is that Facebook advertising is harder to measure accurately than it was four years ago. That does not make it worthless. But it does mean that anyone presenting Facebook ROAS as a clean, reliable number is either not being careful or not being honest.

When You Should Genuinely Reduce or Stop Facebook Spend

There are specific situations where pulling back on Facebook advertising is the right commercial decision, and I want to be concrete about what those look like.

The first is when incrementality testing shows that a material portion of attributed conversions are not incremental. If you run a holdout test and the control group converts at nearly the same rate as the exposed group, the platform is not driving those sales. It is observing them. Paying for observation is not a growth strategy.

The second is when your brand has a significant awareness or consideration problem that Facebook cannot solve. If evidence suggests that large segments of your target market have never heard of you, or hold incorrect perceptions about your product, Facebook’s conversion-optimised delivery is not the right tool. You need reach and frequency among new audiences, which typically means broader media investment.

The third is when the cost of acquisition via Facebook has risen to a point where it is no longer commercially viable, and testing has confirmed that creative and audience optimisation have been exhausted. This happens. Platforms mature, competition increases, CPMs rise. At that point, the question is not how to make Facebook work, but where else to find customers efficiently.

The fourth, and perhaps the most underappreciated, is when Facebook spend is crowding out investment in channels that build longer-term equity. If every pound in the budget is chasing short-term conversion and nothing is going into channels that build awareness, consideration, or category presence, the brand is mortgaging its future. That is a business risk, not just a marketing one.

There is useful context on channel diversification and creator-led approaches in this resource from Later on go-to-market strategies with creators, which is worth reviewing if you are thinking about where else your media investment could go.

The Reframe That Changes Everything

The title of this article is deliberately provocative, but the actual argument is more nuanced. The question is not whether to advertise on Facebook. The question is what you expect Facebook to do for your business, and whether you are measuring it in a way that reflects that honestly.

When I was running agencies, one of the most useful exercises I could do with a new client was to ask them to separate their marketing budget into two buckets: demand creation and demand capture. Most could not do it. Everything was lumped together under “digital” or “performance” with an implicit assumption that it was all working toward the same end. It was not.

Facebook, when used for retargeting and lower-funnel activation, sits firmly in the demand capture bucket. That is legitimate and valuable. But if your growth strategy depends on Facebook to create new demand, to reach people who have never considered your brand and bring them into the funnel, you are asking the platform to do something it is not well-designed to do.

The reframe is this: treat Facebook as one part of a portfolio, not as a growth engine. Use it for what it does well. Build other channels for what it does not. And measure the whole portfolio against business outcomes, not platform-reported metrics.

This kind of thinking connects directly to how growth strategy should be structured at a commercial level. If you want to go deeper on that, the Go-To-Market and Growth Strategy hub is where I work through the frameworks that I have found most useful across the industries and business models I have worked in.

What to Do Instead, or Alongside

If you are reconsidering your Facebook allocation, the question immediately becomes: where does the money go? I want to give a direct answer, because vague advice about “channel diversification” is not useful.

For brands with a genuine awareness deficit, broadcast and video channels, including connected TV, YouTube, and audio, offer reach at scale among audiences who are not yet in market. The attribution is messier. The results take longer. But the long-term payoff in brand equity and category presence is real.

For brands in categories where organic search intent is strong, paid search and SEO investment builds a channel that is structurally different from social. Someone searching for a solution to a problem is expressing intent. That is a different quality of audience from someone who saw an ad while scrolling. Semrush’s analysis of growth marketing approaches covers some of the mechanics here, including how brands have built durable acquisition channels outside of paid social.

For brands where relationship and trust matter, email, content, and community channels build something that paid social cannot: a direct relationship with an audience that you own. Platform algorithm changes do not affect your email list. iOS updates do not affect your newsletter open rate.

None of this means abandoning Facebook entirely. It means being deliberate about the role it plays, the budget it receives relative to that role, and the metrics you use to evaluate whether it is earning its place.

Thinking about how to structure a go-to-market approach that works across channels, not just within one, is something I have written about in more detail in the growth strategy section. The commercial logic is consistent regardless of category: match the channel to the job, measure the job honestly, and do not let platform reporting substitute for business thinking.

A Note on the Brands That Should Genuinely Stop

There is a subset of advertisers for whom the answer genuinely is to stop, at least temporarily. These tend to be businesses where the Facebook account is running on autopilot, where nobody has run a proper incrementality test in years, where creative has not been refreshed in months, and where the ROAS figure is being used to justify continued spend without anyone asking whether the number is real.

I have walked into accounts like this. The platform is spending, the reports look acceptable, and the team has learned not to question it because the number is “good enough.” But good enough compared to what? Compared to a platform benchmark that the platform itself set? That is not a standard. That is a comfort blanket.

If you cannot answer the question “how much of our Facebook-attributed revenue would have happened without Facebook?”, you do not actually know whether the spend is working. You know whether the platform says it is working. Those are different things.

Stopping in that context is not defeatism. It is commercial discipline. Run the test. Understand the baseline. Then make a decision based on evidence rather than inertia.

Forrester’s work on go-to-market struggles across categories highlights a consistent pattern: organisations that rely on a single channel without testing its true contribution tend to discover the problem at the worst possible time, usually when growth has already stalled.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions