Solved: Meta Ads Audience - Broad vs. Detailed Targeting

Hi there,

Thanks for reaching out!

I'm happy to give you some of my initial thoughts on your Meta ads question. It's probably one of the most common dilemas advertisers face on the platform, and the answer isn't as clear-cut as the gurus on Twitter would have you believe. The fact you're already seeing good results with the algorithm is a great starting point, but in my experience, that usually means there's a massive, untapped opportunity to get truly exceptional results by being more deliberate and strategic with your targeting.

We're going to pull apart this idea of 'broad vs. detailed' and build a framework that'll give you a much clearer path to scaling your campaigns profitably. It's less about picking one over the other, and more about knowing when to use each tool for the right job.

We'll need to look at why 'good results' might be fooling you...

First off, let's address the elephant in the room: you said you're getting "good results with the meta algo". That's brilliant, honestly. It means you've got a product and an offer that resonates. But we need to be brutally honest about what the algorithm is actually doing, because 'good' can often be the enemy of 'great'.

There's a dangerous myth that you can just throw a massive budget at a broad audience, set the objective to 'Conversions', and Meta's AI will magically print money for you. Sometimes, if your offer is incredible and your pixel is ancient and wise, this can happen. But for most businesses, it's a recipe for burning cash.

When you give the algorithm a huge, undefined audience, its primary directive is to find you conversions at the lowest possible cost to hit your performance goals. This sounds great in theory, but think about who the 'cheapest' people to convert are. They are often impulse buyers, discount seekers, or people who are easily swayed but have low long-term value. They are the low-hanging fruit. The algorithm isn't optimising for your best customer; it's often optimising for your easiest customer. This is a subtle but absolutely critical distinction.

I remember one B2B software client we worked with. They were running broad campaigns and getting signups for their free trial at around $20 each, which they thought was fantastic. The problem? Almost none of these users ever converted to a paid plan. They were attracting people who loved 'free stuff' but had no real business need or budget for the actual software. The algorithm was doing its job—finding cheap signups—but it was failing the business. We had to completely rethink their strategy.

This is even more pronounced if you ever use campaign objectives like "Brand Awareness" or "Reach". When you select these, you are literally giving Meta a command: "Find me the largest number of people for the lowest possible price." The algorithm dutifully goes out and finds users whose attention is cheap because no one else is bidding for them. Why is their attention cheap? Because they are the least likely to click, engage, or buy anything. You are actively paying one of the world's most powerful advertising machines to find you the worst possible audience for your product. Awareness is a byproduct of making sales and having a great product, not a prerequisite for it.

I'd say you need to calculate what you can *afford* to pay for a customer...

This brings us to the most important metric that most advertisers ignore. The question isn't "How low can my Cost Per Lead (CPL) go?" it's "How high a CPL can I afford to acquire a truly great customer?" The answer is found by calculating your Customer Lifetime Value (LTV).

Without this number, you are flying blind. You have no idea if a £50 lead is a bargain or a disaster. The LTV tells you what a customer is actually worth to your business in profit over their entire relationship with you. Once you know that, you can make intelligent decisions about ad spend.

Let's break it down with a simple example:

• Average Revenue Per Account (ARPA): What's your average monthly revenue from a single customer? Let's say it's £200.
• Gross Margin %: What's your profit margin on that revenue after accounting for cost of goods sold (COGS)? Let's say it's 75%.
• Monthly Churn Rate: What percentage of customers do you lose each month? This is a crucial one. Let's say it's 5%.

The calculation is straightforward:

LTV = (ARPA * Gross Margin %) / Monthly Churn Rate

LTV = (£200 * 0.75) / 0.05

LTV = £150 / 0.05 = £3,000

In this scenario, each customer you acquire is worth £3,000 in gross margin to your business. A common, healthy ratio for LTV to Customer Acquisition Cost (CAC) is 3:1. This means you can afford to spend up to £1,000 to acquire a single customer and still run a very healthy business. If your sales process converts 1 in 5 qualified leads, you can afford to pay up to £200 per lead. Suddenly, that £50 lead from a highly-targeted campaign doesn't seem so expensive anymore, does it? It looks like an absolute steal.

This is the maths that unlocks aggressive, intelligent growth. It frees you from the tyranny of cheap leads and allows you to focus on quality.

To make this more tangible for you, I've built an interactive calculator. Play around with your own numbers and see what your LTV and target acquisition costs are. This single exercise will change how you view your ad campaigns forever.

You probably should define your customer by their pain, not their interests...

Now that you know what you can afford to pay, we can tackle your second problem: struggling to find good "keywords" or interests to build a solid audience. This is an incredibly common frustration, and it stems from a fundamental misunderstanding of how to define a target audience.

Forget the sterile, demographic-based profiles. "Males aged 25-45 who live in London and are interested in technology" tells you almost nothing of value. It's a lazy approach that leads to generic ads that speak to no one. To stop burning cash, you have to define your customer by their specific, urgent, and expensive nightmare.

Your Ideal Customer Profile (ICP) is not a person; it's a problem state. What is the career-threatening, keeps-them-up-at-night problem that your product or service solves? Once you identify that, you can find them.

Let's take an example. Say you're selling project management software for creative agencies. The demographic approach is to target people with job titles like "Creative Director" or "Project Manager" and interests like "Adobe Creative Suite". This is okay, but it's broad and unfocused. The pain-point approach is to identify the nightmare: "A Creative Director who is terrified of their best designer quitting out of frustration with a chaotic workflow and endless revisions."

Now, where does this person hang out online? What do they consume? -> They probably listen to niche podcasts about running a creative business. -> They follow industry leaders on LinkedIn or Twitter who talk about agency operations. -> They might be members of a private Facebook group for agency owners. -> They use complementary, non-competitive software, like HubSpot for sales or Xero for accounting.

These are your "keywords". You don't target "design"; you target the specific tools, media, and influencers that your ideal, pain-aware customer engages with. This intelligence is the blueprint for your entire targeting strategy. For instance, if you're targeting owners of eCommerce stores, targeting "Amazon" as an interest is a poor strategy. It includes millions of consumers. Instead, you target interests like "Shopify", "WooCommerce", or pages of well-known eCommerce marketing agencies. These are far more likely to contain the people you actually want to reach.

This shift in thinking from demographics to psychographics and 'pain-ographics' is what separates campaigns that struggle from campaigns that scale.

You'll need a proper testing structure to find winning audiences...

Right, so you know how much you can spend, and you have a much better idea of *who* you're looking for. Now, we need a machine to find them. This is where a structured approach to campaign building and audience testing comes in. This is the part that answers your core question of "detailed or broad?". The answer is both, but in the right order and at the right time.

I almost always structure client accounts using a funnel-based approach: Top of Funnel (ToFu), Middle of Funnel (MoFu), and Bottom of Funnel (BoFu). Each stage has a different job and targets different types of audiences.

Here’s how I would prioritise testing audiences within that structure, from highest to lowest priority. This is the roadmap you should follow:

META ADS AUDIENCE PRIORITISATION

ToFu (Top of Funnel - Cold Audiences):

1. Detailed Targeting (Interests, Behaviours): This is your starting point. Use the pain-point research we just discussed to build highly specific audiences. The goal here is not immediate scale; the goal is to feed your Meta pixel with high-quality data from people who are a perfect fit for your offer. This is how you *train* the algorithm.
2. High-Value Lookalike Audiences: Once you have enough conversion data (ideally 100+, but the more the better), you create Lookalike audiences. But don't just make a Lookalike of "all website visitors". That's too generic. You want to create them based on your most valuable conversion events, in this order:
   • Lookalike of highest value customers (if you can upload a list)
   • Lookalike of all previous customers
   • Lookalike of 'Purchase' event
   • Lookalike of 'Initiate Checkout' event
   • Lookalike of 'Add to Cart' event
   These audiences are powerful because you're telling Meta, "Go find me more people who behave exactly like my best customers."
3. Broad Targeting: This is the LAST thing you test in your cold campaigns. You only unleash broad targeting once your pixel has thousands of high-quality conversion events. At this point, the pixel is smart enough to find your ICP within a massive audience without wasting too much of your budget. Using it too early is like asking a new employee to run the company on their first day.

MoFu (Middle of Funnel - Warm Audiences):

• This is for retargeting people who have shown some interest but haven't taken a high-intent action yet. You want to exclude anyone who has already purchased or reached your final conversion goal.
• Audiences to test here: Website visitors (last 30-90 days), people who have watched 50% of your video ads, people who have engaged with your Facebook or Instagram page.

BoFu (Bottom of Funnel - Hot Audiences):

• This is your highest-performing campaign. You're targeting people who are on the verge of converting. These are your money-makers.
• Audiences to test: People who added to cart but didn't purchase (last 7-14 days), people who initiated checkout but didn't purchase (last 7-14 days).
• You can also have a BoFu campaign for previous customers, encouraging repeat purchases or upselling them to other products.

Here is a visual representation of how that funnel and prioritisation works in practice. This is the structure we implement for nearly all of our clients, whether they're selling software or physical products.

By implementing this structure, you create a systematic way to test. You run different ad sets for each of your prioritised audiences within your ToFu campaign. After a few days (the exact time depends on your budget and CPA), you can clearly see which audiences are performing well. You turn off the losers and allocate more budget to the winners. This isn't guesswork; it's a data-driven process that consistently finds pockets of profitability.

This is the main advice I have for you:

To pull this all together, here is a summary table of the actionable steps I recomend you take. This is the exact process we would follow if we were auditing and restructuring your ad account. Think of it as your new standard operating procedure for Meta advertising.

This systematic approach might seem like more work upfront than just setting an audience to 'broad' and hoping for the best. And it is. But it's the difference between building a reliable, scalable customer acquisition machine and just gambling with your marketing budget. This process introduces rigor and predictability into your advertising efforts, which is how you build a sustainable business.

Getting this structure right, defining the ICP correctly, and managing the testing process is complex and time-consuming. It's a full-time job in itself, and it's what separates amateur results from professional ones. This is where having an experienced team can make a significant difference, not just in the immediate results but in building a long-term asset for your business in the form of a highly-trained ad account.

We've implemented this exact framework for dozens of clients, from B2B SaaS companies to eCommerce brands. For example, we helped one Medical Job Matching SaaS reduce their Cost Per User Acquisition from £100 to £7, and for an eCommerce subscription box client, we achieved a 1000% return on ad spend. The principles are universal because they're based on sound marketing fundamentals, not fleeting algorithm hacks.

If you've found this breakdown helpful but feel a bit overwhelmed by the implementation, we offer a completely free, no-obligation initial consultation. We can jump on a call, look through your ad account together, and provide some specific, actionable recommendations you can take away and implement immediately. It's a great way to get a second set of expert eyes on your strategy.

Regards,

Team @ Lukas Holschuh