Solved: New lead generation campaign - How to start?

Hi there,

Thanks for reaching out.

Happy to give you some initial thoughts on your new lead generation campaign. The questions you're asking are common, but the answers aren't as simple as just picking an audience type. Honestly, seeing a CPL of $60 right out of the gate tells me there's probably a bigger issue at play than just choosing between broad or detailed targeting. To really fix this and build something that can scale, we need to look at the entire process, not just the buttons you're pressing in Ads Manager.

A high CPL like that usually points to one of a few things: your targeting is way off, your ads aren't connecting, or—and this is the most common one I see—your offer just isn't compelling enough for the audience you're trying to reach. It's a classic sign of a lack of real demand for what you're putting in front of people.

We'll need to look at your offer first...

This is probably not what you want to hear, but the number one reason I see campaigns fail, especially with high costs, is the offer itself. Before you spend another dollar trying to find the "perfect" audience, you need to be brutally honset with yourself about what you're actually offering them in exchange for their details.

A great ad campaign can't fix a bad offer. If people aren't biting, it's because what you're giving them isn't valuable enough to solve an urgent problem they have. I see so many founders and marketers create what they think is a great product or lead magnet, but they've never truly gotten inside the head of their customer to understand their real-world frustrations.

You need to stop thinking in terms of demographics like "men aged 25-40 who like business". That tells you nothing. You need to define your customer by their nightmare. What is the specific, urgent, expensive problem that keeps them up at night? What's the career-threatening issue they're facing?

For example, if you're selling project management software, the nightmare isn't 'needing better organisation'. It's the Head of Operations being terrified of a major project deadline being missed, costing the company a huge client and putting her job on the line. That's a pain point. That's a nightmare. Your offer, your lead magnet, your entire ad campaign needs to be the aspirin for that specific headache.

Is your lead magnet a generic "Download our free guide to X"? Or is it a "Free 5-minute calculator to identify your biggest time-wasting tasks"? One is a nice-to-have vitamin. The other is an urgent painkiller. People will pay (with their email address and attention) for a painkiller. They'll ignore a vitamin nine times out of ten. That's the difference between a $6 CPL and a $60 CPL.

Think about it:
-> Who is the specific audience you're trying to acheive? Not a broad category, but a specific persona.
-> What is their most urgent, costly problem right now?
-> How does your lead magnet provide an immediate, tangible piece of the solution to *that* specific problem?

If you can't answer these three questions with absolute clarity, then no amount of targeting tweaks will save the campaign. You need an offer so relevant and so valuable that the right person feels foolish for not taking it. Only then should you move on to worrying about who sees it.

You probably should rethink your targeting strategy from the ground up...

Okay, let's talk about your direct question: Advantage+ (broad) vs. custom audiences. A lot of people get this wrong, and it costs them a fortune.

Here's an uncomfortable truth: when you set a campaign to be completely broad, especially on a new account without much data, you're essentially telling Facebook's algorithm: "Go find me the cheapest possible people to show this ad to." And the algorithm is very good at its job. It will go and find users who are not in demand. Why? Because their attention is cheap. They are the least likely to click, least likely to engage, and definately the least likely to ever buy anything. You're effectively paying one of the world's most powerful advertising machines to find you the worst possible audience for your offer.

Advantage+ can work, but only when your pixel is already very "smart". That means it has hundreds, if not thousands, of high-quality conversion events from the *exact* type of people you want more of. Without that data, it's just guessing in the dark, and it will always favour cheap impressions over quality leads.

So, how do you start? You have to give the algorithm guidance. You need a structured approach to testing audiences so you can feed the pixel with the right kind of data. I usually prioritise audiences in a specific order, moving from the top of the funnel (cold traffic) to the bottom (warmest traffic). This is how you build a sustainable system.

For an account like yours, here's the structure I would start with:

This structure is designed to do one thing: find out which pocket of people on Facebook actually responds to your offer. You aren't lumping everything together. You are running a clean test. By seperating your audiences into distinct ad sets, you'll be able to see clearly which group gives you the best CPL. The one that works is the one you double down on. The ones that don't, you turn off.

This is the work you have to do to *teach* the pixel who your customer is. Once you have a winning detailed targeting audience that's bringing in leads at a much better cost, you then use that data to build your more powerful audiences: retargeting and lookalikes. I've seen so many clients completely neglect this. They might make a "website visitors" audience but forget all the higher-intent ones like "people who initiated checkout" or "people who spent the most time on the pricing page".

The goal isn't to start broad. The goal is to start specific, find what works, and then use that success to build broader, more powerful audiences later. We ran a campaign for a B2B software client where, by using this exact methodology of starting narrow and then building powerful lookalike audiences from the initial data, we got them 4,622 registrations at just $2.38 per registration. That doesn't happen by just turning on Advantage+ and hoping for the best.

I'd say you need a proper testing framework...

This brings me to your other question: how do you know what creative works? You've made three ads, which is a good start. But how you test them is what matters.

As you can see in the table above, the idea is to put the same set of ads in each of your test ad sets. This is important. It means you are comparing apples to apples. If Ad Set A with your 3 ads performs much better than Ad Set B with the same 3 ads, you know the difference was the *audience*, not the creative.

Once you've found a winning audience (let's say Ad Set A is getting you leads for $15 while the others are at $60+), you can then turn off the losing ad sets and focus your budget on the winner. Now, inside that winning ad set, you can start to properly test your creatives. After a few days, Facebook will likely have started favouring one of your three ads. That's your initial winner.

Now you can start iterating. Take the winning ad and create variations. Test a new headline. Test a different image. Test a video instead of an image. But you only test *one thing at a time*. This is how you learn. If you change the headline and the image at the same time, you'll never know which change made the difference.

Your ad copy itself needs to speak directly to that nightmare we talked about. Stop selling features. Sell the solution to the pain. A good framework to use is Problem-Agitate-Solve (PAS).

Problem: State their pain point directly. "Are your cash flow projections just a shot in the dark?"
Agitate: Poke the bruise. Make them feel the pain more acutely. "Are you one bad month away from a payroll crisis while your competitors are confidently raising their next round?"
Solve: Present your offer as the clear, easy solution. "Get our free projection template and turn uncertainty into predictable growth in under 10 minutes."

Here’s how you could structure your three different hooks to test:

By testing these different psychological angles, you'll quickly learn what resonates most with your audience. Is it their pain? Their aspirations? Or their need for proof? The data will tell you.

You'll need to understand the numbers behind the CPL...

Let's go back to that $60 CPL. Right now, it feels expensive because it's just a number without context. The real question isn't "how low can my CPL go?", it's "how high a CPL can I afford to acquire a great customer?". The answer to that lies in a number you absolutely must know: your Customer Lifetime Value (LTV).

If you don't know this, you are flying completely blind. Let's run through a quick, hypothetical calculation. You'll need to plug in your own numbers, but the formula is the same.

1. Average Revenue Per Account (ARPA): What's a customer worth to you each month? Let's say it's £300.
2. Gross Margin %: What's your profit margin on that? Let's say it's 75%.
3. Monthly Churn Rate: What percentage of customers do you lose each month? Let's say it's 5% (meaning the average customer stays for 20 months).

The calculation for LTV is: (ARPA * Gross Margin %) / Monthly Churn Rate

So, in our example: (£300 * 0.75) / 0.05 = £225 / 0.05 = £4,500 LTV.

Each customer you acquire is worth £4,500 in gross margin to your business over their lifetime. Now we have context.

A healthy business model often aims for a 3:1 LTV to Customer Acquisition Cost (CAC) ratio. This means you can afford to spend up to £1,500 (£4,500 / 3) to acquire a single paying customer.

Now let's work backwards. If your sales process converts, say, 1 in 10 qualified leads into a paying customer, you can afford to pay up to £150 per lead (£1,500 / 10).

Suddenly, your $60 (£48) CPL doesn't seem quite so terrifying, does it? It actually looks potentially profitable, *if* the leads are high quality and convert at that 1-in-10 rate. The problem isn't necessarily the cost. The problem is the *unknown value* of the leads you're generating. Without the LTV math, you're just optimising for a cheap CPL, which often leads to low-quality leads that never convert, meaning your effective CPL is infinite.

I've run campaigns for software companies where we happily pay over $20 for a B2B lead from LinkedIn, because we know they convert into customers worth thousands. I've also run campaigns where we've gotten trials for just $7 on Meta. The "right" price is entirely dependent on your business model. You need to do this maths. It will change your entire perspective on what a "good" campaign looks like.

This is the main advice I have for you:

To pull this all together, getting your lead generation campaign to work isn't about finding a single magic button. It's about building a systematic, logical process. You need to stop randomly testing things and start thinking like a scientist. Define your problem (the offer), form a hypothesis (the audience and ad), run a clean experiment (the testing structure), and measure the results against a meaningful metric (your LTV-driven allowable CPL).

Here’s a table summarising the actionable steps you should take right now.

As you can probably tell, this is a lot more involved than just setting an ad live. It requires a deep understanding of strategy, maths, and human psychology. It’s a full-time job to get this right and optimise it continuously, and it's where having an expert in your corner can make all the difference between burning cash and building a predictable engine for growth.

If you’d like an expert pair of eyes on this and want to walk through a proper strategy tailored to your business, we offer a free, no-obligation initial consultation where we can look at your setup together and give you a clear path forward.

Hope this helps!

Regards,

Team @ Lukas Holschuh