Hi there,
Thanks for reaching out!
That's a really sharp question, and it gets to the heart of what separates successful campaigns from the ones that just burn cash. A lot of people treat the 'objective' setting as just another button to click, but understanding what's happening 'under the hood' is the difference between guessing and actually having a strategy. I'm happy to give you some of my thoughts on this.
The short answer is yes, the system is actively filtering and prioritising, but it's a lot more sophisticated than just looking at someone's recent search history. It's about predicting future behaviour based on trillions of data points. Let's get into it.
TLDR;
- When you choose a 'Conversions' objective, the ad platform's algorithm doesn't just show your ad to your entire target audience. It builds a profile of users who are most likely to take your desired action (e.g., purchase, sign up) and actively prioritises showing the ad to them first.
- Running 'Reach' or 'Brand Awareness' campaigns is often a massive waste of money. You're telling the algorithm to find the cheapest people to show ads to, who are almost by definition the least likely to ever buy anything.
- The algorithm is a prediction engine, not a mind reader. It needs the right inputs: a clearly defined target audience, a compelling ad, and most importantly, a high-value, low-friction offer. A bad offer will fail no matter how clever the algorithm is.
- This letter includes a flowchart visualising the difference between campaign objectives and an interactive calculator to help you understand your own business numbers, like Customer Lifetime Value (LTV).
We'll need to look at the biggest myth first: Paying to find non-customers...
Before we get into how conversion campaigns work, we need to talk about how they *don't* work. And the biggest mistake I see people make, day in, day out, is choosing the wrong objective. Here is the uncomfortable truth about awareness campaigns on platforms like Meta or Google Display.
When you set your campaign objective to "Reach" or "Brand Awareness," you are giving the algorithm a very specific command: "Find me the largest number of people inside my target audience for the lowest possible price."
The algorithm, in its infinite wisdom, does exactly what you asked. It looks at your audience—let's say it's 'small business owners in the UK'—and it thinks, "Okay, who within this group is cheapest for me to show an ad to?" It seeks out the users who are least likely to click, least likely to engage, and absolutely, positively least likely to ever pull out a credit card. Why? Because those users are not in demand. Nobody else is bidding for their attention, so their attention is cheap. You are actively paying the world's most powerful advertising machine to find you the worst possible audience for your product.
Think about it. If a user has a history of clicking on ads, engaging with brands, and buying things online, their attention is valuable. Other advertisers are bidding for it, which drives the cost up. The algorithm knows this. So when you say "find me cheap eyeballs," it deliberately avoids those valuable users. It's a false economy. You might get a low Cost Per Mille (CPM), but you're showing your ads to people who have demonstrated through their behaviour that they don't convert. It's like setting up a stall in a library to sell air horns. You might reach a lot of people, but you're in the wrong place with the wrong crowd.
I'd say you should see it as a prediction engine...
So, when you select a conversion goal—like 'Sales', 'Leads', or even 'Add to Cart'—you're giving the algorithm a completely different, and much smarter, instruction. You're saying, "Don't just find me anyone in this audience. Go through all the data you have and find me the specific people who look like they are about to buy something like what I'm selling."
The system then becomes a prediction engine. It's not just filtering based on interests. It's building a complex probability model in real-time. It looks at hundreds, if not thousands, of signals for every single user in your audience:
- Historical Conversion Data: Has this user purchased from ads like yours before? Have they filled out lead forms recently?
- On-Platform Behaviour: Do they click 'Shop Now' buttons? Do they engage with business pages or just with photos from their friends? -> This is why you must setup your conversion events properly. The algorithm needs this data.
- Website & App Data (via Pixel/SDK): What other websites have they visited? Have they added items to carts on other e-commerce sites in the past 24 hours?
- Contextual Signals: What time of day is it? What device are they on? (e.g., someone on a desktop computer at 2 PM on a Tuesday might be in a 'work research' mode, while someone on a mobile at 9 PM on a Sunday might be in a 'sofa shopping' mode).
- Ad-Specific Data: It also looks at who has already converted from *your specific ad* during the 'learning phase'. It identifies common traits among your first few customers and then aggressively seeks out more people who share those traits.
It takes all these signals and gives each user in your audience a 'conversion score'. This score represents the probability that this specific person will take the action you want them to take, right now. When it's time to show an ad, the algorithm doesn't just randomly pick someone from your audience. It shows it to the person with the highest conversion score who is available at that moment. This is fundementally different from just finding cheap impressions. It's finding valuable actions.
This is why you'll often see that a conversion campaign has a much higher CPM than a reach campaign. You're paying a premium to get in front of these high-intent users. But that's the point. You're not paying for eyeballs; you're paying for results. It's far better to pay £50 to reach 1,000 potential buyers than it is to pay £5 to reach 10,000 people who will never, ever buy from you.
You probably should focus on what you can control...
Now, the algorithm is incredibly powerful, but it's not a magic wand. It can only work with the raw materials you give it. This is where your strategy comes in. Your job is to feed the machine the best possible ingredients. There are three main levers you control:
1. The Audience (The 'Who'): This is your initial targeting. While the algorithm will find the best people *within* this group, you need to provide a sensible starting point. If you sell high-end B2B software, targeting people interested in "video games" is probably not going to work, no matter how good the algorithm is. You need to define your customer not by demographics, but by their pain. What is the expensive, urgent, career-threatening nightmare they are trying to solve? Once you know that, you can find them. Are they in certain Facebook Groups? Do they follow specific industry leaders on LinkedIn? Do they use complementary software tools you can target as an interest?
2. The Creative (The 'What'): This is your ad copy and your visual. It needs to speak directly to that pain point. Don't sell "FinOps software"; sell "the feeling of opening your AWS bill and smiling because you know where every pound went." Your ad's only job is to get the right person to stop scrolling and think, "That's me. They understand my problem." A generic ad gets a generic response (i.e., nothing).
3. The Offer (The 'Why Now'): This is, without a doubt, the most important piece of the puzzle and the number one reason campaigns fail. The algorithm can find the perfect person and your ad can perfectly describe their problem, but if your offer is weak, they won't act. The "Request a Demo" button is the graveyard of B2B advertising. It's high-friction and low-value for the prospect. You're asking for their time in exchange for a sales pitch. It's an arrogant ask. Your offer must provide undeniable value *upfront*. For a SaaS company, this is a free trial (no credit card). For an agency, it could be a free, automated audit tool. For us, it's a free 20-minute strategy session where we actually solve a problem for them. You have to give value to get value.
A weak offer cant be saved by a powerful algorithm. It's like having the world's best delivery driver, but the package is empty. The delivery will be flawless, but the customer will be dissapointed.
You'll need to know your numbers, or you're flying blind...
This all leads to the final peice of the puzzle. If you're going to pay a premium for high-intent users, you need to know what a user is actually worth to you. Otherwise, how do you know if your campaign is profitable? People get obsessed with a low Cost Per Lead (CPL), but it's a vanity metric. A £5 lead that never converts is infinitely more expensive than a £250 lead that becomes a £10,000 customer.
The key is to understand your Customer Lifetime Value (LTV). This tells you the total profit a typical customer will bring to your business over their entire relationship with you. Once you know your LTV, you can determine your maximum allowable Customer Acquisition Cost (CAC).
Here's the basic formula:
LTV = (Average Revenue Per Customer Per Month * Gross Margin %) / Monthly Customer Churn Rate
Let's say you run a SaaS business:
- Average Revenue Per Account (ARPA): £200/month
- Gross Margin: 80% (0.80)
- Monthly Churn Rate: 5% (0.05)
LTV = (£200 * 0.80) / 0.05 = £160 / 0.05 = £3,200
A healthy business model often aims for an LTV:CAC ratio of 3:1 or better. In this case, you could afford to spend up to £1,067 (£3,200 / 3) to acquire a single new customer and still have a very profitable business. If your sales team converts 1 in 10 qualified leads, you can now afford to pay up to £106 per lead. Suddenly, that £75 lead from a targeted LinkedIn campaign doesn't seem so expensive, does it? It looks like a bargain. This is the maths that unlocks aggressive, intelligent growth.
I've included a simple calculator below so you can play with your own numbers. This is often an eye-opening exercise.
This is the main advice I have for you:
To wrap it all up, the system isn't just a simple filter. It's a highly complex prediction and auction system designed to find the outcome you ask for. Your job is to give it the right instructions and the best possible raw materials to work with. If you get that combination right, it can feel like a superpower. If you get it wrong, it's just a very efficient way to set money on fire.
| Actionable Recommendation | Why It Matters | How to Implement |
|---|---|---|
| Always Use a Conversion Objective | This instructs the algorithm to find high-intent users who are likely to take action (buy, sign up), not just cheap viewers. It focuses your budget on potential customers. | In your ad platform (e.g., Meta Ads Manager), select 'Sales', 'Leads', or another relevant conversion goal. Ensure your pixel/tracking is correctly installed. |
| Define Your Customer by Their 'Nightmare' | Generic demographic targeting leads to generic ads. Targeting based on a specific, urgent pain point makes your message powerfully relevant and pre-qualifies your audience. | Interview your best customers. What was the exact problem they had before they found you? Use that language in your ads and target interests/behaviours related to that problem. |
| Create a High-Value, Low-Friction Offer | The offer is the #1 reason campaigns fail. An irresistible offer that delivers instant value makes conversion easy. "Request a Demo" is not a compelling offer. | Replace "Request a Demo" with:
|
| Calculate Your LTV and Allowable CAC | This frees you from the trap of chasing cheap, low-quality leads. Knowing what a customer is worth allows you to confidently invest in acquiring the right ones. | Use the formula (ARPA * GM%) / Churn Rate. Aim for a 3:1 LTV to CAC ratio to determine your maximum spend per new customer. |
It's not just about setting up an ad and hoping for the best. It's about building a complete system where your audience, creative, offer, and campaign objective all work together to feed the algorithm the right signals. That's where professional help can often make a huge difference. With years of experience, we can help identify the best strategies, implement the entire process, and manage the ongoing optimisation that's needed to truly scale.
If you'd like to chat more about your specific situation, we offer a free, no-obligation initial consultation where we can take a look at your account and strategy together. It's usually super helpful for potential clients to get a taste of the expertise they'd be getting.
Hope this helps!
Regards,
Team @ Lukas Holschuh
