01
The customer avatar is the foundation. The rest is improvisation.
Before spending a single euro on ads, we need to know exactly who buys. Not "who could theoretically buy." Who actually buys: age, situation, motivation, the moment they decide to pay.
A real case, anonymized. A client sells custom-order women's footwear, premium pricing. I went through the CRM, the actual orders from the past 12 months, and what customers say in the comments on Instagram. The real avatar comes out clearly: women between 25 and 35, above-average income, who care about design and pay for something unique. That's the definition. Full stop.
When I showed them the data, the answer was: "but we also have male customers, and women over 45 who buy." Yes, they do. But it's statistical noise. Under 5% of orders. If we allocate the Meta Ads budget to serve those 5% as well, we miss the 95% who pay.
At that point I have two options. Either the client accepts the data and we concentrate budget on the real avatar. Or they insist the avatar "is broader," and then we're already on the wrong road before we even start. I refuse the second case. Not from ego. From experience. I know exactly where it leads: three months later we're spending twice as much for the same number of customers, ROI collapses, I'm the one to blame.
02
Limited budget concentrates. It doesn't scatter.
The most common mistake clients with small-to-medium budgets make: they want to be everywhere. Google Ads, Meta, TikTok, LinkedIn, Pinterest, maybe programmatic. "So people see us."
Another real case. A real estate developer with a complex under construction. Budget €8,000/month for ads. He came in with the request: "I want campaigns on Google, Meta, TikTok and Outbrain."
I walked him through the simple math. The highest purchase intent for an apartment shows up when someone actively searches "apartments for sale in [the area where he's building]" on Google. That's the moment the user raises their hand and says "I'm looking right now, I'm ready to talk."
With €8,000 a month, if you put everything on Search plus a few well-controlled Performance Max campaigns on his inventory, you dominate the SERP in that area. Fewer leads per month, but warm leads. Real intent, conversion rates of 8 to 15%.
Scatter €8,000 across five platforms and you have €1,600 each. Below the minimum viable threshold on any of them. On Meta lead ads you get cheap leads, yes. But the conversion rate of those leads into actual customers is under 1%. You're buying data, not customers.
The client pushed for Meta lead ads because "a developer friend does it and gets leads." I turned the project down. Three months later I ran into him by chance. Budget spent, lots of leads, zero sales. Now he's with another agency. Good.
03
Two stores, two strategies. Never a template.
This is the principle that separates real performance marketing from "I run the same ads for all clients and just swap the pixels." Two stores in the same niche can have completely different strategies, depending on what they sell and how the product gets bought.
Two clients, both fashion eCommerce. Similar budgets.
Store 1 sells women's dresses, focused on occasion dresses. The products solve a concrete problem that already exists in the user's head: "I'm going to a wedding and I need a dress." The user is actively searching. Here we concentrate budget on Google Search Ads and Shopping Ads, with clear segmentation by event occasion. Controlled Performance Max. Meta retargeting for people who visit the site but don't buy.
Store 2 sells custom-order women's footwear. The products don't solve an urgent problem. They're impulse purchases, triggered by visual discovery: "those boots are gorgeous, I'm buying them." The user isn't actively searching. You have to meet her where she scrolls. Here we concentrate budget on Meta Ads and TikTok Ads with strong video creative. Very little Google, only for branded search and search retargeting.
Same industry. Similar budgets. Fundamentally different strategies. Because the nature of the purchase is different. Search vs. Discovery. Need vs. Want. That's what real strategy means, not a template applied to everyone.