Notes from running paid campaigns and building sites for owner-led businesses. No growth hacks, no secret formula. Just the things that decide whether the money comes back.
A cheap click is easy to buy. Broaden the targeting, loosen the placements, and the cost per click drops immediately. It feels like the campaign is improving. It usually is not.
The only number that connects an ad account to a bank account is what you paid for each real enquiry. That figure can rise while clicks get cheaper, because you have started buying attention from people who were never going to buy anything.
So we optimise on enquiries, not traffic. Sometimes that means paying more per click for a narrower, more expensive audience that actually converts. The ad account looks worse on paper and the business does better.
Ask your agency what a lead costs you, not what a click costs you. If they cannot answer, the tracking is not set up.
Most campaigns that fail do not fail at the ad. They fail in the three seconds after the click, on a page that was designed to describe a business rather than to get one thing done.
Two things fix most of it. First, message match: whatever the ad promised should be the first thing on the page, in roughly the same words. A visitor who has to hunt for what they clicked on leaves. Second, one action. Not a menu, a chat widget, a newsletter box and a phone number competing for attention. One obvious next step, repeated.
A homepage is built to serve everyone who arrives from anywhere. A landing page is built to serve one visitor who arrived from one ad. They are not the same job, which is why sending paid traffic to a homepage quietly wastes so much of the budget.
If you are paying for traffic, build the page it lands on. Do not borrow one.
Without conversion tracking, every decision after launch is a guess dressed up as a strategy. You can see which ad got clicks, but not which ad produced the customer who spent four thousand dollars.
That means pixels and the conversions API on the Meta side, conversion actions and call tracking on the Google side, and a form that records where each enquiry came from. Phone calls matter here more than people expect: for a roofer or a real estate agent, most enquiries never touch a form at all.
It is unglamorous work and it happens before a single ad goes live, because retrofitting it means throwing away the first month of data.
If phone calls are how your customers get in touch, untracked calls mean your best campaign looks like your worst.
Google is demand capture. Someone types roof repair near me, and you appear. The intent already exists and you are competing to be the one who answers it. It tends to produce fewer, warmer, more expensive leads.
Meta is demand creation. Nobody opens Instagram looking for a conservatory. You interrupt them with something good enough to stop the scroll. Volume is higher, cost per lead is often lower, and the leads need more follow-up.
If people already search for what you sell, start with Google. If your product is a nice-to-have, a new idea, or something visual, start with Meta. Once one channel is producing at a cost you are happy with, add the other rather than splitting a small budget across both from day one.
Splitting a small budget across two platforms usually means neither gets enough data to optimise.
For most small businesses, social media is not where the sale happens. It is where the sale gets checked. Someone sees your ad, likes the look of it, then opens your profile to see whether you are real.
What they are looking for is not a content calendar. It is evidence: recent activity, finished work, actual customers, a page that does not look abandoned in 2023. A profile that fails that check kills a lead you already paid for.
Which is why we treat the profile as part of the funnel rather than a separate service. Consistent beats frequent. Ten posts that show real work will outperform a hundred that show quotes on gradients.
Before spending more on ads, open your own profile the way a stranger would. Would you hire you?
Every one of the notes above is something we apply on client accounts, not theory borrowed from a blog. If any of it sounded uncomfortably familiar, that is usually worth a conversation.
Bring your ad account, your site, or just a description of what is not working. We will tell you what we would change first.