Most paid advertising advice starts with the platforms. That is the wrong end.
Start with arithmetic instead, because the arithmetic decides whether any of this works before a single ad runs. You need three numbers: what a customer is worth to you, how often you close one, and what a lead costs in your category. If those three do not resolve in your favour, no amount of campaign optimisation rescues it.
What a Click Actually Costs
Across 13,474 search advertising campaigns running between 1 April 2025 and 31 March 2026, the average cost per click was $5.42 US, the average conversion rate was 8.18%, and the average cost per lead was $66.69 US (sourced from WordStream).
The spread by category is far more useful than the average. Physicians and surgeons averaged $40.04 US per lead. Home and home improvement, $90.92. Legal services, $131.63.
Two caveats worth stating plainly: those are US campaigns priced in US dollars, and they are averages hiding enormous variation. Canadian costs differ. Treat them as a sense of scale, not a quote.
Now do the arithmetic. At a $90 cost per lead, a $1,000 monthly budget buys roughly eleven leads. Close one in four, with a customer worth $2,000, and that works comfortably. Close one in twenty, with a customer worth $400, and it does not, and it never will.
That calculation takes ten minutes and it is the single most useful thing in this article. Do it before you fund anything.
Google Search: For People Already Looking
Google Search is the channel for businesses where customers actively search for the thing you sell. Plumbers, dentists, lawyers, accountants, emergency anything.
The intent is already there, which is why it is expensive and why it converts. You are not persuading anyone to want a plumber. You are being visible at the moment they already do.
If you run one channel and you are a local service business, run this one. A few things that matter more than bid strategy:
Negative keywords. Without them you pay for "plumber salary," "plumber apprenticeship" and "how to fix a tap yourself." This is the most common budget leak I see and the easiest to plug.
Location targeting set to presence, not interest. The default setting includes people merely searching about your area from elsewhere. Change it.
A landing page that matches the ad. Sending paid traffic to your homepage is the most reliable way to make a small budget vanish.
Conversion tracking that actually fires. An account optimising toward nothing is an account spending toward nothing.
Before you spend anything at all, check whether you can win the same searches without paying. The Maps 3-Pack guide covers the free version of the same visibility.
Meta: For People Who Are Not Looking Yet
Facebook and Instagram do a different job. Nobody opens Instagram intending to buy a service. You are interrupting, which means the creative carries the weight that keyword intent carries on Google.
This suits visual, discretionary and locally social businesses: restaurants, boutique fitness, salons, home renovation, anything where a photograph does the persuading.
What it demands in return is creative volume. One ad running for four months stops working, because the same people keep seeing it. If you cannot produce new material regularly, Meta will be frustrating and Google will not.
Judge it on cost per enquiry, not on likes, reach or engagement. Those metrics are real but they are not revenue, and a campaign can look busy while producing nothing.
LinkedIn: Expensive Per Click, Cheap Per Right Person
LinkedIn has the highest cost per click of the three by a wide margin, which makes it look indefensible until you consider who is on the other end of the click.
If you sell to other businesses, with a deal size in the thousands and a decision maker you can describe by job title, industry and company size, the targeting is precise in a way the other two cannot match. If your deal size is small or your buyer is a consumer, it is the wrong room.
One practical note: content from a named person's profile tends to land better than the same content from a company page, particularly for a business nobody has heard of yet. People engage with people.
The One Rule That Saves the Most Money
Do not add a second channel until the first one has been profitable for three consecutive months.
The most expensive pattern in small paid budgets is three channels running at once, none of them properly, all of them splitting a budget too thin to gather usable data on any one. Each channel individually looks like it might work if you just gave it a bit more. None of them get enough to prove it either way.
Three months matters because one good month is noise. Seasonality, a lucky week, one unusually large customer. Three consecutive profitable months is a pattern, and a pattern is the only thing worth scaling.
If your budget is under a few hundred dollars a month, the honest answer is that paid search will struggle to gather enough conversion data to optimise against at all. That money usually does more in fixing the reasons you are invisible organically first, which lowers what paid has to carry afterwards. When the arithmetic does work, what running the ads involves is written out.
What does your arithmetic say?