Search ads: buying intent

The vending machine with settings

Search advertising is the closest thing marketing has to a vending machine. Someone types their need, in their own words, at the moment they feel it — and you appear with the answer. No other channel sells demand this ripe. That's why you fund it first (Lesson 4 will make this a rule), and why it runs out first: search spend is capped at however many people are searching.

But the machine has settings, and the defaults quietly spend your money on the wrong searches. This lesson is those settings: structure, match types, negatives, and the landing page. None of them requires spending anything to get right.

Structure is relevance

A search account has four layers: campaign › ad group › keyword › ad. One campaign per goal and budget. One ad group per tight theme, so the ad can mirror the search. That mirroring is not cosmetic — Lesson 1 showed you that rank is bid × predicted quality, and nothing raises predicted quality like an ad that repeats the searcher's own words back to them and lands on a page that keeps the promise.

Structure, in other words, is relevance. And relevance is a discount.

Match types: how literally your money is taken

When you buy a keyword, you're really telling the auction how loosely to interpret it. Three settings:

• Exact — [physiotherapie graz]. Searches with that meaning, tightly. Maximum control, minimum reach.
• Phrase — "physiotherapie graz". The phrase's meaning inside longer searches: "gute physiotherapie graz empfehlung" qualifies.
• Broad — physiotherapie graz, no punctuation. The theme, as the algorithm interprets it — which can drift a long way from what you meant, at your expense.

The taught default for small budgets: exact and phrase first. Broad match is not evil — it finds searches you'd never have thought of — but it needs conversion data to steer by and negatives to fence it, and in week one you have neither. Broad is a tool for month three.

Negative keywords: the searches you refuse to pay for

Every keyword list attracts searches you never wanted: free, jobs, training, what is. A negative keyword tells the auction: never enter me for this. Add the obvious ones on day one, before any money moves.

Then — the single highest-value habit in search advertising — read the search-terms report weekly. It lists the actual searches your ads appeared for. Fifteen minutes, once a week: every irrelevant search becomes a negative, every surprising-but-good search becomes a keyword. Your campaign gets cheaper every week it runs, or it would if you were running it — for now, the habit goes in the plan.

The click is a promise your page must keep

Quality Score — the platform's 1–10 read on each keyword — is built from expected clickthrough rate, ad relevance, and landing-page experience. That third one drags Module 4 back into the room: the searcher clicked on a promise, and the page must be that promise, immediately.

Watch what message match is worth in euros, using Storkflow's canonical numbers. Lena pays about €13 a click for "client onboarding software". In the homepage era, the site converted 2.2% of visitors to trials:

€13 ÷ 0.022 = €591  per trial start

Her Module 4 landing page's early baseline is 3.4%:

€13 ÷ 0.034 = €382  per trial start

Same keyword, same bid, same €13 click — and the page difference is worth €209 per trial. Her target CAC from Module 1.5 is €340; at 2.2% the channel is hopeless, at 3.4% it is close, and at 13 ÷ 340 ≈ 3.8% it breaks through the target. (One honesty flag, per the pack: 3.4% is from a fortnight with too few conversions to call a proven number. It's a planning input, labelled as such.) The lesson in one line: the landing page is a price cut, and message match is how you claim it.

Tomas's draft build, end to end

Herzog Physio, built on screen in a free Google Ads account, everything saved and paused. His patients search in German (Module 6 canon), so the keywords are German; translations in brackets are for you, not the account.

Campaign: Check-up Buchungen – Graz — goal: Check-up bookings, geo-targeted to a radius around Graz.

Ad group 1 — physiotherapie graz: [physiotherapie graz], "physiotherapeut graz". Ad group 2 — sport & rücken: [sportphysiotherapie graz], "physiotherapie rückenschmerzen graz" (back pain).

Negatives from day one: kostenlos (free), gratis, jobs, stellenangebote (vacancies), ausbildung (training), gehalt (salary), übungen (exercises — informational searches are Module 5's job, not this budget's).

Responsive search ad, group 1 — you supply up to 15 headlines and 4 descriptions; the system assembles combinations. His lines carry the four jobs: mirror the keyword, state the differentiator, show proof, ask for the action.

Descriptions (≤90 chars): "Wenn Physiotherapie Ihnen nicht hilft, sagen wir es ehrlich. Check-up €80, online buchbar." and "Termine diese Woche an zwei Standorten in Graz. Buchen Sie online, ohne anzurufen."

Landing page: herzog-physio.carrd.co — the Module 4 booking page. Not the homepage. The homepage lists nine pages of services at equal weight; the searcher asked one question.

Pin sparingly, if at all — pinning a headline to position 1 buys control at the cost of the system's ability to find better combinations.

One more setting worth naming: brand bidding. If competitors bid on "herzog physio", a small exact-match brand campaign is cheap insurance. Nobody currently does, and his organic and map results rank first — so he skips it, and writes down why.

What about buying informational searches?

Your Module 6 keyword plan classified intent. Buy the transactional tier first — "physiotherapie graz" before "rückenschmerzen was tun" ("back pain what to do"). Informational searches are real people with real problems, but they are content's job to win (Modules 5 and 6), not your first €500's. Paying transactional prices for informational curiosity is the most common way small search budgets die.

---