GOOGLE ADS KEYWORD TARGETING STRATEGIES: HOW I PLAN BEFORE I BUILD

Most weak accounts I inherit weren't built from a bad keyword list. They were built from no plan at all. Someone opened the campaign wizard, pasted in whatever the Keyword Planner suggested, picked a match type by reflex, and hit publish, then spent the next three months trying to reverse-engineer a structure out of a mess. My google ads keyword targeting strategies exist precisely to avoid that, and the whole discipline lives in one habit: I plan the targeting on paper before I build a single ad group. By the time I'm in the interface, every decision has already been made, and the build is just data entry.

This article is about that plan, the sequence from theory to an operational map. It's not a list of "best keywords," because the best keywords depend entirely on the account. It's the thinking that comes first: how I decide what a keyword is supposed to do, how I group intents, how I assign match types deliberately instead of by habit, and how I leave room for the data to correct me. Get that plan right and the campaign almost builds itself; skip it and you're improvising structure while spending money.

The reason planning matters more here than in most parts of Google Ads is that keyword targeting is the one layer where a small early decision compounds hard. A match type chosen carelessly in week one shapes which queries you pay for in month six. An ad group that mixes two intents poisons the reporting for as long as it exists. So the cost of planning is a couple of hours; the cost of not planning is an account you're perpetually cleaning up.

The keyword targeting theory I plan around

Before any structure, I hold one idea in my head, and it's the core of the google ads keyword targeting theory I work from: a keyword is not a word you bid on, it's a hypothesis about a person. Every keyword you add is a bet that someone, in a particular state of mind, will type something close to it, and that when they do, they'll be worth reaching. Once you see keywords as bets on intent rather than strings of text, everything downstream, structure, match types, negatives, gets easier to reason about.

The second half of the theory is that match types are the width of that bet. Google itself frames the three options as a spectrum of control: exact match gives you the most steering over which searches trigger your ad, while broad match reaches the widest set of related searches and hands the pricing of each one to Smart Bidding. That framing is straight from Google's official documentation on keyword matching options, and I treat it as the physics of the system: exact is a narrow, confident bet on meaning; broad is a wide bet that trusts the machine to find adjacent intent. Neither is "better." They're different widths for different confidence levels.

Put those two together and you get the whole planning principle: I'm sorting intents by how confident I am, then choosing a bet-width to match. High confidence in the intent, narrow match. Low confidence but high potential, wider match with guardrails. This is why the theory has to come before the build, because the structure of the account is just the physical expression of these bets, and if the bets aren't clear in your head, the structure won't be clear on the screen. It's the same reasoning that underpins any sound approach to keyword strategy, applied specifically to the targeting layer.

Step one: sorting intent into tiers

With the theory settled, the first thing I actually write down is intent tiers. I take everything the business could plausibly be searched for and I sort it not by topic but by how close each intent sits to a conversion. This is the single most important move in the plan, because the tier a keyword lands in determines its ad group, its match type, its bid posture, and how patient I'll be with it.

Tier one: transactional intent

These are the searches where the person has already decided and is choosing a provider or a product, "buy," "quote," "near me," "pricing," service-plus-location phrases. I'm most confident about these, so they get the narrowest bets: exact and phrase match, tight ad groups, the first slice of budget. In the theory, these are my high-confidence hypotheses, and I want maximum control over exactly which queries trigger them.

Tier two: commercial-investigation intent

The "best," "vs," "reviews," "alternatives" searches, someone actively comparing but not yet committed. Medium confidence. These get phrase match mostly, sometimes carefully-fenced broad, and their own ad groups so the messaging can speak to comparison rather than closing. I don't fund these as hard as tier one, but they're where a lot of pipeline is built.

Tier three: discovery and broad intent

The wide, category-level searches where intent is real but diffuse. Lowest confidence, highest potential volume. This is the only tier where I'll lean on broad match to let the system find adjacent intent I couldn't brainstorm, and only when there's conversion signal to steer it. Deciding what belongs here versus what's just noise is the same judgment I apply when working out how to choose keywords for Google Ads in the first place, applied at the level of the whole tier rather than the individual term.

The point of tiering isn't tidiness. It's that once an intent has a tier, every other decision about it is nearly automatic. The tier is the strategy for that keyword. That's what turns a vague "keyword list" into a plan with internal logic.

Step two: the operational map

Now the theory and the tiers become a concrete document, the operational map. This is a simple grid I fill in before I open Google Ads, and it's the artifact that makes the build mechanical. For every theme, I decide four things in advance: the intent tier, the ad group it lives in, the match type it launches with, and the negative strategy that fences it.

Match type is decided per tier, never per whim. Tier one launches on exact and phrase because I want steering. Tier three opens to broad only where the account already has conversion data, because Google is explicit that broad match leans on Smart Bidding to price each unique query sensibly, and without that signal you're paying full price for the system's exploration. Planning the match type against the tier, rather than picking it in the wizard, is the difference between a deliberate bet and an accidental one, and it's why understanding the underlying keyword match types in depth is a prerequisite to any real targeting plan, not a footnote to it.

The negatives column is where the map earns its keep. For every theme I write down, up front, what I expect to want to exclude, cross-intent noise, wrong-audience queries, informational searches leaking into transactional groups. I'm not waiting for the account to bleed before I think about negatives; the plan anticipates the noise each tier will attract and pre-loads the defense. Broad-match tiers get the heaviest negative planning because they cast the widest net, and an un-fenced broad keyword is the fastest way I know to burn a budget.

The last column on the map is the ad-group boundary, and the rule is one intent per ad group, no exceptions. The moment two tiers share an ad group, the ad copy can't be sharp for either, Quality Score muddies, and the reporting stops telling you anything clean. The map exists to enforce that separation on paper, where it's cheap to get right, instead of discovering the violation three months into a live account, where it's expensive to unpick.

Step three: closing the loop with data

A targeting plan that never changes is a plan that was wrong and stayed wrong. The last piece of my google ads keyword targeting strategies is the feedback loop, because every keyword was a hypothesis, and the search terms report is where those hypotheses get graded. Once traffic arrives, the report tells me which bets paid off and which intents I misread, and I let it rewrite the map.

Two things happen on every review. Winning queries I didn't anticipate, especially the ones surfacing from broad-match tier-three keywords, get promoted into their own tightly-matched keywords with copy built around them, moving them up the tiers toward tighter control. And the junk each broad keyword drags in gets added to the negative lists I pre-drafted on the map, tightening the fences I already planned. The plan wasn't a prediction to defend; it was a starting structure designed to be corrected.

If you take one thing from how I work: don't build the campaign and bolt targeting on afterwards. Write the map first, sort intent into tiers, assign a match type and a job to every keyword, plan the negatives before the noise arrives, then build. Treated that way, google ads keyword targeting strategies stop being a pile of keywords you're forever cleaning up and become a legible plan where every term has a known purpose, a known width, and a known place, and the search terms report has a clean structure to improve rather than a mess to rescue.