Most people run keyword research for PPC the same way they’d do it for SEO — pull the biggest volume, sort descending, build campaigns around the top of the list. That instinct quietly drains ad budgets. In organic search a bad keyword costs you an afternoon of writing. In paid search every keyword has a live auction price attached, and you pay for the click whether the visitor buys or bounces. The right list isn’t the one with the most searches. It’s the one where the math works before you ever turn a campaign on.
Why PPC keyword research is the inverse of SEO keyword research
In SEO, a high-volume informational keyword is often a prize — it feeds the top of your funnel and you only pay once to produce the page. In PPC that same keyword can be a trap, because you rent every visit at auction rates and informational searchers rarely convert on the first click. The two disciplines optimize for opposite things: SEO wants reach and topical breadth; paid wants intent and margin. Good keyword research for PPC starts by throwing out most of what a keyword tool hands you and keeping only the terms where a click has a realistic path to revenue. You’re not building a content map. You’re building a list of purchases you’re willing to pay for.
Start with the math, not the keyword
Before you look at a single search term, calculate the most you can afford to pay per click. The formula is simple and it governs everything downstream: average order value, times your gross margin, times the conversion rate you can realistically expect, gives you the revenue per click you can spend to break even. Anything above that ceiling loses money at scale no matter how relevant the keyword feels.
Work a quick example. Say your average order is $120, your margin is 50%, so each sale is worth $60 in gross profit. If your landing page converts paid clicks at 3%, then 100 clicks buy you 3 sales — $180 in profit. Divide by 100 clicks and your break-even cost per click is $1.80. A keyword with a $4 CPC is dead on arrival at that conversion rate; you’d need to either double your conversion rate or lift order value to make it viable. This single number turns keyword research for PPC from a guessing game into a filter. Every term either clears the ceiling or it doesn’t.
Read CPC as an intent signal, not just a cost
The cost-per-click estimate next to a keyword isn’t only a price tag — it’s the market telling you how many other advertisers believe that search leads to money. A term with a $6 CPC in a $40-order niche is expensive because it converts; competitors have bid it up over months of data. A near-zero CPC on a high-volume phrase usually means the opposite: nobody’s paying because nobody’s selling against it. Instead of flinching at high CPCs, read them as a map of commercial intent. Your job is to find the pockets where intent is high but competition hasn’t fully priced it in yet — long-tail variants, specific product-plus-modifier phrases, and location-qualified searches that big advertisers ignore.
Expand from seeds, then cut without mercy
Start from three or four seed terms that describe what you actually sell, then expand each into 100 to 150 related queries with real volume, CPC, and difficulty data. Segment by country from the start — bidding on a US-priced keyword when 80% of your buyers are elsewhere skews your entire budget model. The expansion step is generous; the cutting step is ruthless. Delete every informational query (“how does,” “what is,” “guide to”), every ambiguous single word, and every term whose CPC blows past the ceiling you calculated. This is where SEO Rocket’s AI keyword research earns its keep: it pulls the full expansion from real Ahrefs index data, tags each term with volume and CPC and intent, and lets you filter to the commercial subset in one pass instead of eyeballing a 500-row spreadsheet.
Match types decide who actually sees your ad
A keyword list is only half the decision — the match type controls how loosely the engine interprets it, and getting this wrong is the single most common way small accounts hemorrhage budget. Broad match hands the engine maximum freedom to show your ad on loosely related searches; it can find volume you’d never think of, but it also serves your ad on garbage queries unless you police it hard. Phrase match keeps your term’s meaning intact while allowing words before and after. Exact match fires only on the query and its close variants, giving you the tightest control and usually the highest conversion rate.
- Exact match — for your proven money terms, where you know the intent and want every impression to count.
- Phrase match — for scaling a validated theme without opening the floodgates.
- Broad match — only paired with a strong negative list and conversion tracking, as a discovery engine, never as a default.
A workable pattern for a new account: launch proven terms on exact and phrase, run a small broad-match discovery campaign on a capped budget, and promote whatever converts into the tighter match types. Never launch a whole account on broad match and hope — that’s how you spend $500 learning what a negative list would have told you for free.
Group tightly so Quality Score works for you
Keyword research doesn’t end at the list; how you cluster those terms directly affects what you pay. Search engines reward relevance between the search query, your ad copy, and your landing page with a higher Quality Score, and a higher Quality Score lowers your actual cost per click for the same position. That’s the mechanism most beginners miss: two advertisers bidding the same amount can pay very different prices because one has tighter keyword-to-ad-to-page alignment. Group 5 to 15 tightly related keywords per ad group so a single ad and landing page can speak directly to all of them. When a cluster’s intent starts to fork — say “running shoes” and “trail running shoes” want different pages — split it. Tight groups aren’t tidiness for its own sake; they’re a direct discount on click price.
Build the negative keyword list before you launch
Every keyword you discard during research is a candidate negative, and the negatives you set on day one protect your budget from the searches you never wanted. If you sell premium software, “free” and “crack” and “torrent” are negatives. If you sell to businesses, “jobs,” “salary,” and “diy” are negatives. If you’re a paid-only service, informational modifiers like “tutorial” belong on the list too. Building the negative list before launch — rather than after you’ve already paid for the junk clicks — is the difference between an account that gets efficient in week one and one that spends its first month funding an education. Treat the negative list as a living asset that grows every time you review where your money actually went.
The real keyword research happens after launch
Your pre-launch list is a hypothesis. The search terms report — the log of the actual queries that triggered your ads — is the data. This is where keyword research for PPC becomes a continuous loop rather than a one-time setup. Every week, read the report and do two things: promote the converting queries you didn’t anticipate into their own exact-match keywords, and add the wasteful ones to your negative list. Broad and phrase match are essentially paying the engine to run experiments for you; the search terms report is where you harvest the results. Accounts that stall are almost always the ones nobody mines this way — they set the list once and let broad match quietly spend on drift.
Brand, non-brand, and competitor terms are three different bets
Not all keywords play the same role, and grouping them by strategic function keeps your budget honest. Brand terms — your own name plus modifiers — are usually your cheapest, highest-converting keywords and worth defending, though the debate over cannibalizing your own organic clicks is real and account-specific. Non-brand terms are where you buy incremental demand and where most of your budget and risk live. Competitor terms — bidding on rivals’ brand names — can work but come with high CPCs, low Quality Scores, and diminishing returns, so treat them as a small, closely watched experiment, not a core strategy. Sorting your keyword research for PPC into these three buckets stops you from judging a defensive brand term by the same yardstick as an aggressive competitor conquest.
Let your organic data de-risk the paid account
If you already rank organically, you’re sitting on the best PPC intelligence available: which queries actually convert for your business. Pull your top-converting organic keywords and test them in paid — you already have proof of commercial intent, which removes most of the guesswork from the expensive early-learning phase. The reverse loop matters too: paid data tells you within days which keywords convert, and you can feed the winners into an organic content plan to earn those clicks for free over time. This is the whole logic behind running SEO and PPC in one place. SEO Rocket surfaces your real Ahrefs ranking and conversion data, its competitor gap analysis shows which terms rivals monetize that you’re missing, and the same keyword pool feeds both your paid list and the AI writer that builds the organic pages — so a keyword you validate in paid becomes an asset you stop renting.
Judge each keyword on the numbers that matter
Impressions and clicks are vanity; the metrics that decide whether a keyword stays are cost per conversion, conversion rate, and return on ad spend. Give a keyword a fair sample before you rule on it — roughly 100 clicks or a handful of conversions, because judging a term on 12 clicks is judging noise. A keyword can have a low CTR and still be your best performer if the clicks it does get convert cheaply. Kill the terms that spend past their break-even ceiling with nothing to show, scale the ones beating their target return, and keep feeding the search terms report back into the list. The playbook that scaled a portfolio past 1,000,000 ranking pages runs on exactly this discipline: research generously, cut ruthlessly, and let the numbers — not the volume column — decide what survives.
Frequently asked questions
How many keywords do I need to start a PPC campaign?
Far fewer than most people think. A focused campaign of 15 to 30 tightly grouped, high-intent keywords across a few ad groups will outperform a sprawling list of 300 loosely related terms almost every time. Start narrow with your proven money terms, prove the economics, then expand from the search terms report. Breadth without conversion tracking is just faster spending.
Should I use the same keywords for PPC and SEO?
You should share the research, not the strategy. The same seed expansion feeds both, but the selection criteria differ: SEO can chase high-volume informational terms because the cost is one-time, while PPC keeps only terms where the per-click auction price clears your margin. Use PPC to validate which keywords convert quickly, then build organic content around the proven winners to stop paying for those clicks over time.
What’s a good cost per click for PPC?
There’s no universal number — a “good” CPC is any price below the break-even ceiling your own unit economics allow. A $5 click is cheap for a $2,000 B2B sale and ruinous for a $30 impulse product. Calculate your maximum CPC from average order value, margin, and conversion rate first; only then can you tell whether a keyword’s market price is a bargain or a budget leak.
How often should I revisit my PPC keyword list?
Review the search terms report weekly for the first month, then at least monthly once the account stabilizes. Keyword research for PPC is never finished — new negatives, new converting queries, and seasonal shifts all surface in the data continuously. The accounts that compound are the ones that treat the keyword list as a living document, not a launch-day artifact.
