Keyword Research for PPC: Find What Pays, Not Just Volume

Most keyword research for PPC ends with a spreadsheet: a few hundred ideas from Keyword Planner, sorted by search volume, and a guess at the budget. Then the campaign launches. Many of the keywords turn out to have the wrong intent. Some never get enough searches to run. The budget was built on bid estimates that were never your real cost per click. You find out by paying for it.

The problem isn’t a lack of data. Google, Microsoft and Pinterest all tell you how people search. But their numbers mean different things, they can’t be compared with each other, and none of them tells you what your campaign will get. Reading them correctly takes time, and most teams don’t have it.

Kyō, Kyōsey AI’s AI marketing manager, does this research the way an experienced PPC strategist would. She works across the platforms you advertise on and filters ideas down to what you actually sell. You get themes ready to become ad groups, the months when demand rises and peaks, and a budget range with every assumption shown. When a campaign already runs, she starts from what it has already learnt.

What Keyword Research for PPC Should Actually Answer

Search volume is an input, not an answer. Good research answers four practical questions:

• Which themes are worth running? Themes that match what you sell, that you have a page for, and that have enough demand to measure.
• Where? Some themes have plenty of demand on Google and little on Microsoft. Others belong on Pinterest, where people plan before they buy.
• When? The rise before the peak is often where the opportunity is.
• Roughly at what price? A range you can defend, not a single number from a planner.

Kyō treats the planning tools as evidence about demand, not as a forecast of results. Once a campaign runs, she replaces the estimates with its own data: real cost per click, real conversion rates, real impression share.

What You Get from Kyō’s Keyword Research

Depending on the job, Kyō delivers:

• Keyword themes grouped into ad groups, one theme per ad group with one landing page, so every ad fits every keyword in it.
• A match type recommendation per theme, based on intent and on whether your conversion tracking and bidding are ready for broader matching.
• A negative keyword list from wrong-intent ideas and from search terms that spent without converting, including the singulars, plurals and synonyms that negatives don’t catch on their own.
• Brand and competitor terms kept separate, with competitor terms used only with your go-ahead and with trademark rules in mind.
• Monthly demand per theme, read from several years of months rather than one annual average.
• A low, expected and high budget for a stated period, with the clicks needed, the cost range and the market’s ceiling spelled out.
• A plain statement of what each number is: which platform, which market, and whether it is a count, a rounded average or a range.

That last point matters more than it sounds. Google’s figures are rounded averages that fold close variants together. Microsoft’s are counts for one market. Pinterest’s are broad buckets. Kyō never adds them up or compares them directly, and she tells you when a number is an estimate.

What A Theme Looks Like When Kyō Hands It Over

Here is an illustrative example for a regional HVAC company. The placeholders stand in for the figures Kyō pulls from your own account and market:

Theme:          Emergency AC repair (illustrative example)
Ad group:       AC repair - emergency
Landing page:   /ac-repair/
Match types:    phrase by default; exact for the proven top-intent term
Negatives:      diy, how to, car, cars, auto, job, jobs, salary, parts
Demand:         Google, your service area, English: [monthly searches]
Seasonality:    rise starts [month], peaks [month]
Budget [period]: low $[a] / expected $[b] / high $[c]
Assumptions:    conversion rate [x]%, impression share [y]%,
                CTR [z]%, CPC $[low]-$[high]
Source note:    Google figures are rounded exact-match averages;
                Microsoft demand checked separately

Every theme comes with the same parts, so you can approve, adjust or drop it without rebuilding the logic yourself.

Starting A New Search Campaign

Kyō first writes down your market, your offer, your pages, your brand and competitor names, and what you don’t sell. That last list is where a lot of wasted spend begins. Keyword ideas follow the seed words, not your business: an HVAC company seeding with air conditioning will be offered DIY repair guides, car AC parts and HVAC job listings.

If your account has run search ads before, Kyō starts from the search terms that actually converted, the best seeds there are, because someone already paid to prove them. She sets the campaign’s real locations and language on every request, so the numbers aren’t secretly worldwide or in every language. Then she filters, groups and prices the result.

Adding Microsoft Ads, Pinterest Or ChatGPT Ads to A Campaign That Already Works

This is where most of the value is. A campaign that already runs has learnt which themes sell and which searches waste money. Starting a new platform from a fresh seed list throws that away. Kyō carries it over, adapted to how each platform works:

For B2B advertisers on Microsoft, she also suggests LinkedIn profile targeting by company, industry and job function. On social platforms, she turns converting themes into creative angles and targeting ideas, and tells you that part is judgement rather than published practice.

Finding The Keywords You’re Missing

A coverage audit asks: what relevant demand exists that none of your keywords or search terms covers? Kyō pulls a year of keywords and search terms with cost, conversions and impression share, generates ideas from your website and your best-converting terms, and reports the missing themes. The same audit finds the opposite problem:

• Spend on keywords with little or no search volume.
• High-volume keywords where you rarely show.
• Keywords Google has made inactive for low volume.

She splits brand from non-brand and compares Google and Microsoft theme by theme, using shares rather than raw numbers.

Planning Around Seasonality

An annual average hides the shape of the year. Kyō reads several years of monthly data per theme to find the peak and, more usefully, when the rise begins. She checks Microsoft’s recent weekly and daily figures to see whether this year’s ramp has started. For home, gifts, food and events, she looks at what is trending on Pinterest, where planning starts earlier, when Pinterest gives access to its trend data. And she raises cost assumptions for peak months, because a busy month brings more competition as well as more searches.

Sizing A PPC Budget from Keyword Data

Kyō builds the budget from the result you want, not a round number. She works out the clicks needed to reach your target conversions at a realistic conversion rate, prices them as a range from a quiet auction to a busy one, and caps them at what the market can supply. For campaigns on automated bidding such as Demand Gen with a target CPA or ROAS, she checks the daily budget is large enough for the bidding to learn.

You get low, expected and high scenarios, rechecked after a few weeks of live data. Budget moves between platforms are judged on cost per conversion, never on cost per click alone.

Who This Is for

• Business owners who want to know whether there is demand for what they sell, and what reaching it will cost, before they spend.
• In-house marketers launching a campaign or adding Microsoft, Pinterest or ChatGPT Ads to one that already works.
• Agencies that want thorough, consistent keyword research on every account without a strategist spending a day in planner exports.

Kyō’s skills are the written-down methods of experienced strategists, more than 100 person-years of agency experience, so she follows them every time. She connects to Google Ads, Microsoft Advertising, Meta, TikTok, GA4, Search Console, Shopify and more. Nothing goes live without your OK, and your data stays yours.

How to Get Started

Kyōsey AI is not another tool. It’s your A-Team: an AI team that does the work, not just recommends it, with elite agency performance for $6/day (see Pricing).

The best way to see it is with your own account. Book a demo, bring a Google Ads campaign that already works, and watch Kyō carry what it has learnt into Microsoft Ads, Pinterest or ChatGPT Ads: themes, negatives, timing and a budget you can plan around.