SaaS PPC keyword research

Most SaaS keyword lists get built in the wrong order. Someone exports 400 terms from Ahrefs or Semrush, sorts the sheet by monthly volume, keeps the top 50, and hands the file to whoever runs the account. Six weeks later 70 percent of the spend sits behind three category terms, and the demo requests came from a student, a consultant, and a product manager at the company you are trying to beat.

The list was fine. The sort order killed it.

Sort every keyword into one of four buying-intent tiers

A SaaS search term belongs in exactly one of four tiers, defined by how close the searcher is to an approved budget. The tier decides the bid, the ad, the landing page and what you are allowed to expect from the campaign in reporting.

Those ranges come from mid-market B2B accounts running in North America, and they move a long way by category. Security and data infrastructure sit at the top of the CPC bands, developer tools sit lower, and HR software sits in the middle. Treat the table as a starting prior, then overwrite every cell with your own numbers within 90 days. The wider distribution lives on our SaaS PPC benchmarks page.

Tier 2 is where most SaaS accounts should spend first. Someone typing monday.com alternatives or Asana vs ClickUp has a live shortlist, a problem serious enough to compare two vendors, and in most cases a manager who has already agreed to look at options. That is worth far more than a browser typing best project management software for the first time.

Tier 4 confuses people because it looks like content territory. It is, mostly. Paid works there only when the problem query maps to a product action you can demonstrate in one screen, like deduplicate salesforce contacts or soc 2 evidence collection. Vague problem queries belong in organic search, not in an auction.

One rule cuts through a lot of argument: a pricing modifier promotes a term one tier. Help desk software is tier 3. Help desk software pricing behaves like tier 2, because nobody researches pricing for fun. The same promotion applies to cost, plans, quote and for enterprise.

How to price a tier instead of guessing a max CPC

Work backwards from a gross-margin-adjusted allowable CAC, not from what the auction is charging. The formula has four inputs and takes ten minutes.

Start with annual contract value, multiply by gross margin, decide what share of that first-year gross profit you will spend to acquire the customer, and you have the allowable paid CAC. Then divide by the number of clicks it takes to produce one customer at that tier.

Take a product at 24,000 dollars ACV with 78 percent gross margin. First-year gross profit is 18,720 dollars. If paid search may spend 60 percent of first-year gross profit to win the account, allowable CAC is about 11,200 dollars. Round it to 11,000 and hold that number constant across every tier.

Now price each tier using its own funnel rates.

Read the last column against the CPC table above and the strategy writes itself. Brand clears with room to spare, which is why the brand argument is about incrementality rather than affordability and belongs in branded search defense. Rival demand clears at 41 dollars against a market price of 8 to 30 dollars, so it is the profit centre. Category clears at 7.92 dollars against a market price of 25 to 70 dollars, so at this ACV it loses money at every click.

That single comparison is the whole discipline. You are not asking whether a keyword is relevant. You are asking whether the auction price is below the price your funnel can pay, and the answer is different for every tier in the same account. Run your own version in the SaaS max CPC calculator before you argue about it with anyone.

Two honest caveats. Funnel rates at the bottom of the table are unstable when volume is low, so a category tier priced on 11 opportunities is a guess wearing a suit. And allowable CAC is a policy decision, not a fact: a Series B company with 18 months of runway and a board pushing for growth may set it at 90 percent of first-year gross profit and accept a longer payback.

Where the list actually comes from, and it is not a keyword tool

The best keyword sources in a SaaS account are already inside the company. Keyword tools estimate demand for terms that already exist; your own systems record the language of people who already bought.

Four sources, in the order I would work them:

• The search terms report, 12 months, all campaigns. Export it, strip to queries with at least one click, then run an n-gram count on two-word and three-word fragments. The fragments that repeat across converting queries are your real keyword themes.
• Closed-won records in the CRM. Pull the original query or the self-reported source field from every deal that closed in the last four quarters. You are looking for the words buyers used before they knew your product name.
• Sales call recordings. Gong or any recording tool will let you search transcripts for the phrase ‘we were using’ and ‘we looked at’. Those two phrases produce your competitor list, including the ones nobody on the marketing team has heard of.
• G2 and Capterra category pages. The category names review sites assign to you are the category names buyers search. If G2 files you under Revenue Intelligence and your ads target Sales Analytics, one of you is wrong, and it is not the site the buyer is reading. Vendors buying placement there should read G2 paid advertising for SaaS vendors before adding review-site terms to a search budget.

Keyword tools still earn their place for two jobs: sizing an unfamiliar category and finding query patterns you have never served. Use Ahrefs or Semrush for pattern discovery, not for prioritisation, and remember that Keyword Planner reports a rounded 12-month average rather than the demand you will see next month. Seasonality in B2B software is brutal in late December and again in August.

Support tickets deserve a mention too. The phrasing customers use when something breaks is the phrasing prospects use when they shop for a replacement. A two-hour read through Zendesk or Intercom macros produces better tier 4 candidates than any tool export.

The kill list: query patterns that reliably waste SaaS budget

Some modifiers signal a searcher who will never buy software, and they show up in every SaaS account that has not been cleaned. Add these as campaign-level or account-level negatives on day one.

Two of those are worth arguing about. Free is not always waste: if you run a genuine free tier, free CRM is a real acquisition term, and HubSpot built a business on it. The test is whether the free product converts to paid at a rate you can measure, not whether the traffic exists.

Template is the other one. Template seekers convert badly on a demo form and convert well on a product signup if the template lives inside the product. Canva and Notion make that work because the template is the product surface. If your template is a PDF gated behind a form, you are buying an email address at software CPC prices, which is a losing trade.

The pattern kill list is only half the job. The other half is watching search terms weekly for the first two months and adding negatives before the pattern spreads. Our SaaS negative keyword list ships with roughly 600 pre-built terms by vertical, and the broader failure modes are catalogued in SaaS PPC mistakes that waste budget.

The spreadsheet: six columns, and one of them is decoration

Build the sheet with these columns, in this order, and make the verdict column the only one anyone is allowed to argue about.

Break-even click-to-customer is just estimated CPC divided by allowable CAC. At 11,000 dollars allowable, a 45 dollar click needs one customer per 244 clicks. That is the number the account has to beat, and it is the only threshold that matters.

Observed click-to-customer takes 90 days and a working offline conversion feed to populate honestly. Until you have it, use the tier averages from the pricing table above and mark the column as estimated. Never present an estimate as measured to a CFO who will remember.

Why 40 searches at $45 beats 5,400 searches at $38

Run the two rows from the table above against a month of real spend and the argument ends. This is the single most useful calculation in SaaS paid search, and almost nobody does it before launching.

Spend 2,000 dollars on rival alternatives at 45 dollars a click and you buy 44 clicks. At the observed 0.62 percent click-to-customer rate, that is 0.27 customers a month, or roughly one customer per 3.7 months, at a paid CAC of 7,258 dollars against an 11,000 dollar allowance. It clears.

Spend the same 2,000 dollars on best project management software at 38 dollars a click and you buy 53 clicks. At 0.06 percent, that is 0.03 customers a month: one customer every 31 months, at a paid CAC of roughly 63,000 dollars. It does not clear, and no amount of ad copy testing closes a gap that size.

The high-volume term is not worse because it is high volume. It is worse because volume attracts every competitor with a credit card, which sets the auction price, while intent stays low because the query includes browsers, students, analysts and people writing listicles. Volume is a proxy for competition at least as much as it is a proxy for opportunity.

There is a real limit here, and it is the reason nobody runs a SaaS account on tier 2 alone. Named rival demand is capped by your competitors’ brand volume, which you do not control. A company spending 8,000 dollars a month can live inside tier 2. A company that needs to spend 90,000 dollars a month will exhaust it in week one and has to buy category terms at a worse ratio, which is a budget allocation problem rather than a keyword problem. That trade-off is worked through in SaaS PPC budget allocation.

What changes between an $8K ACV and a $90K ACV

Allowable CAC scales with ACV, so the same keyword is profitable for one company and ruinous for another in the same category. Here is roughly where each tier lands by ACV band.

ACV band	Tiers that usually clear	Tiers to avoid	Sensible monthly floor
Under $3K, self-serve	Brand, integration pairs, narrow problem queries	Category head terms, most versus terms	$3K to $6K
$8K to $25K, hybrid	Brand, alternatives, versus, pricing-modified category	Broad category head terms	$10K to $25K
$30K to $80K, sales-led	All four tiers, with category restricted by industry modifier	Unqualified problem queries	$25K to $60K
$100K+, enterprise	Brand, rival demand, category, plus paid social ABM	Anything without firmographic targeting	$50K+

The self-serve row is the one people get wrong most often. At an 1,800 dollar ACV and 80 percent margin, allowable CAC at 60 percent of first-year gross profit is about 864 dollars, which at a 2 percent click-to-signup rate and a 25 percent signup-to-paid rate means roughly 200 clicks per customer and a max CPC near 4.30 dollars. That rules out almost every category term in software and explains why Calendly and Loom grew on product loops rather than on scheduling software clicks.

At the enterprise end the arithmetic inverts. A 120,000 dollar ACV at 75 percent margin supports an allowable CAC near 54,000 dollars, and a 70 dollar click on data observability platform is cheap. Datadog and Snowflake can pay prices that would bankrupt a 20-person company selling into the same accounts.

Build the list in ten working days

Landing page discipline in step 8 is the part teams skip, and it costs more than any bidding error. A tier 2 click that lands on a homepage converts at roughly a third of the rate of the same click landing on a genuine comparison page. Ad copy that matches the tier matters almost as much, and the patterns that hold up are collected in the SaaS ad copy templates.

Who owns the sheet, and when it gets rebuilt

Give one named person the keyword sheet and give it a review calendar, because keyword lists decay in three predictable ways. Competitors launch and die, your ACV moves, and Google changes what a match type means.

Quarterly is the right cadence for a full re-price, tied to the finance close so win rates are current. Monthly is right for negatives. Weekly is right for the search terms report in the first two months of any new campaign and after any match type change.

Two events should trigger an unscheduled rebuild: a pricing change on your side of more than 15 percent, and a competitor acquisition that removes a brand from the market. When Salesforce or Atlassian absorbs a product you were conquesting, that term’s volume redirects within about two quarters and your best-converting ad group quietly dies. The tooling that spots these shifts early is covered in PPC tools for SaaS teams, and the account-wide view sits on the SaaS PPC and paid ads hub.

Do this before your next budget meeting

Open your search terms report, export the last 12 months, and tag every query that produced an opportunity with one of the four tiers. That exercise takes about two hours and usually shows that 60 to 80 percent of pipeline came from tier 1 and tier 2 while 60 to 80 percent of spend went to tier 3.

Then set your allowable CAC with finance, fill the break-even column, and bring the kill list to the meeting. Arguments about keywords end quickly when the sheet shows the conversion rate a term would need and the rate it actually delivers.