Most advice about Google Ads for SaaS turns out to be advice about Google Ads with the word SaaS attached to it. Match types, negative keyword lists, ad copy tests, landing page speed. All of that applies, none of it is specific to selling software, and none of it explains why an account that looks correctly built can still spend a quarter’s budget on signups that never turn into customers.
The thing that makes this category different is not in the keyword plan. It is that the conversion your account optimizes toward is not the conversion that pays you. A free trial signup happens minutes after the click, on your own site, where a tag can see it. Revenue happens weeks later, in the billing system, after a person has decided to keep using the product. Every automated part of the account learns from the first event and knows nothing at all about the second.
The conversion you report is the conversion you buy
A modern Google Ads account is a machine for buying more of whatever you have told it worked. That is not a complaint about the platform. It is the stated design, and Google is direct about the dependency between its two most automated instruments. On its page about broad match, Google Ads Help says: “It’s critical to use Smart Bidding with broad match.” The reasoning given is that broad match widens which auctions you enter, and the bidding strategy is what decides which of those auctions are worth entering.
Follow that one step further and the SaaS problem appears on its own. Broad match is steered by Smart Bidding. Smart Bidding is steered by the conversions you send it. So if the conversion you send is a free trial signup, the system will do exactly what you asked and get very good at buying free trial signups, including every one that was never going to pay. The waste people blame on broad match is usually a faithful execution of a conversion definition nobody revisited.
The second half of the problem is volume. Google’s current page for Target CPA tells you that you must set up conversion tracking to use the strategy, and it sets no numeric floor on how many conversions you need first. That absence gets filled by folklore, and every advertiser you ask will quote a different threshold with equal confidence. What is actually true is narrower and more useful: a bidding strategy learns from the events it receives, so an account sending a handful of paid conversions a month is asking a statistical system to work from a sample that would not persuade a person either.
That leaves two honest options. Send the system the event that matters and accept slow learning, or send it an earlier and more frequent event and accept that you are optimizing toward a proxy. Both are defensible. Choosing between them without saying which one you chose is what produces the quarter nobody can explain.
Your billing system knows something the account does not
Before restructuring campaigns or rewriting ads, get the paid conversion back into the account. This is the piece of work that changes what every other decision is based on, and it is usually the piece that has never been done.
The mechanism has been in the platform for years. Google Ads Help describes it plainly: “Google Ads provides you with unique IDs, called Google Click ID (GCLID), for every click that comes to your website from an ad.” You store that ID with the lead or the trial account it produced, and when that account later converts to a paying customer, you send the ID back with the conversion and its value. Google now points advertisers toward enhanced conversions for leads as the upgraded version of the same idea, using hashed first-party data such as an email address alongside the click identifier, so treat GCLID as the mechanism to understand rather than the current default to implement.
What matters is not which of the two you pick. It is that the trial-to-paid step stops being invisible. Once a paid conversion carries a value, the account can be told that ten signups from one query are worth less than two from another, which is a sentence it currently has no way to hear. In our experience the trial signup is the conversion already configured when we open a SaaS account, and it is frequently the only one.
This is also not a marketing task on its own. Somebody has to store an identifier against a customer record and schedule the job that sends conversions back. That belongs to whoever owns the billing system, which is why it tends to sit undone through several rounds of campaign optimization that were all guessing.
Narrower targeting does not buy a cheaper click
Once the loop is closed, the ordering question becomes answerable rather than ideological. The usual instinct is to buy the term that names your category, because it has the demand and it feels like the whole market is behind it. The usual correction is to buy narrower comparison queries instead, on the theory that they cost less. The second piece of advice is more sensible than the first and it is not correct for the reason people give.
Data table
| Query | Estimated cost per click |
|---|---|
| email marketing software (201,000/mo) | $23.47 |
| mailchimp alternatives (2,400/mo) | $24.04 |
| hubspot alternatives (1,900/mo) | $36.71 |
Three queries from one category, pulled from Semrush’s US database on August 4, 2026. “email marketing software” draws 201,000 searches a month at an estimated average cost per click of $23.47. “mailchimp alternatives” draws 2,400 a month at $24.04. “hubspot alternatives” draws 1,900 a month at $36.71. Each figure is a Semrush estimate of recent auction prices, not a rate card you can hold anyone to.
Monthly demand falls by roughly 99 percent across those three queries while the cost of a click rises by more than half. Narrowing your targeting buys you a different intent, not a discount. A query that names a competitor draws bids from every vendor that competes with them, which is one reason an auction like that clears where it does.
That is an argument for a priority order, and the order is not about price. Fund the queries your own pages can answer better than a review site can, and where the outcome is legible in your own data: brand, competitor and alternative queries, and the specific job your product does stated the way a buyer would state it. The category head term is a volume purchase against every funded competitor in your market, and a paid conversion imported from billing is the only thing that will ever tell you whether it worked. Buy it once you can measure it. Our note on PPC lead quality covers how to judge what arrives.
What to stop doing
Four habits recur in the SaaS accounts we open, and all four follow from the same root.
- Stop treating the trial signup as the goal. It is a step. Naming it the conversion tells every automated system in the account that it is the finish line, and they will believe you.
- Stop running broad match on a signal you have not closed. Google’s own guidance ties broad match to Smart Bidding, and Smart Bidding is only as good as the events feeding it. Broad match on a trial signup conversion is an efficient way to buy volume you cannot evaluate.
- Stop reading cost per acquisition as though it meant cost per paying customer. If the conversion action is a signup, the column labelled cost per acquisition is the cost of a signup. Two campaigns with identical numbers in that column can be far apart on what they actually cost per paying customer, and nothing on the screen says so.
- Stop launching every campaign type at once. Search, Performance Max and Demand Gen all draw on the same conversion signal. Starting them together on a signal you are not confident in multiplies one measurement problem across three surfaces and leaves you unable to say which one to keep.
The pattern underneath all four is the same: each one lets an automated system act on a definition that was set once, early, by whoever installed the tag.
The numbers worth putting on the report
Cost per trial belongs on the report, because it is real and because you have it. It just cannot be the number that decisions are made from.
The set that survives a question from a finance team starts with cost per paying customer by campaign, which requires the imported conversion and is the reason the loop comes first. Next to it, put the trial-to-paid rate by campaign, since that is what separates two campaigns whose cost per trial looks identical. Then measure the whole thing on the cohort of clicks from a given month rather than on what converted during that month, because a paid conversion arriving in August from a July click will otherwise be credited to the wrong period and flatter whatever you happened to change in August.
Give the measurement enough time to be readable. A payback figure calculated on a few weeks of a product sold on an annual contract is arithmetic, not evidence. If you want the campaign work run against these numbers rather than signup counts, that is what our Google Ads management and wider PPC work is built around.
FAQ: Google Ads for SaaS
Is Google Ads worth it for SaaS companies?
It is worth it when the paid conversion can be measured, and it is a gamble when it cannot. Search advertising reaches people who have already described their problem in their own words, which suits software with a defined job. The risk is not the channel, it is running it against a trial signup and discovering late that the cost per paying customer was never in view.
Should SaaS companies bid on free trial or demo keywords?
Both, in separate campaigns. The two intents are not interchangeable: a trial query asks to try the product, a demo query asks to be shown it, and the accounts they produce can differ in value. Which way they differ for you is something to measure rather than assume, and running them together averages the answer away before you can see it.
Does broad match work for SaaS?
It is worth using once your conversion signal reflects revenue. Before that it widens what you buy without improving how you judge it. Google Ads Help states that it is critical to use Smart Bidding with broad match, and Smart Bidding learns from the conversions you send. Broad match against a trial signup conversion buys more trial signups, which is the outcome you asked for and often not the one you wanted.
Are competitor brand names a sensible buy?
Usually yes, with the understanding that the click can be expensive. In Semrush’s US database on August 4, 2026, “hubspot alternatives” carried an estimated average cost per click of $36.71 against 1,900 searches a month. What decides whether that price is worth paying is whether the paid conversion is in your account, because cost per signup will not tell you.
How many conversions does Smart Bidding need before it works?
Google’s current Target CPA page does not publish a number. It states that you must set up conversion tracking to use the strategy and leaves the volume question open, so any specific threshold quoted to you is convention rather than documentation. The practical version is that a strategy learns from the events it receives, so an account sending very few paid conversions a month should expect slow learning and should not read early swings as results.
What should a SaaS company measure in Google Ads?
Cost per paying customer by campaign, and trial-to-paid rate by campaign. Read both by the month the clicks happened in, not the month the conversions landed in. Cost per trial stays on the report as a diagnostic. It is a fine early-warning number and a poor basis for deciding where budget goes.
COSEOM Team
COSEOM is a multilingual B2B growth agency. Since 2008 the COSEOM team has run international SEO, generative engine optimization, paid media and demand generation programs for B2B software and technology companies, in English, German, Spanish and Portuguese. This account publishes our research, playbooks and product guides. More about COSEOM