If you run a product-led SaaS, your Google Ads account should be optimizing toward activated sign-ups and product-qualified leads — not raw trial starts, not form fills, and definitely not clicks. The single highest-leverage change most PLG accounts can make is moving the conversion they bid to downstream, from "someone gave us an email" to "someone actually used the product." Get that one thing right and the rest of this post is detail.
I manage paid search for a lot of SaaS companies, and the PLG accounts almost all arrive with the same symptom: cheap sign-ups, flat revenue. The dashboard looks great — cost per trial is down, conversion volume is up and to the right — but new paid customers are not moving. That gap is not a bidding failure. It is a measurement failure, and it is entirely fixable.
The sign-up is the wrong conversion
Here is the trap. You drop a conversion tag on your free-trial thank-you page and point Smart Bidding at it. Google does exactly what you asked: it finds you the cheapest possible trial sign-ups. The problem is that the cheapest sign-ups are also the worst — tire-kickers, throwaway emails, people who will never log in a second time. In most PLG funnels, the majority of free trials never activate, and a large share of those that do never pay.
When you optimize to the sign-up, you are explicitly telling the algorithm that all of those users are equally valuable. They are not. A trial that creates a project, imports data, and invites a colleague is worth fifty trials that bounce after the welcome screen. Bidding to the sign-up flattens that difference to zero, and your CPA looks fantastic while your revenue does nothing.
Define activation before you touch a bid
The fix starts off the ad platform entirely. Before you change a single campaign setting, define your activation event: the in-product action that most strongly predicts a user becomes and stays a paying customer. This is product analytics work, not ads work. Look at users who converted to paid and find the early behavior they share — the aha moment.
- Collaboration tools: inviting a second teammate, or a third person accepting an invite.
- Analytics / dev tools: connecting a data source or firing the first real API call.
- Project / workflow apps: creating a first project and completing the core action inside it.
Pick the event with the cleanest correlation to retention, then treat that as your true conversion. A user who hits it is a product-qualified lead. Everything before it — the click, the visit, the sign-up — is just the on-ramp. This is the same discipline that separates a healthy account from a bleeding one in our B2B SaaS CAC benchmarks: you cannot benchmark or bid to a number you are measuring at the wrong point in the funnel.
Closing the loop: enhanced and offline conversions
Once you know your activation event, you have to get it back to Google, and there is a timing problem: the click happens today, but activation might happen three days later, after the user has wandered off and come back. The bridge is the Google Click ID. Capture the GCLID (and enhanced-conversion identifiers like a hashed email) at sign-up and store it against the user record.
Then, when that user activates, fire the conversion — ideally server-side — and stamp it with the stored GCLID via an offline conversion import. Now Google knows that this specific click led not just to a sign-up but to a real activated user days later. With enhanced conversions for leads and offline conversion import wired up, you can even feed conversion values: a higher value for activation, a higher one still for first payment. Switch the campaign to a value-based strategy like Maximize Conversion Value or tROAS, and Smart Bidding starts hunting for users who use and buy the product, not users who abandon a tab.
This is the highest-ROI plumbing in a PLG ad account, and it is the first thing I check in a Google Ads audit. If the only conversion firing is a front-end trial tag, the account is flying blind no matter how good the bidding looks.
Structure campaigns around the funnel a keyword attracts
PLG does not mean "no sales." It means self-serve is the default path and a human gets involved only when the deal is big enough to justify one. Your campaign structure should mirror that. Different keywords pull different funnels, so let intent decide the offer:
- Self-serve / bottom-funnel terms ("free [category] tool", "[competitor] alternative", "sign up for X") point at the free trial and optimize to your activation event. This is the engine of PLG growth.
- High-ACV / enterprise-intent terms ("[category] for enterprise", "[category] security / SSO", "[category] pricing for teams") point at a "talk to sales" or demo offer and optimize to qualified opportunities, not trials.
Blending those under one campaign and one bid target is a classic mistake — the algorithm averages across two completely different value profiles and serves neither well. Split them, give each its own conversion goal, and you get a self-serve machine and a sales-assist machine running in parallel instead of one muddy compromise.
When PLG needs a sales-assist layer
Pure self-serve has a ceiling. As your average contract value climbs, the accounts worth the most are exactly the ones that want a conversation before they commit — security review, procurement, multi-seat rollout. That is where the sales-assist layer earns its keep. On paid search it shows up as that second campaign group: fewer keywords, higher CPCs, a demo or contact-sales offer, and bidding tied to pipeline value pulled from your CRM rather than to trial starts.
The trick is to keep the two motions measured separately. Self-serve campaigns answer to cost per activated user and trial-to-paid rate. Sales-assist campaigns answer to cost per qualified opportunity and pipeline created. If you judge a demo-request campaign by its (lower) trial volume, you will kill the thing that brings in your biggest accounts. For the broader attribution picture across both motions — and why last-click will mislead you here — a consultant who specializes in SaaS is worth more than a generalist agency that treats every funnel like a lead-gen form.
The bottom line
Running Google Ads for PLG SaaS comes down to one principle applied relentlessly: bid to in-product value, not to the front door. Define the activation event that predicts revenue, close the loop with enhanced and offline conversions so Google can see it, structure campaigns around the funnel each keyword attracts, and add a sales-assist layer where deal sizes justify it. Do that and your cheap sign-ups stop being a vanity metric and start being a leading indicator of real, paying users.
This is most of what I do day to day across a portfolio of SaaS accounts. If you want a second set of eyes on whether your PLG account is optimizing to activation or just to email captures, that is exactly what a managed engagement or a one-off audit is for — and you can see the kind of work it produces in our case studies.