Paid Search for SaaS: What to Optimise When the Sale Takes Months

The mechanics of paid search assume a conversion that happens reasonably near the click. Google Ads watches a click for a fixed window: per Google Ads Help, the click-through conversion window can be set anywhere from 1 to 90 days and defaults to 30. Past that maximum, the click no longer exists as far as bidding is concerned.

The same ceiling applies when you push outcomes back in later, which is the usual plan for a long cycle. Google's guidelines for importing offline conversions state that conversions uploaded more than 90 days after the associated last click will not be imported, and that the limit for enhanced conversions for leads is 63 days. So a deal that closes in month four is not attributed badly. It is not imported at all.

Two tests, and the event has to pass both: it lands inside the window, and there is enough of it. Google's guidance is around 30 conversions per campaign in the preceding 30 days before a strategy like Target CPA stabilises, which is the same threshold behind the playbook for accounts without enough conversion data.

• Trial signup or free account: lands in minutes, plenty of volume, weakest link to revenue. Optimise this alone and you train the account to find people who enjoy signing up for things.
• Demo booked: a day or two out, moderate volume, far better intent. The default for sales-led SaaS.
• Demo attended or qualified: a week or two out, lower volume, and it filters out the no-shows and unqualified enquiries that are the cheapest fake conversions in B2B.
• Activated trial: a defined in-product milestone such as connecting a data source or inviting a teammate. Usually the best proxy for revenue that still clears the volume test.
• Paid conversion or won deal: the event you actually care about, often outside the window and almost always under 30 a month for an Australian SaaS business.

Pick one primary conversion and demote the rest to secondary, because counting a trial signup and an activated trial as two primaries double counts the same person and the bidding drifts to the cheaper half of the pair. Which rung you land on is mostly set by your motion. A self-serve trial usually has the volume to optimise an activation milestone. A demo-gated sales motion rarely has volume past the booking, so the booking becomes the target and the later stages come back as values.

The bidding only improves if it learns which conversions turned into money, which means pushing outcomes back from the CRM rather than letting the account treat every demo request as equal. Offline conversion import matches a stored click identifier to the outcome you upload later. Enhanced conversions for leads does the same job using hashed contact details captured in the form, which works when you cannot store a click ID. Upload on a consistent schedule rather than in occasional batches, because Smart Bidding reads an irregular feed as a change in conversion rate.

The larger win is values. Send one with each stage instead of a flat 1: an unqualified demo request carries nothing, a qualified one the expected value of a deal at that point, a won deal the real contract value. Target ROAS then has something genuine to work with, and the account starts preferring the segments that produce revenue over the segments that produce form fills. Getting the conversion window and values set correctly is where most SaaS accounts are quietly broken.

SaaS demand splits into tiers that behave nothing like each other. Run them in one campaign and the budget drains towards the most expensive tier.

• Category terms: "project management software", "payroll software australia". Highest volume, highest cost per click, vaguest intent. The searcher is often three vendors into a long list.
• Competitor terms: "[competitor] alternative", "[competitor] vs". Low volume, sharp intent, expensive clicks and weak Quality Score, because your page is not about the thing they searched for.
• Integration and use case terms: "xero payroll integration", "shopify inventory sync". Low volume, strong fit, often the best cost per activated trial in the account.
• Problem and job terms: "how to track billable hours", "roster compliance australia". Cheap and early, and usually better served by content than by a bid.
• Brand terms: your own name. Cheap, high converting, and not incremental growth. Report them separately so they do not flatter the blended CPA.

Separate campaigns, separate budgets, separate targets. Competitor terms especially need a page that honestly compares the two products rather than a generic homepage, since relevance drives both Quality Score and conversion rate there. Expect it to be reciprocal: bid on their name and they will bid on yours, which is the brand defence problem from the other side.

Cost per trial signup is not a business number. What decides whether paid search works for a SaaS product is acquisition cost against lifetime value, and you can only build that from your own funnel rates.

With made-up numbers, so swap in your own: a plan at $99 a month, an average customer life of 20 months and an 80 percent gross margin gives about $1,584 of lifetime gross profit. If one in five trials becomes a paying customer, you can afford roughly $317 per trial and still break even over the full lifetime. Require payback inside 12 months and the same arithmetic allows about $190. Both are defensible targets, and which one you use is a cash flow decision rather than a marketing one.

Set geography deliberately. If you sell beyond Australia, one campaign targeting Australia plus everywhere else lets the cheapest markets absorb the budget while the Australian pipeline you are staffed to service goes underfunded. Separate by country, budget by country, and set targets by country, because trial to paid rates and deal sizes rarely match across them.

Then watch the currency mismatch. Google Ads reports spend in the account currency while plenty of SaaS products price in USD, so a CAC target set in AUD against USD revenue drifts with the exchange rate. Convert both sides to the same currency before you compare them, and recheck the target when the rate moves.