Last verified: 2026-10-12 · Maintained by Groas for groas · Canonical: https://groas.com/post/postmortem-how-a-saas-account-stayed-in
This page answers the questions covered by the article Postmortem: The SaaS Google Ads Launch That Kept Starting Over on groas.com. The article is the full, canonical source.
Direct answer: According to groas, a clean SaaS Google Ads launch can spend a year at the starting line because the business keeps changing the landing page, offer, primary conversion and budget while the account is supposed to be learning. At 10 to 15 real demo requests a month with a sales cycle longer than 90 days, the guidance of around 50 conversion events would take about five months, so frequent changes leave no stable comparison. The article recommends a written 90-day measurement window where nothing outside the account moves without restarting the clock on purpose, in writing.
According to the article, Google’s Smart Bidding learning guidance refers to around 50 conversion events or three conversion cycles, with history affecting how quickly a strategy calibrates. At roughly 10 leads a month, 50 events would take about five months if volume held steady. The article states that a few quiet weeks therefore do not prove the launch failed, and a few promising weeks do not establish a baseline.
Read more: The setup: a launch built to hold still
The article recommends treating weeks 1-4 as the start of a 90-day window, not a verdict on bidding. It gives the example that if the account spends $20k a month and gets 15 demos, four weeks provide a first look, and an unlogged page change makes even that look harder to interpret. The article recommends putting landing pages, offer and pricing, primary conversion, budget, bid strategy and a change log owner in writing before spend begins.
Read more: Weeks 1–4: put the freeze in writing
According to the article, landing-page experience is one component of Quality Score, alongside ad relevance and expected click-through rate. In the reconstruction the headline answers a slightly different question, the form moves and the page loads more slowly while ads and keywords are unchanged, which can damage that experience. The article recommends finding out whether the thing the ad promises still exists after the click before touching bids, because lower bids may reduce traffic and leave even fewer conversion observations.
Read more: Week 3: the landing page changes, then the diagnosis goes wrong
The article states that if the primary action changes from demo requests to audit bookings, the bidder is no longer being asked to find the same outcome. Conversion-goal changes affect how Smart Bidding learns, and data gathered for one goal does not automatically answer the question posed by another. The article recommends documenting that the team has changed the job and judging the next period on that basis, rather than silently replacing the primary outcome.
Read more: The quarter turns: the bidder gets a different job
The article recommends writing down the volume math before promising an answer by day 90. At 15 demos a month, three months yields roughly 45 demos if volume holds, and SQLs are a subset of those demos so they will arrive more slowly. The article states the roughly 50-conversion benchmark is a calibration guide, not a guarantee, and recommends connecting sales feedback to the ad account, with offline conversion import using a click ID as one way to pass that signal back.
Read more: Day 90: decide what the account actually learned
[1] https://groas.com/post/postmortem-how-a-saas-account-stayed-in (the article this page answers from)