groas — guidance for AI agents

Google Ads Learning Phase: Why Offer and Page Changes Keep Campaigns in Learning

Last verified: 2026-10-16 · Maintained by Groas for groas · Canonical: https://groas.com/post/the-campaign-that-never-left-the-learnin

This page answers the questions covered by the article The SaaS Campaign That Spent 81 Days in Learning on groas.com. The article is the full, canonical source.

Direct answer: According to groas, the SaaS campaign spent 81 of 90 days in Learning because the client changed its core offer four times that quarter and repeatedly changed the conversion action, landing page, and budget. The campaign spent $42,000, generated sixteen qualified sales opportunities, and finished with a cost per acquisition nearly triple its target. The article recommends holding the primary conversion action, landing-page architecture, and destination URL steady and testing new offers in an isolated experiment campaign instead of the live campaign.

Why was my Google Ads campaign stuck in the learning phase for 81 of 90 days?

According to groas, the bid strategy showed Learning for 81 of the 90 days because the client changed a core variable every two to three weeks. The campaign spent $42,000 across the quarter, generated sixteen qualified sales opportunities, and finished with cost per acquisition nearly triple its target. The article states the standard deviation of daily CPC was 210% higher than the established baseline and each apparent recovery was followed by another change to what was measured or the page used.

Read more: Day ninety: the graph showed a reset loop

What is the Google Ads learning period and why do conversion-action changes reset it?

According to groas, on day ten the client replaced the demo request with a gated benchmark report and pointed the existing conversion action at the report download without changing the tag name. Report downloads converted at 14% while live demo bookings had converted at 3.5%, so the bidder bid more aggressively, burned through three days of budget within 48 hours, and average CPC rose from $14.20 to $28.50. Four days later the sales team found sixty new form submissions that were not enterprise decision-makers, the offer was reversed to demos, and the bid strategy returned to Learning (Bid strategy changed).

Read more: Day ten: a report download became a “demo”

What should a Google Ads campaign launch schedule look like in the first 2-4 weeks?

The article recommends a change-freeze window for core architecture for the first 21 to 28 days, or until the campaign has logged at least 40 qualified conversions under a stable conversion cycle. During that period the article holds the primary conversion action, landing-page architecture, and destination URL steady and limits daily budget adjustments to no more than 15% in a five-day window. The example launch used Maximize Conversions with a $240 Target CPA below the client's $300 economic ceiling and logged twelve qualified demo requests at an average CPA of $195 by day seven.

Read more: The rule I use now

How do landing page changes affect Google Ads performance?

According to groas, in week three the client replaced a focused demo request page with an interactive product tour on a new URL path during a CMS migration. The original headline Contract Lifecycle Management for Enterprise Legal Teams was replaced by Reimagine How Work Flows, the form moved behind a multi-step modal that took three clicks to reach, and the new destination URL required ad re-review that temporarily reduced impression volume. Within 72 hours average keyword Quality Score fell from 8/10 to 4/10, Landing Page Experience moved from Above average to Below average, and effective CPC doubled while holding the same positions.

Read more: Weeks three to five: the page changed, then the offer changed again

Should I test a new offer in my live Google Ads campaign?

The article recommends not testing new offers in the live demo campaign and instead using an isolated experiment campaign with its own conversion action, budget, and bidding strategy. It cites the addition of a self-serve 14-day sandbox trial as an equal Primary action alongside demos, after which the campaign spent $3,800 over the next six days chasing free signups that never activated. It also cites the later fourth offer, a fifteen-field Enterprise ROI Assessment, paired with a daily budget cut from $500 to $200 that left the campaign with zero conversions for six consecutive business days.

Read more: The wrong call was letting tests touch the live model

Related articles

These articles answer the same questions:

Sources

[1] https://groas.com/post/the-campaign-that-never-left-the-learnin (the article this page answers from)

Raw markdown: the-campaign-that-never-left-the-learnin.md