--- title: "Six Weeks in Learning: The SaaS Campaign That Kept Resetting Itself" description: "A B2B SaaS Google Ads postmortem: how one conversion-goal change started a six-week cycle of reactive edits, and the week-one decision that could have stopped it." image: "https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb477254ba684cf486d41b_5bbd3271-5221-4fc4-ac14-6970853c511a.png" --- September 29, 2026 • min read # Six Weeks in Learning: The SaaS Campaign That Kept Resetting Itself ![Young man with curly hair wearing a black shirt outdoors against green foliage background.](https://cdn.prod.website-files.com/6821efca072e48f6f495a47e/68562d390107b3921a6e3d68_1743932904108.jpg) **Alexander Perleman**, Head Of Product @ groas Ex-Goldman Sachs and Stanford Computer Science alex@groas.ai [LinkedIn](https://groas.com/post/the-account-that-spent-six-weeks-stuck-i#) ![Cover image for: Six Weeks in Learning: The SaaS Campaign That Kept Resetting Itself](https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb477254ba684cf486d41b_5bbd3271-5221-4fc4-ac14-6970853c511a.png) Forty-two days into a Google Ads campaign, the Bid Strategy Status column still said **Learning**. Roughly $22,000 had gone out the door. The client blamed Google, the agency blamed the sales team, and the CMO wanted to know whether automated bidding worked for business software at all. This is an **explicitly typical composite**, not an account I personally managed. It is the version of this failure that B2B growth teams run into when a slow sales cycle meets a fast-moving dashboard. The [Google Ads learning phase](https://support.google.com/google-ads/answer/6263057) was not a mysterious punishment. The team kept changing the conditions under which the bid strategy was supposed to learn. What looked like one six-week algorithm problem was a sequence of human decisions. The first decision, on day eight, mattered most. Everything after it was an attempt to repair what the team thought they were seeing. #### The setup: a demo campaign with a slow feedback loop The company sold a workflow automation platform with an average contract value of $14,000 and a 45-day sales cycle. It launched a targeted Search campaign on a $15,000 monthly budget. The bid strategy was Maximize Conversions with a $160 Target CPA, and the primary conversion goal was a “Book a Demo” form submission. That goal made sense. A completed demo request was closer to a sales conversation than a casual page visit. It also meant the campaign would not produce a neat stack of conversions every morning. Buyers might click an ad, look at security certifications, check pricing, talk to a colleague, and return four to nine days later to book. Google notes that [Smart Bidding needs conversion data and time to calibrate](https://support.google.com/google-ads/answer/7074566). Here, the team had neither much volume nor a completed view of its recent clicks. By day seven, the campaign had recorded six demo submissions at an average CPA of $310. That was nearly double the $160 target, and executive nerves snapped. Six demos were a reason to investigate. They were not enough to declare that the bid strategy had failed, especially while buyers who had clicked during launch week could still convert. **The first week’s dashboard was an incomplete account of the first week.** #### Day 8: the wrong conversion became the goal At an emergency review, the team decided the model needed more conversions. The idea had a certain tidy logic: if demo requests were scarce, give Smart Bidding something easier to find. The media buyer changed the campaign’s conversion setup so that a whitepaper download and a product tour pageview joined demo bookings as primary actions. That was the wrong call. Google’s distinction between [primary and secondary conversion actions](https://support.google.com/google-ads/answer/11464374) is not a filing preference. Primary actions feed the core Conversions column and can guide Smart Bidding. Secondary actions remain visible for observation without becoming bidding goals. ![Diagram showing demo requests mixed with lower-value download signals in Google Ads Smart Bidding.](https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb477254ba684cf486d41f_d05d74b0-f936-4156-b31e-af1971871729.png) Before the change, a conversion meant someone had asked to speak with the company. After it, a conversion could also mean someone had viewed a tour page or taken a whitepaper. Those actions might be useful to track, but they did not show the same buying intent. The team had not solved a shortage of demo data. It had **changed what counted as success**. That distinction matters more than the badge. Smart Bidding now had to work with a different set of goals, while the humans looking at the account could no longer read its top-line conversion count as a count of demo requests. As [signal engineering discussions](https://www.searchenginejournal.com/how-to-fix-google-ads-smart-bidding-issues-with-primary-vs-secondary-conversions/517674/) point out, mixing lower-value actions into primary goals can send optimization toward the easier action instead of the one the business needs. The week-one decision that could have ended this story was simple: leave demo bookings as the primary goal, track the other actions as secondary, and wait for the launch clicks to finish converting before making a judgment. Instead, the team treated an unfinished measurement window as an emergency. #### Days 14–20: better-looking numbers, worse decisions By day 14, the campaign showed 42 conversions. On a hurried status call, that sounded like progress. But 36 of those actions were whitepaper downloads, while demo bookings for the period had fallen to two. The sales pipeline was flat. The account manager pointed to the rising conversion count and said the bidding model was waking up. I understand the temptation. A green line is easier to defend than a discussion about delayed demo requests. But the line had changed meaning on day eight. **More conversions did not mean more demos.** The next idea was to “feed the model data” by raising the daily budget from $500 to $1,100 overnight. More budget could buy more traffic; it could not make an enterprise buyer book a demo before they were ready. With the bidding goal already broadened to include easier actions, the extra spend gave the campaign more room to pursue those actions. Guidance on [B2B SaaS campaign scaling](https://aimers.blog/b2b-saas-google-ads-best-practices/) warns against treating a sharp budget increase as a substitute for qualified demand. Three days at the higher daily budget produced no *recorded* demos. On day 18, the client panicked. That word “recorded” is doing work here. A buyer who clicked during those three days might still have been evaluating the product. Google Ads can [attribute a later conversion back to the original ad interaction](https://searchengineland.com/what-your-google-ads-hour-of-day-report-doesnt-tell-you-446098), so recent performance can look worse before those conversions arrive. The team did not wait to find out. On day 20, it cut the daily budget to $600 and lowered the Target CPA from $160 to $110 to “stop the bleeding.” Now the bid strategy had a changed goal, a budget that had jumped and fallen, and a more restrictive CPA target. The people managing it still wanted a clean read on whether the original setup worked. They had made that read impossible. Google lists a [setting change among the reasons a bid strategy may show Learning](https://support.google.com/google-ads/answer/6263057). Not every edit has an identical effect, and a status label alone cannot tell you the cost of each one. The practical problem was clearer: whenever the team saw a short run of uncomfortable numbers, it changed the conditions again. #### Days 24–27: a new offer, then new keywords By day 24, the argument had moved from bidding to the offer. The marketing VP thought an executive demo was too big an ask for a first visit. Without first resolving what the campaign should count as a primary conversion, the web team replaced the landing-page destination across the ad groups. Traffic that had gone to a demo calendar now went to a self-serve 14-day free trial sandbox. The offer might have deserved a test. It did not deserve to be swapped into the middle of this one without a plan for reading the result. A trial visit and a demo request are different asks. Changing the destination and offer changed the experience behind the ad while the team was still trying to judge the old experience. [B2B SaaS campaign guidance](https://tntgrowth.com/blog/google-ads-for-saas-scaling-rules/) treats offer changes as meaningful changes to campaign performance, not cosmetic page maintenance. ![Chalkboard showing a bidding trajectory crossed out and overwritten with “START OVER”.](https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb477254ba684cf486d423_fa222449-205d-4bd9-98ed-761092a23644.png) Three days later, the agency added 45 broad match keywords to the primary ad group. The argument was familiar: more traffic would compensate for erratic performance. But the campaign now had to handle a wider set of searches alongside the new destination, altered conversion goals, revised target, and budget swings. Google also identifies [composition changes as a possible Learning trigger](https://support.google.com/google-ads/answer/6263057). That was the damage. In less than a month, the team had changed the definition of a conversion, the amount it was willing to spend, the CPA it was aiming for, the offer people saw, and the searches the campaign could enter. By day 42, after roughly $22,000 in spend, the account still showed Learning. Nobody could point to a stable stretch of the original campaign and say what it had achieved. The algorithm did not need a pep talk. It needed an account that would stop moving long enough to be measured. #### The fix was available in week one The campaign needed a primary goal that meant the same thing on day one and day 42. For this setup, that meant keeping **demo bookings as the primary conversion action** and leaving the whitepaper download and product tour pageview as secondary observations. Their numbers would still be available in “All conv.” without making those lighter interactions the outcome Smart Bidding pursued. That single decision would not guarantee a $160 CPA. It would not create demand or shorten a buyer’s evaluation. It would, however, have preserved a consistent signal and spared the team the misleading day-14 celebration over 42 mixed conversions. With no apparent surge to celebrate, the case for doubling the budget would have been weaker. With a clear view of demos, the team could have investigated the $310 early CPA without pretending downloads had fixed it. The other half of the fix was timing. A learning phase is not a secret 14-day countdown. Its duration depends in part on how much useful conversion data arrives and how long it takes to arrive. If buyers commonly take several days between click and demo request, judging the most recent few days as though every potential conversion has landed is bad arithmetic. A campaign with sparse demo bookings needs more patience than one receiving a steady stream of prompt conversions. When the status says **Learning**, the team should inspect the stated reason before editing the account. Google’s [learning-status guidance](https://support.google.com/google-ads/answer/6263057) distinguishes triggers such as a new strategy, setting changes, composition changes, and conversion delay. Those are not interchangeable diagnoses. A change made yesterday calls for a different conversation from conversions that have not arrived yet. The [time-lag report](https://groas.com/post/google-ads-learning-phase-how-long-it-la) gives operators a way to check how much of a recent period may still be maturing. None of that means sitting on a broken campaign forever. It means separating a verified problem from the discomfort of waiting. If the primary goal is wrong, correct it. If the offer needs a test, plan one that can be read. But do not call three days of unconverted clicks proof that a nine-day buyer journey has failed. That is where an autonomous growth engine like [groas](https://groas.com/) has a better operating model than the check-in-and-tweak routine. It runs continuous execution within guardrails, keeps a human strategist responsible for direction, and records its actions. The point is not that a machine is magically patient. It is that campaign decisions should answer to the business goal and the available signal, not to someone’s need to bring a fresh adjustment to Thursday’s call. I used to give account changes too much credit simply because they looked decisive. This composite is the cost of that instinct laid out on a calendar: an incomplete first week, a goal changed on day eight, attractive but misleading numbers on day 14, budget and target swings by day 20, then an offer and keyword overhaul before the month was out. The campaign never got the quiet stretch its operators kept asking the platform to provide. My rule now is one sentence: **keep the real business conversion primary, and do not change the campaign to answer for clicks that have not had time to convert.** ## Related Posts [![Cover image for: Stop Asking When Google Ads Will Finish Learning. Stop Resetting the Conditions.](https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb4bda3748f4695a8d3309_5fc122e6-dd3d-4a83-87da-ca6f1fd63488.png) ##### Stop Asking When Google Ads Will Finish Learning. Stop Resetting the Conditions. Budget swings, tCPA cuts, and Friday asset dumps can keep Smart Bidding recalibrating. Before you blame the learning phase, check your change history. September 29, 2026 • min read Written by](https://groas.com/post/stop-asking-how-long-the-learning-phase) [![Cover image for: ROAS vs. CPA: The Margin Mistakes Beginners Make in Google Ads](https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb49c21470bda4cd649a58_cb8a62fb-d905-43b9-a16e-399a855282fa.png) ##### ROAS vs. CPA: The Margin Mistakes Beginners Make in Google Ads ROAS and CPA are different bets about your margins, not interchangeable bidding settings. Here’s how the wrong target can starve volume or buy unprofitable conversions. September 29, 2026 • min read Written by](https://groas.com/post/what-i-got-wrong-about-roas-and-cpa-when) [![Cover image for: Target CPA vs. Target ROAS: Which Goal Protects Profit?](https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb499dcae7506325addb77_7abaaade-6db5-4e1b-9fd4-069892192d8e.png) ##### Target CPA vs. Target ROAS: Which Goal Protects Profit? Target CPA controls acquisition cost; Target ROAS bids for conversion value. Here is the strongest case for each—and the default I would choose when the data cannot settle it. September 29, 2026 • min read Written by](https://groas.com/post/roas-or-cpa-the-bid-goal-argument-argued) ![](https://cdn.prod.website-files.com/6821efca072e48f6f495a47e/6823c03839b2b618430d5ae6_Light-p-500.png) [![White stylized owl eyes with green background icon.](https://cdn.prod.website-files.com/6821efca072e48f6f495a47e/6a68a036a5dd5627409d3579_Groas%20Icon.png)](https://groas.com/post/the-account-that-spent-six-weeks-stuck-i#) [contact](mailto:scale@groas.ai) Explore [Home](https://groas.com/)[Philosophy](https://groas.com/our-philosophy)[Blog](https://groas.com/blog) What We Do [Paid Search](https://groas.com/paid-search)[Earned Search](https://groas.com/earned-search) Who We Serve [Businesses](https://groas.com/for-businesses)[Agencies](https://groas.com/for-agencies) MCP [ChatGPT](https://chatgpt.com/plugins/plugin_asdk_app_6a9c2d7108f08191beb5b3d736be8655?q=groas)[Claude](https://claude.ai/directory/mcp-groas-ai) Legal [Privacy Policy](https://groas.com/legal/privacy-policy)[Terms of Service](https://groas.com/legal/terms) Get Started [Apply](https://groas.typeform.com/to/xC1bQNUT) © 2026 groas 🇺🇸 [![](https://cdn.prod.website-files.com/62434fa732124a0fb112aab4/62434fa732124a389912aad8_linkedin%20small.svg)](https://www.linkedin.com/company/groas/) ```json { "@context": "https://schema.org", "@graph": [ { "@type": "", "@id": "https://www.groas.com/post/the-account-that-spent-six-weeks-stuck-i#article", "headline": "Six Weeks in Learning: The SaaS Campaign That Kept Resetting Itself", "description": "A B2B SaaS Google Ads postmortem: how one conversion-goal change started a six-week cycle of reactive edits, and the week-one decision that could have stopped it.", "url": "https://www.groas.com/post/the-account-that-spent-six-weeks-stuck-i", "image": { "@type": "ImageObject", "url": "https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb477254ba684cf486d41b_5bbd3271-5221-4fc4-ac14-6970853c511a.png" }, "thumbnailUrl": "https://cdn.prod.website-files.com/6823bbd57170ea42b357cf81/6abb477254ba684cf486d41b_5bbd3271-5221-4fc4-ac14-6970853c511a.png", "datePublished": "2026-09-29T05:06:59.932Z", "dateModified": "2026-09-29T05:06:59.932Z", "inLanguage": "en-US", "isAccessibleForFree": true, "articleSection": "", "keywords": "Google Ads, ", "mainEntityOfPage": { "@type": "WebPage", "@id": "https://www.groas.com/post/the-account-that-spent-six-weeks-stuck-i" }, "author": { "@id": "https://www.groas.com/author/#person" }, "publisher": { "@id": "https://www.groas.com#organization" }, "about": [ { "@type": "Thing", "name": "Google Ads" }, { "@type": "Thing", "name": "Pay-per-click advertising" }, { "@type": "Thing", "name": "Performance marketing" } ], "mentions": [ { "@type": "Organization", "name": "groas", "url": "https://www.groas.com" }, { "@type": "Thing", "name": "Google Ads" } ], "speakable": { "@type": "SpeakableSpecification", "cssSelector": ["h1"] } }, { "@type": "Person", "@id": "https://www.groas.com/author/#person", "name": "", "description": "", "jobTitle": "", "email": "", "url": "https://www.groas.com/author/", "image": { "@type": "ImageObject", "url": "" }, "sameAs": [ "" ], "knowsAbout": [ "Google Ads", "Performance Max", "AI Max", "Pay-per-click advertising", "Conversion tracking", "Bid management", "Search advertising" ], "worksFor": { "@id": "https://www.groas.com#organization" } }, { "@type": "Organization", "@id": "https://www.groas.com#organization", "name": "groas", "alternateName": "groas.com", "url": "https://www.groas.com", "logo": { "@type": "ImageObject", "url": "https://cdn.prod.website-files.com/6821efca072e48f6f495a47e/6834a970e03a7016560e3515_Logo%20Design%20256x256.png" }, "description": "Dedicated strategists run your account and a proprietary engine trained on $500B in profitable ad spend optimises the execution underneath them", "sameAs": [ "https://www.linkedin.com/company/groas" ], "knowsAbout": [ "Google Ads", "Performance Max", "AI Max for Search", "Pay-per-click advertising", "Autonomous campaign management", "AI advertising agents" ], "areaServed": "Worldwide" }, { "@type": "BreadcrumbList", "itemListElement": [ { "@type": "ListItem", "position": 1, "name": "Home", "item": "https://www.groas.com" }, { "@type": "ListItem", "position": 2, "name": "Blog", "item": "https://www.groas.com/blog" }, { "@type": "ListItem", "position": 3, "name": "", "item": "https://www.groas.com/category/" }, { "@type": "ListItem", "position": 4, "name": "Six Weeks in Learning: The SaaS Campaign That Kept Resetting Itself", "item": "https://www.groas.com/post/the-account-that-spent-six-weeks-stuck-i" } ] } ] } ```