A shop switches to a new agency and cost per lead spikes within the first few weeks. Nobody explains why until someone digs into what actually changed on the account.

Google Ads and organic search rankings both run on systems that take time to learn your business. Any significant change ripples through those systems, and the disruption doesn't show up right away. It shows up weeks and months later, once the original cause is easy to forget.

Your ad account has to relearn you

Every time a significant change touches your Google Ads account, whether that's a new person taking it over, a restructure, or a shift in strategy, the account doesn't just continue as it was. Google's Smart Bidding system treats a meaningful change as a reason to recalibrate. While it recalibrates, the account carries a "Learning" status, and performance during that window is unpredictable by design.

Google's own guidance sets a two week minimum for this. In practice, most accounts don't settle into stable performance until four to six weeks out, and it takes longer if conversion volume is on the lower side. Campaigns generating 30 or more conversions a week tend to clear learning faster. Campaigns generating fewer than 10 can take the full six weeks or longer, and fleet and diesel accounts, which often run on longer B2B sales cycles, commonly sit in that lower-volume range.

The type of campaign matters too. Performance Max and Shopping campaigns pull signals from multiple channels at once, which typically stretches their learning period longer than a standard search campaign. And if the typical conversion window runs seven to fourteen days, which is common for higher-consideration B2B purchases like fleet services, the algorithm can't fully judge its own bidding decisions until those delayed conversions come in. That pushes the effective learning period well past the two-week minimum, even before accounting for the disruption of a major account change.

This is also why any new hand on the account can make things worse without meaning to. A learning reset isn't triggered by any one specific cause. It's triggered by any of the following: changing a target CPA or target ROAS by more than 15 to 20%, a significant budget change, or restructuring campaigns. Anyone unfamiliar with an account's history is more likely to make one of these changes early, often to "clean things up," which resets the clock a second time right as it was starting to settle.

For context on how sensitive this system is: even a 20% change to daily budget alone can trigger a new learning phase. A full account restructure, new targeting, new creative, is a far bigger change than that.

On accounts that go through a major change mid-year, cost per lead has typically run 20 to 35% above baseline for the first three to five weeks before it settles back down. It’s important to make adjustments slowly and with this in mind. 

Your site has to relearn you too

The same pattern shows up in search rankings, but on a longer clock. When a significant change touches a website, whether that's a new CMS, a redesign, a full platform change, or a large content restructure, rankings don't drop the moment the change goes live. They erode three to four weeks later, once Google finishes reprocessing everything that changed. That delay is exactly why the risk gets missed: the site looks fine at launch, and the damage only becomes visible once it's too late to easily undo.

Part of that delay is mechanical. Google has to recrawl and re-index every page that changed, and how fast it does that is governed by a site's crawl budget, essentially how many pages Google is willing to check per day. If that process isn't managed deliberately, it's left to Google to eventually get to everything, which is a slower and less predictable path than actively resubmitting sitemaps and prioritizing high-value pages.

The damage also isn't evenly distributed. Brand searches, people searching a business name directly, tend to recover fastest, because that traffic doesn't depend heavily on technical signals. Non-brand rankings, the searches that actually bring in new customers who don't already know the business, are far more sensitive to structural changes: URL changes, content restructuring, and shifts in how a page's content aligns with what someone was actually searching for. If page structure changes or content gets consolidated without accounting for search intent, restoring the old wording won't fix it. The alignment between the page and the search itself has to be reassessed.

Metadata is another place this quietly breaks. Page titles, descriptions, and structured data don't carry over automatically through a platform change. They have to be rebuilt and verified, and a missed or default-generated title tag on even a handful of high-traffic pages can suppress rankings without ever throwing an obvious error.

How long recovery takes depends on what actually changed. An HTTP to HTTPS switch usually recovers in two to four weeks. A CMS platform change runs four to eight weeks. A full domain change can take six to eighteen months, though a properly executed one can recover in as little as thirty to sixty days.

Redirects themselves don't always cost ranking:  Google has confirmed that a correctly implemented redirect doesn't lose PageRank. It's redirects that are missing, mismapped, or chained through multiple hops, each of which weakens the signal Google is trying to pass through.

One shop who came to us last year lost roughly a third of its organic leads because an old redirect map missed a batch of service pages. Rankings took over six months to recover once the gaps were fixed. These are significant, and often silent errors that can have real repercussions on your bottom line.

Questions worth asking before any major account or site change

Before making a significant change to either system, it's worth having real answers to:

How long should it take for performance to stabilize after this change, and what should I expect to see in the meantime? Anyone who can't give a real timeframe here hasn't managed this kind of change before, or isn't being straight about the dip.

Who specifically is managing the technical side of this, and what happens to existing redirects or account settings? Ask for a name, not a team. If nobody can answer specifically, that's the answer.

What's the plan if rankings or ad performance drop more than expected? Anyone who's never had this happen won't have a real answer. Anyone who has will.

Can you show me what happened the last time a change like this was made? Not a highlight, the actual before and after. If that can't be shown, that's worth noting.

Any change to a working system has a cost

Given both of the above, a significant change to how an account or website is set up carries the full cost regardless of whether it turns out to be the right call. The reset happens either way: weeks of degraded ad performance, weeks to months of ranking risk. None of that is refunded just because the change was well-intentioned. This isn't an argument against ever making changes. It's an argument for being deliberate about them, not reactive.

If a current setup isn't performing the way it should, that's a real cost too, and leaving it alone isn't automatically the safer option. The point isn't act versus wait. It's that how a change gets managed determines whether that cost gets paid once, cleanly, or paid twice because the first attempt wasn't planned well.

Before making any significant change, get real answers to two things: who is actually managing it, and what the plan is for the weeks where performance dips no matter how well it's executed. And be sure the change is worth it before making it. It isn't something that can be undone for free.