You launched the new site last Tuesday. Your organic traffic dropped 40 percent overnight, which you expected. You turned on paid search to backfill the loss, which is what the contingency plan said to do. Six days later your branded paid search spend is half what it was before the launch, your nonbrand campaigns are still in the learning phase, and your agency wants another week before it will diagnose the problem. The marketer whose boss approved the migration budget is going to ask why the backup plan is not working. The honest answer is that paid search does not have a site migration playbook, and Google is not going to write one for you.

Paid search got reset the day the new domain went live. Your Quality Score got reset. Your historical conversion data got reset. Your automated bidding algorithms got reset. Your ad review pipeline got reset, because new ads and new URLs trigger another round of policy checks. None of this is documented in the place your team would look for it, because the playbook everyone uses is the organic search one.

What actually resets during a paid search migration

A domain migration changes more than URLs. It changes the destination every paid click lands on, which means every ad has to be reviewed again under Google's destination policy. It changes the conversion tracking path, because the new site ships new analytics events and often a new tag manager configuration. It changes the landing page experience signals Google uses to score ad relevance. It changes the auction dynamics, because the historical performance data attached to the new URLs is zero.

Your Quality Score drops on launch day and stays low until the new URLs accumulate enough click and conversion history to recalibrate. Your automated bidding strategy enters a learning phase that Google estimates at five to seven days but in practice can run for two to three weeks when the conversion volume is unstable. Your campaign-level history -- the month over month delivery patterns your team spent quarters tuning -- does not transfer. The new campaigns start from a clean slate, even if you cloned them from the old ones.

The bid recalibration is the most visible problem. Google told your agency that learning takes five days, and the agency told your CMO the same thing. The agency did not tell the CMO that the five day window is an algorithmic forecast based on stable conversion volume, not a deadline. When conversions drop because organic traffic is down and the new landing page is still being tested, the learning phase extends. The longer it extends, the more conservative the algorithm becomes, and the less spend it commits.

Why Google will not write this playbook

Alphabet reported $403 billion in revenue in 2025, with about 73 percent from advertising. Paid search during migration is a meaningful fraction of that. Google has the data, the modeling, and the engineering resources to publish a paid search migration playbook that mirrors the organic search one it already maintains in Search Central. It has chosen not to.

One reason is that publishing a playbook would commit Google to timelines and benchmarks it does not want to be measured against. The organic playbook is forgiving -- site migrations are messy and variable, so Google can recommend best practices without promising outcomes. A paid playbook would have to acknowledge that the platform sometimes underdelivers during migration, which would invite advertiser lawsuits and regulatory attention. The safer move for Google is to say nothing and let advertisers figure it out.

The other reason is structural. A migration playbook that recommended specific actions -- pausing campaigns, rebuilding ads in a specific order, holding bids at a floor for a defined period -- would expose how much control the algorithm actually has versus how much the advertiser retains. The migration experience is one of the clearest illustrations that the advertiser is no longer operating the campaigns, the algorithm is. Google does not want that lesson taught.

What to actually do during the next migration

Stop treating paid search as the contingency plan for organic volatility. Treat it as a separate channel that also breaks during migration, and budget for both being broken at the same time. That means holding back a meaningful reserve -- typically 20 to 30 percent of the migration quarter's paid budget -- to fund the second wave of campaigns after the initial reset clears.

Build the paid search campaign structure before launch day. The mistake most teams make is launching the new domain and then rebuilding the paid campaigns against it. The campaigns should exist in paused form on the new URLs for at least two weeks before launch, with historical conversion data imported from the old domain where possible. The day you go live you flip them on, not build them.

Set an explicit decision window with the agency. Five days is not a deadline, but it is a checkpoint. If the campaigns have not recovered 70 percent of pre-migration performance by day ten, escalate to a manual bidding strategy and rebuild the algorithm from a known good baseline. Waiting another month for the algorithm to figure it out is the most common and most expensive mistake.

Document the migration in the same way you document an incident. The teams that learn from migrations are the ones that record what reset, how long it took to recover, what the spend variance was, and what they would do differently. The teams that treat each migration as a one-time event end up relearning the same lesson every launch.

Your CMO approved the migration expecting paid search to backfill the organic drop. That expectation is wrong, and the trade press is not going to tell them. The next migration you plan should budget for organic and paid to both be degraded for two to four weeks, not assume paid will save you.

Sources