A poorly structured account isn’t just unpleasant to look at: it literally prevents the algorithm from learning correctly, for lack of data volume per campaign.
Step 1: start from intent, not from the catalog
Rather than creating one campaign per product or per site page, group by similar search intent. Two queries that reflect the same purchase intent, even for different products, can often coexist within the same ad group.
Step 2: separate Search, Shopping, Display and Performance Max
Each campaign type follows a different logic and deserves distinct budget and monitoring. Mixing these objectives in an overly merged structure makes performance analysis illegible.
Step 3: limit the number of ad groups per campaign
Too many ad groups dilutes data volume per group. Below a certain click threshold per week, the algorithm can’t distinguish a well-performing group from a weaker one for lack of statistical signal.
Step 4: a living negative keyword list
The negative list isn’t built once and for all at launch. It must be enriched every week from the search terms report, to prevent budget leaking toward off-topic queries.
Step 5: document changes
Noting the date and reason for each major structural change (campaign merges, bidding strategy changes) helps connect a performance variation to its likely cause, rather than guessing after the fact.
The trap to avoid: over-restructuring
Frequent restructuring resets campaigns to their learning phase each time, temporarily degrading performance. A structure should be designed to last several months, not tweaked at every minor dip in results.
For a full diagnostic of your current structure, see our Google Ads account audit guide.