Campaign and ad group structure

Part 2 started Bright Leaf Coffee with two ad groups. This part covers how that structure should actually grow — the single decision that most directly determines how well part 9's ad relevance component of Quality Score can perform.

The old approach: Single Keyword Ad Groups (SKAGs)

For years, the standard advice was one keyword per ad group — a Single Keyword Ad Group, or SKAG — maximizing relevance between keyword and ad copy by making them essentially identical. This genuinely worked well under the match-type behavior of that era, but it produces an enormous number of ad groups to manage, and modern broad and phrase match (part 5) already cast a wider net than SKAGs were designed to control precisely — Google's own machine learning-driven ad matching has made the extreme granularity of true SKAGs less necessary, and often actively harder to manage than it's worth.

The modern approach: tightly-themed ad groups

Ad group: "Coffee Subscription — General"
  Keywords: coffee subscription, monthly coffee delivery, specialty coffee subscription
  Ads: 2-3 RSAs, all speaking to general subscription value

Ad group: "Coffee Subscription — Gift"
  Keywords: coffee subscription gift, gift for coffee lover, coffee gift box
  Ads: 2-3 RSAs, all speaking specifically to gifting

Ad group: "Coffee Subscription — Single Origin"
  Keywords: single origin coffee subscription, ethiopian coffee subscription
  Ads: 2-3 RSAs, specifically mentioning single-origin sourcing

This is Bright Leaf Coffee's account grown to a realistic, mature structure — a handful of keywords per ad group, all sharing the same specific theme and intent, with ad copy written to speak directly to that shared theme. This keeps part 9's ad relevance component high (every keyword in an ad group is genuinely close to what the ads actually say) without the management overhead of hundreds of one-keyword ad groups.

Splitting by intent, not just by product

The three ad groups above aren't split by product feature alone — "General" targets someone deciding whether to subscribe at all, "Gift" targets someone buying for someone else, "Single Origin" targets someone already sold on subscribing but choosing a specific style. Each represents a genuinely different mindset a searcher is in, which is what should drive ad group boundaries — not an arbitrary keyword count target.

Campaign-level splits: budget and settings, not just theme

Campaign: Search — Coffee Subscriptions (daily budget $30)
Campaign: Search — Coffee Beans, One-Time Purchase (daily budget $15)

Beyond ad groups within one campaign, separate campaigns make sense when settings genuinely need to differ — a different daily budget priority, a different bidding strategy (subscriptions might use Target ROAS once mature, one-time purchases a simpler Maximize Conversions), or different location targeting entirely. Cramming products with meaningfully different business priorities into one campaign forces them to share a budget and bidding strategy that may not actually fit both equally well.

How big is too many ad groups

There's no fixed correct number, but a genuine warning sign is an ad group with keywords that don't obviously share one intent, or an ad group so narrow it gets negligible search volume — both waste management effort without a corresponding benefit. The actual test: could one set of Responsive Search Ads genuinely, specifically speak to every keyword in this ad group? If not, it's a sign the ad group is covering more than one real intent and should split.

Next: audience targeting — reaching people based on who they are and what they've done, layered on top of the keyword targeting covered so far.