Google Ads campaign structure: how many campaigns and ad groups?

The short answer on Google Ads campaign structure: run the fewest campaigns you can, and keep each one large enough to produce at least 30 conversions per month. Opening more campaigns than your budget can feed splits the signal automated bidding needs to learn, and that drop is measurable. On the ad group side, 3 to 7 themed ad groups per Search campaign is enough for most accounts under 2026 conditions.

What is Google Ads campaign structure?

Google Ads campaign structure is the architectural decision that determines how budget and conversion data are divided across the four-level hierarchy of account, campaign, ad group, and ads plus keywords. The campaign level carries budget, bid strategy, location and network settings; the ad group level carries search intent and ad copy. That is why campaign count is not a reporting preference: it is a decision about how sparse your data becomes.

How many campaigns is the right number?

The right number is the smallest one your budget and conversion volume allow. Practical rule: divide your total monthly conversions by 30, and the result is the ceiling on campaigns you can support. An account producing 90 conversions a month feeds three campaigns comfortably; split the same account into ten and each one averages nine conversions, which is not enough for automated bidding to read a pattern.

The only valid reason to open a new campaign is that the split needs a setting that only exists at campaign level. These five qualify:

• Separate budget required: brand searches should never eat from the same budget as generic searches.
• Different goal type: a sales campaign and a lead generation campaign cannot share one bidding target.
• Different margin or ROAS target: a 15 percent margin product and a 60 percent margin product cannot run on one ROAS target.
• Different geography or language: location and language settings only exist at campaign level.
• Different campaign type: Search, Shopping, Video and Performance Max are separate campaigns by definition.

What is data sparsity and how does it damage automated bidding?

Data sparsity is what happens when the same conversion volume is divided across so many campaigns and ad groups that no single unit holds statistically meaningful data. Automated bidding looks for patterns in order to predict per auction value. When it finds too little to read, it behaves conservatively, moves bids around, and stays in learning longer. The result is a higher CPA on the same budget and violent week to week swings.

Platform documentation commonly cites 30 conversions per campaign in the trailing 30 days as the working threshold for target CPA and target ROAS strategies, with a learning period that typically runs 7 to 14 days. Keeping the daily budget at 10 to 20 times your target CPA helps that learning finish in reasonable time. We compared the strategies themselves in our target CPA versus target ROAS guide; what matters here is how much data your structure leaves for whichever strategy you pick.

How many ad groups: themed grouping or SKAGs?

Three to seven themed ad groups per Search campaign is the right band for most accounts. A themed ad group is a group that maps 3 to 20 keywords sharing one search intent to a single set of responsive search ads. A single keyword ad group, or SKAG, is the older method of opening one group per keyword, and in most accounts today it does more harm than good.

• Close variants widened: the same query already lands in several groups, so the control a one-keyword group promised has largely disappeared.
• Responsive search ads: the system tests headline and description combinations; a one-keyword group leaves almost no room for that test.
• Signal splitting: in an account cut into hundreds of groups, no group collects enough conversions to learn from.
• Upkeep cost: every new group means another negative list, another set of ad copy and another landing page to maintain.

The one place SKAGs still hold up is a handful of keywords that carry high volume and high value on their own. Everywhere else, build themes around intent: brand, category, product attribute, problem-led queries. Treat your negative keyword strategy as part of the structure so themes do not leak into each other, and remember that copy plus landing page alignment feeds straight into Quality Score.

When does campaign consolidation pay off?

Consolidation pays off when two campaigns share the same goal type, margin band, geography and bid strategy. In an illustrative mid sized account that moves the same budget from twelve campaigns into four, monthly conversions per campaign rise from 13 to 39. Even if total conversions stay flat, the pattern automated bidding can read gets sharper and learning finishes sooner.

1. Measure monthly conversions per campaign over the trailing 90 days.
2. Flag every campaign under 30 conversions a month as a consolidation candidate.
3. Match candidates by goal type, margin band and geography; do not merge anything that fails the match.
4. Do the merge in a single move; migrating piece by piece resets learning several times over.
5. Do not read results for at least 14 days after the move; the first week is the learning period.

Three example skeletons: small, mid and large accounts

The three skeletons below are illustrative and scale with monthly budget and conversion volume. Your own numbers will differ; what does not change is that every campaign has to carry enough data to feed its own learning.

Small account (20 to 40 conversions a month): 2-3 campaigns are enough.

• Brand search: one campaign, one ad group, small budget.
• Generic search: one campaign, 3-4 themed ad groups.
• One Shopping campaign if you sell products; if you sell services, postpone remarketing at this stage.

Mid account (40 to 150 conversions a month): 4-6 campaigns.

• Brand search: separate campaign with its own budget.
• Generic search: 1-2 campaigns, 4-6 themed ad groups each.
• Shopping: two campaigns split by margin band, high and low.
• Performance Max: one campaign with asset groups per product group.
• Video or Demand Gen: only if a separate budget is genuinely required.

Large account (150+ conversions a month): 8-12 campaigns.

• Brand in its own campaign, competitor queries in another.
• Generic search: 3-4 category level campaigns with 5-7 ad groups each.
• Shopping: 2-3 campaigns by margin band.
• Performance Max: 2 campaigns for product groups with different ROAS targets.
• Remarketing and customer lists: separate campaign.
• A separate campaign set per country or language where those differ.

The cost of a structure decision comes less from the decision itself than from noticing it too late. Ads Sensor shows conversions per campaign, blended cross platform ROAS and anomaly signals in one panel; the AI analysis names which campaign is starved of data and why, then tracks the before and after of any change you apply. Join the beta to look at your own account structure the same way.