A lot of Google Ads account structure advice is outdated.

The old goal was to force precision through segmentation. One keyword per ad group. Separate campaigns for each match type. Device splits on top of that. The theory was that more control equals better performance.

The new goal is to give Google enough clean signal to optimise while keeping control at the points where the business actually needs it. Those are different objectives — and they lead to different structures.

The shift in one sentence

The modern structure is generally: fewer campaigns, clearer business segmentation, themed ad groups, less pointless splitting by match type or device, and stronger negatives, messaging, landing pages, and measurement.

What still deserves separation

Separate campaigns when there is a real business reason:

  • –Different geographies with meaningfully different economics or language
  • –Different conversion targets or CPA/ROAS goals
  • –Different service or product lines with different margins or sales cycles
  • –Different inventory or budget constraints
  • –Branded vs non-branded — different roles, different budget logic

Do not separate just because old PPC habits taught you to. Ask whether the separation serves a real business need or just creates reporting noise.

Broad match and modern structure

Broad match works best when:

  • –Conversion tracking is trustworthy and tied to real outcomes
  • –Negative keywords are maintained and reviewed regularly
  • –Landing pages are relevant across the intent range being matched
  • –Ad groups are tightly themed around real intent families

Broad match in a messy account is not "AI optimisation" — it is expensive ambiguity. The keyword type is only as useful as the structure and measurement around it.

Lead generation example

For lead gen, a strong modern structure often looks like:

  • –One campaign per major service line or geography where economics differ
  • –Themed ad groups organised by buyer problem — not by match type
  • –Offline conversion logic guiding Smart Bidding toward real lead quality
  • –Negative lists that prevent irrelevant query expansion

Not:

  • –One campaign for exact, one for phrase, one for broad
  • –Device split layered on top of that
  • –Duplicate terms scattered across multiple ad groups
  • –Twenty campaigns each with one or two ad groups

eCommerce example

For eCommerce, structure decisions should reflect:

  • –Category economics — segment where margins differ materially
  • –New vs returning customer logic where budget or ROAS targets differ
  • –Brand vs non-brand — different roles, different bid strategies
  • –Inventory constraints — clearance, promo, and evergreen may need different handling

How to clean up an over-segmented account

If you inherit a messy account, clean in this order:

  1. Conversion setup — fix tracking before changing anything else
  2. Campaign purpose — document what each campaign is supposed to do
  3. Redundant splits — identify campaigns that can be merged without business loss
  4. Ad-group theming — consolidate single-keyword ad groups into intent themes
  5. Negative strategy — rebuild shared lists with clear ownership logic
  6. Landing-page alignment — confirm each ad group has a relevant, strong destination
  7. Reporting views — build views that reflect business outcomes, not vanity metrics

What not to clean first

Do not start by changing bid strategies or consolidating budgets before fixing conversion tracking. Changes made on bad data produce bad results even if the structural logic is sound. Tracking first — always.