Account Structure & Tools

Hagakure Structure vs SKAG

In short: These sit at opposite ends of the same spectrum: how many keywords share an ad group, and how much control you hand to the algorithm versus keep for yourself. Hagakure pushes ad groups to be few and broad, feeding a small number of high-volume containers with broad match and trusting Smart Bidding to sort relevance. SKAG does the opposite - one keyword walled off in its own ad group so a human decides match by hand. Most real accounts land somewhere between these two poles. If you are rebuilding a legacy account, Hagakure is a reasonable destination and SKAG is what you are leaving behind.

By the AdFlint research team · Fact-checked against current Google and Meta platform behavior · Last reviewed July 2026

Hagakure Structure

Google's consolidation philosophy: few campaigns, broad keywords, and responsive ads, so machine learning sees pooled data instead of scattered fragments.

Originating from Google Japan, Hagakure argues that automated bidding performs better when traffic concentrates in a small number of high-volume ad groups rather than spreading across many thin ones. Advertisers adopt it when moving to Smart Bidding and broad match. The misunderstanding is reading it as an instruction to delete all structure. You still need separation where budgets, geographies, or margins genuinely differ, and negatives still do real work.

Full definition

SKAG

A legacy Google Ads structure putting one keyword in its own ad group so every query maps to tightly matched copy.

Single keyword ad groups gave advertisers manual control over message match back when exact match meant exact. Close variants ended that: the same query can now enter several ad groups anyway, and splitting keywords fragments conversion data that Smart Bidding needs. Treat SKAGs as history rather than a plan. Accounts still built this way usually show hundreds of low-volume ad groups, none with enough data to bid intelligently.

Full definition

Side by side.

The differences that actually change what happens in your account.

 Hagakure StructureSKAG
Ad group countDeliberately small - a handful of broad, high-volume ad groups per campaign.Deliberately large - one ad group per keyword, so count scales with your keyword list.
Match type usedBroad match by default, assuming Smart Bidding and query matching will find relevant traffic.Historically exact match, chosen specifically to control which query triggers which ad group.
Who decides relevanceThe bidding algorithm and auction-time signals, informed by pooled conversion data.The advertiser, by hand-picking the one keyword each ad group is allowed to match.
Conversion data per ad groupHigh - broad ad groups pool traffic from many related queries into one learning set.Low - one keyword's worth of volume, often not enough for Smart Bidding to leave Limited status.
Negative keyword roleLoad-bearing - broad match needs an active negative list to keep the algorithm inside relevant territory.Used mainly to stop near-duplicate SKAGs from triggering each other's ads.
OriginA structural philosophy documented by Google's Japan team, built around Smart Bidding and broad match.A manual-control technique from the era when exact match still meant exact.
Where it failsSmall accounts with too little budget to ever leave the algorithm's learning phase, or a weak negative list.Any account past a trivial size, where hundreds of thin ad groups become unmanageable and starve Smart Bidding.

What actually separates them.

01

Hagakure deliberately funnels traffic into a small number of ad groups to maximize data per container; SKAG deliberately isolates traffic to one keyword per container, minimizing data per container.

02

Hagakure assumes broad match plus Smart Bidding will find relevant queries the advertiser never typed; SKAG assumes the advertiser has already enumerated every relevant query and just needs an ad group to hold each one.

03

A Hagakure-built ad group can safely absorb an entire product category's traffic; a SKAG account needs a fresh ad group every time a new keyword variant gets added.

04

Negative keywords do fundamentally different jobs in each: in Hagakure they are the main lever constraining a broad, algorithm-driven ad group; in SKAG they mainly prevent sibling ad groups from stealing each other's traffic.

05

Rebuilding a SKAG account toward Hagakure principles means merging ad groups and widening match types at the same time - doing only one half does not produce either structure's benefit.

Which one should you use?

Use Hagakure Structure when

  • You are moving an account onto Smart Bidding and want ad groups with enough historical conversions for the algorithm to learn from quickly.
  • You maintain an active, well-tuned negative keyword list and are comfortable trusting broad match to find relevant queries within it.
  • You are rebuilding a legacy account and want a philosophy to rebuild toward, not just a smaller version of the old structure.
  • Your budget is large enough that a handful of broad ad groups will each still get meaningful daily traffic.

Use SKAG when

  • You are documenting or maintaining a pre-existing account exactly as it was built, without a mandate to restructure it.
  • You need one specific keyword, such as a trademark term, isolated from all others for legal or bidding reasons.
  • You are auditing an account and need to identify why ad groups are stuck in Smart Bidding's learning phase.
  • You are explaining to a client or teammate why an old account's structure is the reason automated bidding is underperforming.

Common questions.

Is Hagakure structure the opposite of SKAG?

Functionally yes, they sit at opposite ends of the same spectrum: how many keywords share an ad group, and how much control you hand to the algorithm versus keep for yourself. Hagakure pushes toward few broad ad groups and heavy reliance on Smart Bidding; SKAG pushes toward one keyword per ad group and full manual control.

Can I apply Hagakure principles without switching to broad match?

You can consolidate ad groups on phrase or exact match and get some data-pooling benefit, but you lose the part of Hagakure that finds query variations you never listed. Most of what people mean by Hagakure structure assumes broad match is doing real work, not just that there are fewer containers.

Why did merging my SKAGs into fewer ad groups hurt performance at first?

Merging resets some of the ad-group-level history Smart Bidding had built up, so expect a short adjustment period before the pooled data starts helping. If performance does not recover after that window, check whether the merged ad group's theme is actually coherent or whether you combined queries that don't belong together.

Or stop choosing between them.

AdFlint picks the setting, writes the ads, and keeps optimizing inside the Google and Meta accounts you already own.

Try AdFlint freeSee how it works

Related comparisons

All comparisons