Every decision about Google Ads account structure is a trade between two things you want and can't fully have at once. Splitting things apart gives you control — separate budgets, clean reporting, the ability to steer spend precisely. Keeping things together gives you signal density — the concentrated conversion data automated bidding needs to actually perform. For years the industry optimised almost entirely for the first. Automation moved the optimum sharply toward the second, and a lot of accounts are still built for a world that no longer exists. This guide is about finding the right point on that dial: consolidating enough to let the machine work, separating only where you genuinely need the control.
The hierarchy, briefly
For orientation: an account contains campaigns, campaigns contain ad groups, and ad groups contain keywords and ads. Budgets and targeting settings like location and language live at the campaign level, which is why campaigns are the meaningful unit of control. Ad groups exist to group closely related keywords with the ads that match them. If you're setting up your first account rather than restructuring one, our guide to structuring your first profitable campaign covers the basics before the strategy below.
Why the old playbook now backfires
The classic advice — dozens of tightly themed ad groups, single keyword ad groups, separate campaigns by match type and device, bid adjustments by hour — was rational when humans set bids. Granularity was control: the smaller the bucket, the more precisely you could price it.
Automated bidding inverted that logic. Smart Bidding models learn from conversion data within a campaign, and splitting one functional audience into ten segments hands each fragment a tenth of the evidence. The result is an account that looks meticulously optimised and quietly underperforms, because every layer of clever segmentation degrades the thing now doing the actual optimising. Responsive search ads suffer the same way — they need query volume to work out which headline and description combinations win, and micro-segmentation starves them of it.
The trade-off Automation didn't remove the tension between scale and control. It moved the optimum — hard — toward consolidation. Structures built before that shift are now working against you.
The governing rule: conversion volume
If you take one operating principle from this guide, make it this one. Before creating any new campaign or ad group, ask whether it will generate enough conversions to be worth isolating.
Roughly 30+ conversions per month per campaign is the level widely cited as letting Smart Bidding operate reliably, with more supporting faster, steadier optimisation.
Below about 15 per month, a campaign is generally considered starved — the algorithm has too little evidence to learn from, and it may never exit learning properly.
The test: if splitting would push either side below the threshold, don't split. If a campaign already sits below it, it's a consolidation candidate.
Google doesn't publish a single official number, so treat these as practitioner guidance rather than hard limits.
A common failure pattern makes this concrete: an ecommerce account with a dozen search campaigns, one per product category, each producing eight to twelve conversions a month, all on Smart Bidding, none ever settling. Nothing is wrong with the ads or the keywords. The structure is the problem — and merging those categories into a few well-fed campaigns typically fixes what months of bid tinkering couldn't.
The four legitimate reasons to split
Consolidation is the default, but "control" is a real requirement, not a nostalgic one. These four reasons justify a separate campaign, and they're the control side of the dial made rigorous.
Instapage
Premium landing page platform for paid-traffic teams with A/B testing, personalization and ad-to-page mapping.
Best for: Paid-traffic teams & agencies building high-converting landing pages
1. Budget separation is required. Budgets live at campaign level. If two things must not eat each other's spend, they need their own campaigns.
2. Geography or language differs. Location and language are campaign settings, so genuinely different markets require separation.
3. Goals or margins genuinely differ. A high-margin product line needing a different return target can't share a bid strategy with a low-margin one.
4. It's a different campaign type. Search, Shopping, Performance Max and video are separate by definition.
If your reason isn't on this list, consolidate.
What to stop splitting on
| Old practice | 2026 approach |
|---|---|
| Single keyword ad groups | Themed ad groups with related terms grouped together |
| Separate campaigns per match type | Match types coexist; broad match paired with Smart Bidding |
| Device-split campaigns | Let automated bidding handle device signals |
| Dozens of micro ad groups | Roughly 3–10 themed ad groups per Search campaign |
| Bid adjustments by hour and location | Signals the bidding model already incorporates |
| A campaign per keyword theme regardless of volume | Group themes until each campaign clears the volume test |
On ad groups specifically, published guidance in 2026 clusters somewhere between three and ten per Search campaign, but the number itself is the least important part. The real question is whether each group generates enough conversions to contribute useful signal — a group producing one or two a month is decoration, not structure. Broad match now carries much of the variation that granular ad groups used to handle, provided you pair it with Smart Bidding and genuinely active negative keyword management, a direction we cover in bidding without keywords.
Where Performance Max fits
PMax changes the shape of the question because it uses asset groups rather than ad groups, and it spans inventory that would otherwise sit in several campaign types. The productive way to think about it is complementarity rather than competition: PMax should cover territory your Search campaigns don't, with asset groups organised around distinct product categories or audience segments, each given tailored creative and audience signals.
Two practical cautions. First, PMax and Search campaigns targeting the same terms will interact, so decide deliberately which should own which queries rather than discovering the overlap in a report. Second, PMax is only as good as what you feed it — creative assets and, for retail, product feed quality do much of the work, which is why product page and feed quality is effectively part of your ad structure now. The same applies to creative: with the algorithm assembling combinations, systematic creative testing becomes a bigger lever than bid management.
The brand versus non-brand question
This one is genuinely contested, so it's worth presenting both sides rather than pretending there's a settled answer. The traditional case for separating brand campaigns is strong: brand terms convert cheaply and in volume, so mixing them into a blended report flatters your numbers and hides how non-brand is really doing. Separation also lets you cap brand spend deliberately.
The counterargument, which has gained ground, is that brand traffic behaves so differently from cold traffic that feeding both into shared automated bidding distorts what the model believes a typical conversion costs — and that separation creates smaller, less well-fed campaigns. In practice most advertisers still separate brand, accepting the smaller data pools in exchange for clarity, because the reporting distortion is immediate and obvious while the bidding distortion is subtle. That's a defensible choice; just make it knowingly rather than by habit, and recognise it as an explicit purchase of control at the cost of some density — a measurement judgement that connects to the wider difficulty of attributing performance accurately.
The control layer that costs you nothing
Consolidation loses you some visibility, and there are two cheap ways to get it back without fragmenting anything. Naming conventions are the first: a consistent scheme encoding campaign type, market, and theme means anyone can read your account at a glance, and it makes filtered reporting possible without structural splits. It takes an hour and pays for years.
Shared negative keyword lists are the second. Rather than maintaining negatives campaign by campaign, shared lists let you apply exclusions across the account centrally — genuinely important now that broad match does more of the matching. Both give you control at the reporting and exclusion layer instead of the structural layer, which is exactly where control should live once the algorithm is doing the bidding.
Restructuring an existing account safely
If you've read this far and recognised your own account in the "old playbook" column, resist the urge to rebuild everything on Monday. Restructuring resets learning, and doing it all at once means a simultaneous performance dip across the account with no way to tell what caused what.
Work in stages instead. Start with the clearest wins — the campaigns sitting well below the volume threshold — and merge those first. Give each change three to four weeks before judging it, since the learning period genuinely distorts early results. Keep budgets steady through the transition so you're not confounding a structural change with a spend change. And document what you did and when, because the alternative is a performance shift nobody can explain three months later. Spending decisions like these deserve the same scrutiny as the rest of your budget, especially in a year when every increase is funded by a cut and paid media competes with fast-growing retail media for the same pounds.
One last thing worth stating plainly: structure can't rescue a weak destination. Consolidated campaigns feeding a slow or unconvincing page will simply buy poor outcomes more efficiently, which is why landing page quality deserves at least as much attention as your campaign tree — and where a capable web team often unlocks more return than another restructure.
The bottom line
Structuring a Google Ads account for both scale and control means consolidating by default and separating on purpose. Give every campaign enough conversion volume to let automated bidding learn — around thirty a month is the commonly cited working target, and below fifteen you're starving it. Split only for budgets, geography, genuinely different goals or margins, and campaign types. Stop splitting by match type, device, or individual keyword, because those are signals the machine already reads better than you can. Then recover the visibility you gave up through naming conventions and shared negative lists rather than structural fragmentation. Fewer, better-fed campaigns with clean naming will beat an elaborate account tree almost every time.
Is your Google Ads account structured for the algorithm — or against it?
We audit, restructure, and run paid search accounts built to feed Smart Bidding while keeping the control you actually need.
Explore Performance Marketing →Frequently asked questions
How should you structure a Google Ads account in 2026?
Consolidate by default and split only when you have a specific business reason. Automated bidding needs enough conversion data per campaign to learn, so the granular structures that worked under manual bidding now degrade performance. Practical guidance widely cited in the industry is that a campaign wants roughly thirty or more conversions per month to let Smart Bidding operate well, and campaigns below about fifteen are usually starved. Structure should serve two things only: giving the algorithm dense signal, and giving you the budget and reporting separation you genuinely need.
How many ad groups should a Google Ads campaign have?
Far fewer than the old playbook suggested. Common guidance in 2026 lands somewhere between three and ten tightly themed ad groups per Search campaign, with the exact number mattering less than whether each group generates enough conversions to be useful to the algorithm. If an ad group produces only a handful of conversions a month, it isn't contributing meaningful signal and is a candidate for merging. High-volume accounts can support more segmentation; low-volume accounts should consolidate aggressively.
When should you split into separate Google Ads campaigns?
There are four defensible reasons: you need separate budgets that can't be allowed to cannibalise each other, you're targeting different geographies or languages, the segments have genuinely different goals or profit margins requiring different targets, or they're different campaign types. Anything else — match types, devices, individual keywords, minor thematic variations — should generally be consolidated, because splitting fragments the conversion data your bidding depends on.
Are single keyword ad groups still a good idea?
No. Single keyword ad groups made sense when advertisers set bids manually and wanted maximum control over each term. In an automated account they actively harm performance by splitting conversion signal into fragments too small for Smart Bidding to learn from, and by starving responsive search ads of the query volume they need to work out which asset combinations perform. If you're still running that structure, consolidation is the single highest-impact change available to you.
Should you separate brand and non-brand campaigns?
It's genuinely debated. Separating them gives you clean reporting, since brand terms convert cheaply and can flatter blended performance, and it lets you cap brand spend. The counterargument is that brand traffic behaves very differently, so mixing it into shared automated bidding can distort targets and mislead the algorithm about what a typical conversion costs. Most practitioners still separate brand for control and reporting clarity, accepting the smaller data pools that result.
