Most Performance Max debates I see on LinkedIn miss the point entirely. Practitioners argue about consolidation versus segmentation like it's a philosophical stance. It's not. It's a math problem with a clear answer for your specific account, and that answer changes as your conversion volume, margin mix, and brand exposure shift.
The real question isn't whether to consolidate or segment. It's whether your current structure is generating enough signal for the algorithm to learn while preserving the budget controls your finance team needs to forecast accurately. Get that wrong in either direction and you're either starving campaigns of data or letting Google optimize for easy conversions instead of profitable ones.
The Constraint That Determines Everything
Here's the technical reality that should anchor every structural decision: bids and budgets in Performance Max live at the campaign level only. Asset groups cannot carry different targets, different budgets, or different bidding intentions. Once you internalize that, the whole segmentation question becomes clearer, because every "how do I control X" question resolves to "do I need a separate campaign for X."
This is the opposite of how Standard Shopping worked. You could set bids at the ad group level, which meant one campaign could genuinely contain several different bidding intentions. Performance Max abstracts that control away. The algorithm decides which auctions to enter, which creative combinations to serve, and how to allocate spend across inventory signals.
That abstraction is intentional. Google's conversion optimizer is designed to find conversions wherever they exist within your campaign's reach, which means it can outperform keyword-level control when the data volume supports genuine learning. But "when the data volume supports genuine learning" is doing a lot of work in that sentence.
The Data Threshold Most Accounts Miss
Smart bidding needs a minimum of 30 conversions per campaign per 30 days to function properly. If your total monthly budget is $3,000 and you're running five campaigns, each gets $600 per month. If your CPA is $80, that's 7.5 conversions per campaign per month. Smart bidding is effectively flying blind.
The fix isn't to increase the budget. The fix is to reduce the number of campaigns until each one can generate enough signal. Two campaigns at $1,500 per month each will outperform five campaigns at $600 per month each in most situations, because the algorithm has enough data to actually learn from.
I've audited accounts where practitioners built 15 campaigns split by every product line, audience, and one-off idea. Fragmentation means no campaign gets enough conversion data to exit learning, so everything sits half-optimized while the budget gets spread too thin. The reporting looks sophisticated. The results don't.
When Consolidation Becomes a Liability
The opposite failure mode is just as common. One campaign ends up owning the whole account, and because Performance Max is algorithmic, you can't control where the spend goes or what gets promoted over something else.
The classic version is branded Search getting cannibalized by Performance Max. Your blended ROAS looks great, but a big chunk of that demand was already owned. Google can't tell the difference between a brand conversion and a net-new one, so it takes credit for a 10x ROAS while your actual prospecting efficiency suffers.
This is the number I'd want to see in any pipeline review: what percentage of Performance Max conversions came from branded queries versus net-new demand? If you can't answer that question with confidence, your reported efficiency is probably inflated.
Google spends where selling is easiest, not where your margin is. A single campaign with one target and one budget will naturally gravitate toward products that convert quickly, regardless of whether those products carry the margin your finance team modeled. If your catalog has meaningful margin variance across categories, a single campaign structure is actively working against your unit economics.
Five Signals That Should Trigger a Structure Review
Brand traffic cannibalization, feed quality, channel performance, conversion volume, and product margins should guide structural decisions. Here's how I'd operationalize each:
Brand cannibalization: Check the search term insights and impression share on your brand terms. If Performance Max is grabbing branded searches, your blended ROAS is overstated. The inability to set separate bids for branded and generic terms means you may need a separate brand Search campaign with higher priority.
Conversion volume per campaign: If any campaign is below 30 conversions per month, it's not learning. Either consolidate it into another campaign or increase budget allocation until it crosses the threshold.
Margin variance across products: If your high-margin and low-margin products sit in the same campaign with the same ROAS target, the algorithm will optimize for conversion volume, not contribution margin. Segment by margin tier when the variance is material to your P&L.
Category-level performance divergence: Stable cost-per-conversion drift and clear category-level performance divergence are restructuring triggers. If one product category consistently outperforms at a different efficiency level, it may warrant its own campaign with a different target.
Seasonal or promotional inventory: Products with time-bound demand patterns often need separate budget protection. A single campaign will deprioritize seasonal items in favor of evergreen performers until it's too late.
The Hybrid Structure That Actually Works
For accounts spending under $3K per month, two to three campaigns maximum: Performance Max plus brand Search. Add nonbrand Search only if the category demands it.
For larger accounts, the default for mature accounts is hybrid: protect high-intent Search (brand and key nonbrand terms) while using Performance Max for incremental demand capture and cross-network reach.
The structural decision is a tradeoff between control, efficiency, and what you can actually measure. Conversion definitions, lead quality feedback loops, URL control, and creative approvals determine whether automation helps or hurts.
The Two-Week Pilot Before Any Major Restructure
Any major structural change to a Performance Max campaign restarts Google's learning phase. Plan for a stabilization window before evaluating new performance.
Before restructuring, document your current baseline: conversion volume by campaign, brand versus nonbrand split, margin-weighted ROAS, and cost per incremental conversion. Run the new structure for at least two weeks before drawing conclusions. Compare against the baseline on the same metrics, not just the ones that improved.
The same account that should consolidate today might need to diversify six months from now. Build the review cadence into your operating rhythm. Quarterly structure audits, tied to the signals above, will catch drift before it compounds.
Model or it didn't happen. The Performance Max structure question isn't about ideology. It's about whether your current architecture is generating enough signal to learn while preserving the controls your CFO needs to sign off on the forecast.
