Demand Gen + Search + Shopping vs Performance Max: are agencies giving up too much control?
Retailers can build demand and capture sales without handing the entire spend distribution to Performance Max. For agencies, the controlled alternative deserves a proper commercial test.
If a retailer asks an agency how to grow Google Ads profitably, “put it into Performance Max” should not be the automatic answer.
PMax can perform well. It can also make an account easier to operate. Those are different benefits, and the client deserves to know which one is driving the recommendation.
My preference is to start with the commercial jobs: capture product demand through Shopping, capture relevant searches through Search, and use Demand Gen to introduce the business and its products to potential customers. Give each job an intentional budget and a measurement framework. Then test whether PMax improves the result.
Three campaigns. Three distinct jobs.
Standard Shopping puts products, prices and imagery in front of shoppers. Segment the feed around margins, stock, category priorities and product performance. Review search terms and use negatives to reduce irrelevant traffic. Shopping does not use positive keyword targeting like Search; the feed and campaign structure matter.
Search provides a separate way to target queries and communicate a relevant offer. Brand, non-brand category searches, specific product needs and local showroom searches can have different budgets, landing pages and commercial expectations.
Demand Gen brings visual creative and audience-led discovery into the mix. Use it to introduce a range, show products in context or support consideration. Google provides channel controls at ad-group level, so you can choose eligible surfaces such as YouTube, Discover and Gmail. This is a different buying context from somebody actively searching for a product.
These are intended roles, not rigid funnel boundaries. Demand Gen can generate direct purchases, and Search and Shopping can reach people early in their research. The structure makes the decisions explicit.
The real advantage: control over spend distribution
Imagine a retailer with an illustrative £100,000 monthly Google Ads budget. A starting plan could allocate £60,000 to Shopping, £25,000 to Search and £15,000 to Demand Gen.
| Campaign type | Planned budget | Commercial job |
|---|---|---|
| Standard Shopping | £60,000 / 60% | Capture product demand |
| Search | £25,000 / 25% | Capture priority queries |
| Demand Gen | £15,000 / 15% | Build and convert consideration |
This is a planning example, not a recommended split for every retailer or a forecast. Actual spend can differ from budget, and daily budget rules still apply. The useful point is that the business chooses how much funding each campaign type receives and can change it deliberately.
If Shopping can absorb more investment profitably, increase its budget. If Demand Gen needs a bounded creative test, protect that allocation. If non-brand Search is expensive, tighten its targeting without changing the funding for the other jobs.
Within PMax, Google allocates spend across eligible channels. Its channel report now makes that distribution more visible, but it does not give you a separate budget dial for each channel. Visibility helps you understand a decision. Control lets you make it.
“We spent £15,000 on building demand because we chose to test it” is a different conversation from “Google moved spend there.”
PMax reporting has improved. The distinction still matters.
It would be wrong to say PMax offers no channel visibility or search-term reporting. Google now provides both, alongside controls such as negative keywords and brand exclusions. A serious comparison should acknowledge those improvements.
The remaining trade-off is budget ownership. You can influence PMax through its settings, goals, assets and exclusions. You cannot directly set how much of the campaign budget goes to each channel. If your strategy depends on a deliberate balance between demand creation and demand capture, separate campaigns give you a clearer way to fund that balance.
Clearer reporting does not automatically mean better measurement
Separate campaigns make spend, targets and reported outcomes easier to discuss. They do not remove attribution bias or prove that every recorded sale was caused by advertising.
A customer may see a Demand Gen ad, return through Search and buy after a Shopping click. Campaign reports can divide credit across that journey; they cannot, on their own, tell you whether the purchase would have happened anyway.
For Search and Shopping, look at purchase revenue, contribution after advertising, product performance and the difference between brand and non-brand demand. For Demand Gen, review purchase outcomes alongside qualified visits, new-customer signals and assisted journeys. Cheap clicks are a diagnostic, not proof of profitable growth.
Check conversion windows and distinguish click-through, engaged-view and view-through reporting where available. Do not compare different attribution definitions as though they were identical. Allow for conversion delay before deciding a test has failed.
Across the account, reconcile platform reporting with first-party revenue, including phone and showroom orders where those matter. Track advertising spend as a percentage of turnover on a consistent basis. £100,000 of spend against £1 million of total turnover is 10% S%T; it is not automatically a 10x Google Ads ROAS because the revenue bases differ.
Where feasible, use a holdout, geographic test or suitable experiment to assess incremental impact. A rise in branded searches after Demand Gen launches is a useful signal, but seasonality, promotions and other marketing can also explain it.
The pros and cons of each approach
| Approach | Advantages | Trade-offs |
|---|---|---|
| Separate Shopping, Search and Demand Gen | Distinct campaign budgets and targets; deliberate product and query strategy; clearer ownership of creative tests and spend decisions. | More planning and maintenance; smaller data pools; risk of over-segmentation, overlap and poorly chosen budget splits. |
| Performance Max | Broad reach in one campaign; automated allocation; less structural administration; potential to find profitable opportunities across inventory. | No direct per-channel budget allocation; attribution still needs scrutiny; requires reliable tracking, feed quality, assets and commercial oversight. |
The controlled approach can fail too. Splitting a small budget into dozens of campaigns can starve bidding of useful data. Over-aggressive negatives can block valuable demand. Ring-fencing an inefficient channel can waste money. More manual decisions are only valuable when the decisions are good.
Separate campaigns can still use Smart Bidding. Commercial control does not require somebody changing every bid by hand. It means deciding the boundaries, objectives and funding within which automation operates.
How an agency should test the alternative
- Set the commercial baseline. Agree the revenue basis, margin constraints, cost-of-sale target and new-customer objective. Audit tracking before comparing structures.
- Choose a meaningful test scope. Use an eligible experiment or carefully designed product or geographic split. Avoid unrestricted duplicate targeting that makes the comparison hard to interpret.
- Keep the comparison fair. Account for promotions, stock, seasonality, budget, creative and conversion delay. Set the evaluation period and success criteria before launch.
- Judge the business result. Compare contribution, first-party revenue, acquisition quality and total cost of sale, alongside platform metrics. Explain uncertainty instead of declaring a winner from a short ROAS snapshot.
There is no universal test duration. The account needs enough purchases and enough time for delayed revenue to appear. A low-volume retailer may need a simpler structure and a longer evaluation than a high-volume business.
Where Performance Max makes sense
For a retailer with limited experience, expertise and management resources, PMax can be the more practical starting point. One consolidated campaign may be easier to maintain than an ambitious multi-campaign structure that nobody has time to manage properly.
That does not make it maintenance-free. The retailer still needs accurate conversion values, a healthy feed, suitable creative and regular commercial review. Poor inputs do not become good strategy because the campaign is automated.
PMax can also earn its place in an expertly managed account. If a fair test shows it produces better profitable growth, use it. The choice should follow evidence, not an ideological preference for manual or automated structures.
Agencies should be held to a higher standard
A retailer buying agency expertise is paying for judgement, management capacity and commercial accountability. Limited internal resources may explain why the retailer needs help; they should not become the agency’s excuse for skipping strategic work.
My position is that PMax is often better suited as a practical starting point for retailers with limited experience, expertise and resources. Agencies should be capable of building and managing the controlled alternative, or testing it seriously before making PMax the immediate default.
That means explaining why a particular campaign structure fits the client’s economics, how spend will be distributed, what the reporting can establish and how the recommendation will be challenged.
If the agency’s recommendation is PMax, the next question is simple: what evidence makes it a better commercial choice than separate Shopping, Search and Demand Gen?
Related analysis: Google Shopping isn’t dead. Lazy PPC strategy is.
Sources and platform details
Platform capabilities checked on 7 October 2026. The strategic conclusions above are my judgement; they are not claims that Google guarantees either structure will perform better.