Should I use performance max or standard shopping?

Both campaign types sell the same products from the same feed. What differs is how much you can see and steer.

Standard Shopping keeps the decisions with you. You choose which products sit in which group, what each group bids, and which searches to exclude. You can read the search terms and act on them. The trade is that the campaign only appears on the Shopping surfaces it is built for, and it relies on you to do the optimisation.

Performance Max hands most of those decisions to the system in exchange for reach across more places at once and bidding that adjusts on signals you cannot see. The trade is visibility. You get less detail about where spend went and less ability to say no to a specific slice of it.

So the question is not which is better. It is which trade suits the account right now. A store that cannot yet explain which products make money needs to see. A store that already knows, has stable conversion data and wants more volume can benefit from handing over the steering.

One more consideration is time. Automation reduces the hours spent adjusting bids, but it increases the hours spent guarding the inputs: the feed, the product groups, the assets and the conversion values. The work moves, it does not vanish.

Choose it when you need evidence more than you need volume.

New accounts. With little conversion history, an automated campaign has nothing to learn from and will spend its first weeks finding out things you could have told it. A manual structure buys the same information more cheaply and leaves you with a search terms report you can act on.

Catalogues with wildly different margins. When one product can carry a large cost per order and another cannot carry any, explicit product groups are the cleanest way to make spend follow profit.

Stores with a feed that is not yet clean. Automation amplifies whatever the feed says, so poor titles and missing attributes turn into poorly matched traffic faster.

Accounts where somebody needs to justify the spend. Search terms, product level results and clear bids are much easier to defend and to learn from.

Seasonal or fast changing stock, where you need certainty that budget moves off a discontinued line the day it goes. Explicit control makes that immediate.

And stores with a small budget. When there is not much money to spend, spreading it across many surfaces at once produces a thin layer of data everywhere and a conclusion nowhere.

Choose it when the account already knows what works and the goal is more of it.

Steady conversion volume. Automated bidding improves with examples, so an account producing consistent orders every week gives it something real to work with.

A feed you trust. When titles, categories, images and identifiers are correct, wider matching is an advantage rather than a leak.

A catalogue where products are broadly comparable in margin, or where you are willing to split it into separate campaigns so the comparable ones sit together.

A store that wants presence beyond the Shopping grid, where the same budget can reach people on other surfaces without running separate campaigns for each.

Good conversion values passed back to the platform. This is the part stores skip. If every order is counted as the same value, the system optimises for order count and will happily buy cheap orders. Sending order value, and where possible profit, changes what it aims at.

Also have assets ready. Images, text and video are part of what the campaign uses, and leaving them empty limits it to a fraction of what it can do while still spending like the full version.

The mistake is putting the same products in both campaign types and hoping the account sorts it out. It will, but not in a way you chose.

Decide who owns what. The cleanest arrangement is to split the catalogue: one campaign type takes a defined set of products, the other takes the rest, with explicit exclusions on both sides so nothing sits in two places. Then each campaign's numbers mean something, and you can compare like with like.

If you do overlap them, expect the automated campaign to take most of the impressions for shared products. Plan for that rather than being surprised by it, and do not read the standard campaign's collapse as evidence that it stopped working.

When you test a switch, change one product set at a time and keep the rest as a control. Run the test long enough to cover a full sales cycle, and judge it on cost per order and on total store revenue, not only on what each campaign reports about itself. Automated campaigns tend to claim credit generously, so the store's own revenue is the tiebreaker.

Keep a manual campaign running somewhere, even a small one, as a window into search terms. Losing that view entirely makes every future diagnosis slower, because you no longer know what people are typing when they find you.