Is Performance Max Worth It for Small Catalogues?

Is Performance Max worth it for a shop with forty products? The question gets asked the wrong way round. Retailers ask whether their catalogue is big enough for Performance Max, when the number Google’s automation actually cares about is how many conversions arrive each month. A shop with forty products and two hundred orders is a better candidate than one with four thousand products and eleven orders.

That distinction matters in Malaysia because most sellers here are small by design, not by accident. The Department of Statistics Malaysia counts micro, small and medium enterprises as a rising share of national output, and a large slice of them sell a tight range of products very well. Telling that retailer they need a bigger catalogue is unhelpful advice. Telling them what the automation actually needs is useful advice.

A note on our position before you read on. IZI Digital Marketing runs both Performance Max and standard Shopping for clients, so we gain nothing from talking you into either. What follows is the test we apply before recommending it, and the conditions under which we tell a small retailer to wait. If you want the service view, our Performance Max and Shopping page sets out scope and pricing. The video below is Google’s own walkthrough of the experiment that answers the question with data rather than opinion.

PART 1 · DIAGNOSE

What Counts as a Small Catalogue Anyway?

IN BRIEFThere is no official threshold, and nobody at Google publishes one. In practice a catalogue is small when one person can still review every product’s margin in an afternoon. Our comparison of Performance Max against standard Shopping uses the same working definition.

Ask three agencies and you will get three cut-offs, usually fifty, one hundred or two hundred products. None of them is documented anywhere. A more useful test is what the catalogue does to your working week.

• You can name your top sellers from memory — if six products carry most of the revenue and you know which six, manual control is still cheap for you. Automation is buying something you already have.
• Margins differ sharply between shelves — a small range often has one high-margin hero and several near-break-even lines. Automation optimising for revenue will happily sell the wrong ones.
• Stock moves faster than the feed — small sellers frequently run out and restock by hand. A feed that lags reality wastes clicks whichever campaign type spends them.

Only the second and third points actually predict whether Performance Max will hurt you. The first one just tells you how much work you are currently doing yourself.

PART 2 · DIAGNOSE

The Real Constraint Is Sales Volume, Not Product Count

IN BRIEFAutomated bidding learns from conversions, not from products. A narrow range that sells steadily gives the system plenty to work with, while a wide range that sells rarely does not. That is also why conversion measurement has to be right before anything else is discussed.

Google’s own advice points the same way. Its retailer guidance asks you to consolidate campaign structure so the system can optimise across channels using a unified budget, and to give transaction-specific values to each conversion rather than a flat number. Both instructions are about concentrating signal, not about having more shelves.

This is where small sellers quietly go wrong. They split a modest budget across three campaigns to keep categories tidy, then wonder why none of them settles. A single campaign carrying the whole range usually reaches a readable result faster.

BENCHMARK BRIEFING 1 OF 4

What Does Performance Max Need Before It Can Work?

IN BRIEFGoogle publishes a foundation checklist for retail advertisers, and most small Malaysian stores meet about half of it. The gap is almost always creative assets and first-party data, not products. Whoever handles your product feed management should be able to show you this list ticked off.

Each row below sets one documented requirement against what a typical forty-to-one-hundred-product Malaysian store usually has in place.

PART 3 · DESIGN

When a Small Catalogue Is Actually the Better Candidate

IN BRIEFA narrow range with even margins and steady repeat sales is close to an ideal Performance Max account. The automation cannot waste money on a shelf you do not stock, and every sale is a good sale. Compare that against the profiles in our review of Malaysian Performance Max agencies before you brief anyone.

Three conditions turn a small catalogue from a liability into an advantage: margins that sit close together, orders that arrive at least weekly, and a product people buy more than once. Meet all three and the automation has a narrow, well-lit room to search rather than a warehouse.

BENCHMARK BRIEFING 2 OF 4

Do Google’s Published Uplift Figures Apply to a Small Store?

IN BRIEFEvery uplift figure Google publishes comes attached to a condition, and most conditions cost money or effort. Read the attached requirement before you read the percentage. The same discipline applies to any agency quoting results at you, as our guide to Google Ads management sets out.

The grid below pairs each published figure with the input it depends on, and asks whether a small Malaysian store can realistically supply that input.

PART 4 · DESIGN

The Feed-Only Route, and What It Quietly Costs

IN BRIEFA feed-only build skips text, image and video assets entirely, so the campaign shows product listings much like standard Shopping. It is the safest first step for a small catalogue and the cheapest to run, though it gives up the video and Search formats. Your store setup has to be clean for it to work at all.

Three things change when you build without assets, and it is worth being clear about which of them you actually mind.

• Your ads look like your feed — no generated headlines pulled from a landing page, no auto-built video. For a small brand with tight photography, this is usually a relief rather than a loss.
• The video uplift goes away — the 12 per cent lift Google attaches to having at least one video simply does not apply, because there is no video. That is a fair trade if producing one is not realistic this quarter.
• The decision is hard to reverse — once manual assets are added, the minimum asset requirements are enforced. Start feed-only if you want feed-only.

BENCHMARK BRIEFING 3 OF 4

What Does a Small Catalogue Get for Each Monthly Budget?

IN BRIEFBudget decides how quickly the bidding gets something to learn from, and small budgets often cannot produce a readable signal inside a sensible window. Knowing that in advance stops you reading noise as a verdict, which is the most common failure across Malaysian retail marketing programmes.

The model below assumes a cost per sale of RM 90 and asks what each monthly budget produces before Google’s conversion delay is even accounted for.

PART 5 · DEPLOY

How to Answer This With a Test Instead of an Opinion

IN BRIEFGoogle ships an uplift experiment that measures what Performance Max adds on top of your existing campaigns, with a fifty-fifty traffic split by default. It is free and documented, and it settles the argument better than any store-side reporting can.

How to run a Performance Max uplift experiment

The steps below follow Google’s documented setup path for measuring incremental benefit.

1. Open Experiments. Go to Experiments inside the Campaigns menu in Google Ads.
2. Choose the uplift test. Select the Campaign types card, then the Performance Max Uplift card, and continue.
3. Select the campaign. Pick an active Performance Max campaign, or create a new one for the test.
4. Confirm the comparison set. Your campaign is tested against an automatically selected list of comparable campaigns, which you can edit while the test runs.
5. Set the traffic split. The split defaults to fifty-fifty between treatment and control; keep both arms large enough to read.
6. Schedule and leave it alone. Name it, set the start date, schedule it, and resist editing anything mid-flight.

If your question is specifically whether to move off standard Shopping rather than to add reach, Google documents a separate Shopping-against-Performance-Max experiment instead.

BENCHMARK BRIEFING 4 OF 4

Who Is Actually Asking This Question in Malaysia?

IN BRIEFSmaller firms now produce a rising share of Malaysia’s output, but their productivity per person improves slowly. That combination is why a wasted quarter of automated learning hurts more here than the global averages suggest, and why choosing the right partner matters as much as choosing the right campaign type.

Official Department of Statistics Malaysia figures, arranged so the direction of travel is visible rather than the headline size.

FAQ

Frequently Asked Questions

1. Is Performance Max worth it if I only sell twenty products?

Often yes, provided those twenty products sell steadily. It depends on your monthly order count rather than your product count, because the bidding learns from conversions and not from catalogue breadth. Twenty products with sixty orders a month is a better candidate than five hundred products with fifteen.

2. What is the minimum budget for Performance Max in Malaysia?

There is no published minimum, but below roughly RM 3,000 a month it rarely settles. It depends on your cost per sale, since the campaign needs enough conversions each month to have something to optimise against. Spend a thin budget on feed quality and conversion tracking first.

3. Should a small store run feed-only Performance Max?

Usually yes, at least to begin with. It depends on whether you can fund proper creative, because a campaign built with weak stock images turns those images into visible ads. Build feed-only deliberately from the start, since adding assets later enforces the minimum requirements permanently.

4. Will Performance Max waste money on my low-margin products?

It can, and this is the real risk for a small range. It depends on how far apart your margins sit, since automation optimising for revenue treats a thin-margin sale as a win. Use custom labels to separate priority products before you hand over control.

5. How long before I know whether it was worth it?

Plan for at least six to eight weeks at a moderate budget. It depends on your conversion delay, because Google advises excluding recent days from any evaluation when customers typically take time to buy. Judging a campaign at three weeks tells you about noise, not performance.

THE VERDICT

Buy the Readiness, Not the Campaign Type

Whether Performance Max is worth it for a small catalogue is decided before the campaign is built. It is decided by how many orders arrive each month, how evenly your margins sit, and whether your feed and conversion values describe reality. Nothing about the campaign type changes those three facts.

So do the unglamorous work first. Rank your products by profit rather than revenue. Fix the feed and give each sale its real value. Then either run Google’s documented uplift experiment, or if your volume cannot support one, choose on margin spread and review it next quarter.

Settle those and the answer arrives on its own, whoever ends up running the account.