tROAS vs Maximise Clicks vs Manual CPC
When each Shopping bidding strategy actually fits, the data each one needs, what it optimises for, and the transition path from a brand-new product to a proven tROAS earner.
Three bidding strategies, three different jobs. Manual CPC gives you full control but no automation. Maximise Clicks buys traffic cheaply and is how you gather data on new products. Target ROAS chases a profit ratio but needs a conversion history to work. Most accounts don't pick one, they move products through them in sequence as data accumulates.
Three strategies, three jobs
The strategies aren't ranked best-to-worst. They optimise for different things, and each needs a different amount of data to work at all.
| Strategy | Optimises for | Data it needs | Control you keep |
|---|---|---|---|
| Manual CPC | Whatever you decide, bid by bid | None: you set the bids | Total, and total responsibility |
| Maximise Clicks | Cheapest possible traffic within budget | Almost none | Budget and an optional max CPC cap |
| Target ROAS | Revenue ÷ spend hitting your target | A real conversion history | The target; Google handles the rest |
Manual CPC does nothing you don't tell it to. Maximise Clicks buys the most clicks it can for your budget, ignoring value entirely: which is the point when the goal is data, not profit. Target ROAS predicts conversion value per auction and bids to hit a profit ratio, which only works once there's history to predict from. For the full mechanics of that last one, see Target ROAS explained.
The data requirement is the whole story
The single fact that decides which strategy fits is: how much conversion data does this product have?
- Zero to a handful of conversions: tROAS has nothing to work with. It will either spend erratically or refuse to spend. Use Maximise Clicks to buy data cheaply.
- Thin but non-zero (say, under 15 conversions): still too noisy for a stable target. tROAS set on this data whipsaws. Keep gathering, or pool the product with statistically similar siblings.
- A real history (roughly 15–30+ conversions): enough to compute a target the product can actually sustain. This is where tROAS earns its keep.
This is why "just switch everything to tROAS" fails on real catalogues: a 5,000-product feed always contains hundreds of products below the threshold, and forcing them onto a target starves them. The data-thinness problem is the constraint every Shopping account runs into.
Where Maximise Clicks fits (and where it doesn't)
Maximise Clicks is a data instrument, not a profit instrument. It's the right tool for:
- New products with no history: buy traffic, watch conversion behaviour emerge.
- Incubation: a dedicated Maximise Clicks campaign where new and thin-data products ramp until they've proven themselves.
- Reviving zombie products that never got enough impressions to prove anything either way.
It is the wrong tool for anything you're running for profit, because it will happily spend your budget on the cheapest clicks regardless of whether they convert. Always cap the max CPC and keep the budget deliberately modest: the goal is cheap data, not volume for its own sake.
Deep dive The transition path, product by product
Real accounts don't choose a strategy once; they move each product along a path as its data matures. The path looks like this:
- New product → Maximise Clicks incubator. No history, so the job is buying data cheaply in a separate campaign with its own modest budget and a max CPC cap. Efficiency is explicitly not the goal here.
- Enough conversions accumulate → graduate to the tROAS campaign. Once the product clears a conversion threshold and shows stable behaviour, it moves to the main campaign at a target derived from its own history: not the campaign average, which would over- or under-price it.
- Proven product → tROAS, tuned in small steps. From here it lives on a value-based target, adjusted in 10–20% relative steps so learning never resets.
Two things make this workable rather than a spreadsheet nightmare. First, the move between campaigns is a feed-label change, not a rebuild: lifecycle-stage custom labels carry the product from one structure to the other. Second, BidSmart's incubator automates the whole path: it runs the separate Maximise Clicks campaign, watches each product against the graduation criteria, and proposes the move to tROAS when a product is ready. Nothing executes on its own: the graduation, the target, and every step in between are staged for your explicit sign-off first. That approval-first model is covered in how BidSmart works.
Picking a strategy: the short version
- Brand-new product, no data → Maximise Clicks, capped, in an incubator.
- Proven product, real conversion history → Target ROAS at a target derived from that product's own numbers.
- A strategic override you genuinely want to hand-manage → Manual CPC, sparingly.
- A whole mixed catalogue → not one strategy, but a path: incubate the thin ones, run tROAS on the proven ones, and graduate products between them as data accumulates. See ramping new products for how the incubation half works.
Frequently asked questions
Which strategy should a new campaign start on?
Maximise Clicks, almost always. A brand-new product has no conversion history, so tROAS has nothing to predict from and Manual CPC has nothing to guide your bids. Maximise Clicks buys traffic cheaply while you accumulate the conversions that later strategies depend on.
Is Manual CPC dead?
No, but its niche is narrow. It's useful when you have a specific reason to override the algorithm, a strategic product you want visible regardless of short-term ROAS, or a tiny account where you genuinely watch every product. For most catalogues, a value-based strategy outperforms hand-set bids.
How do I know a product is ready to move from Maximise Clicks to tROAS?
Enough conversions to compute a stable ROAS, roughly 15–30 in the product's own history, not the campaign average. Below that, tROAS is guessing. Graduating on a fixed conversion threshold is exactly what an incubator structure automates.