Imagine you’ve launched a personal care range. Ten products. Your monthly Google Ads budget is £2,000. You’ve got a friend who tells you to chuck the lot into Performance Max and let Google’s AI work it out. You’ve got another who says spend hard, get the brand out there, ride the awareness wave, and the snowball will follow. And somewhere in your head is the cautious voice saying — push one hero product, do it well, build from there.

Three plausible strategies. They cannot all be right. So which one actually works on £2k a month?

Here’s the honest answer, after digging through Google’s own guidance, third-party case studies, and the data agencies have been quietly publishing all year.

The strategy people most often pitch — and why it’s risky on a small budget

The “all products into Performance Max, set ROAS or max clicks, let Google figure it out” approach is the default a lot of advertisers reach for. It’s the path of least effort, and PMax is genuinely powerful when you feed it well. The problem is that PMax has a minimum effective dose, and £2k/month sits right on the edge of it.

According to Google’s own published guidance for Demand Gen and PMax, the algorithm enters a learning phase the moment you launch, and it needs around 50 conversions before it can optimise reliably. Smarter Ecommerce, who analysed over 4,000 PMax retail campaigns, found that campaigns “need at least 30 monthly conversions (ideally 60+) for optimal performance.” Below that, you’re starving the system.

If your average product sells for £30 and your blended conversion rate is 2%, £2,000 of ad spend at a £1.50 CPC gets you ~1,300 clicks, which gets you ~26 conversions a month. You’re under threshold. The algorithm is making decisions on too little data, and the first 2–4 weeks of any new PMax campaign typically run 30–50% above target CPA while it learns.

This is the part nobody mentions when they tell you to “just put it in PMax.” There is no fast boost. There’s an expensive learning period followed by performance — if you’ve fed it properly.

The “burn money, get famous, ride the snowball” strategy

This is the more interesting question, because it’s the one funded brands have been doing for years and it does sometimes work. The logic: spend hard, accept poor short-term ROAS, get your product in front of enough eyeballs that brand recognition kicks in, then scale on cheaper branded search and repeat customers.

Here’s what’s true about it: branded search ROAS is genuinely transformational once it builds. One agency analysing their accounts found that branded keywords averaged 1299% ROAS while non-branded averaged 68% — roughly 19 times higher. So if you can buy enough awareness to create branded search demand, you eventually unlock a much cheaper traffic source.

Here’s what’s also true. That strategy works when you have venture money behind you. As Trendtrack put it bluntly in their 2026 benchmark guide: funded companies might accept 2:1 ROAS while investing in awareness; self-funded businesses generally need 5:1 to maintain cash flow. Every pound has to earn its place.

The snowball is real. But you can’t roll it with £2k unless you genuinely have eighteen months of runway and you’re prepared to accept that months one to six are net losses. Most small founders don’t have that, and pretending otherwise is how startups die.

The hero product strategy

This is the cautious version, and it has more going for it than people realise — but it has a specific failure mode too.

The case for it: with ten products and limited budget, picking one and pushing it hard concentrates your spend, gives that single product enough conversions to clear PMax’s learning phase quickly, and lets you build a clean dataset for the algorithm. Blue Wheel describe exactly this approach with one of their clients: they noticed the brand’s hero product was outperforming and absorbing most of the spend, so they broke it into its own campaign with a more aggressive ROAS target. That allowed them to scale it specifically.

The case against starting that way: you don’t actually know which of your ten products is the hero until the market tells you. Founders are nearly always wrong about this. The product you assume will sell is rarely the one that does. Picking a hero on instinct and putting £2k behind it for six months can mean six months of pushing the wrong product while your real winner sits invisible in the catalogue.

So what actually works on £2k/month and ten products?

None of the three strategies above, in their pure form. The approach that the data and case studies keep pointing to is a sequence — and it’s less exciting than any of them, but it’s the one that doesn’t waste your money.

Months 1–2: Fix the feed before you spend

Before any campaign, your Merchant Centre product feed has to be right. Store Growers, who specialise in ecommerce PMax, are blunt about this: “Weak feed = weak Performance Max results. No amount of campaign structure fixes that.” Titles with the keywords people actually search, clean images, accurate descriptions, GTINs, brand fields, and product type all properly filled in. This is unglamorous and it’s where most accounts have the biggest hidden gains.

If you skip this step, every other decision you make is built on sand.

Months 1–3: One campaign, all products, gather data

For a ten-product range on £2k/month, the consensus across the agencies I trust is: don’t over-segment. Store Growers’ rule of thumb is don’t create more campaigns than your conversion volume can support. At 30–50 conversions a month, that’s one campaign — possibly two. Five campaigns split by product category will starve every one of them.

So: one PMax campaign with all ten products, or — and this is worth considering for very small budgets — Standard Shopping with manual control while you build conversion history. Store Growers’ explicit recommendation: “If you’re just starting, or you’re spending less than £1k/mo, stay away from Performance Max. Stick to Standard Shopping and Search campaigns.” £2k is borderline. It can support PMax if your feed is strong; if it isn’t, Standard Shopping gives you more control while you fix things.

Bid strategy: Maximize Conversions for the first 30 conversions, then graduate to Target ROAS or Target CPA once you’ve got enough data. Switching too early is the single most common way to kill a campaign — the algorithm needs the historical signal to bid sensibly.

Months 4–6: Find your real hero, then split it out

By month four, you should have enough data to see which products actually pull. Not which ones you thought would. The ones the market keeps buying. This is where the hero strategy comes in, but earned rather than guessed.

Once you’ve got a clear hero — or two — you do exactly what Blue Wheel did with their client: pull it into its own campaign with a more aggressive ROAS target. Now you’re scaling something you know converts, with the rest of the range continuing to gather data in the general campaign. This is also when the Hero/Zombie framework that Smarter Ecommerce write about becomes useful: identify products that have zero impressions in PMax (zombies) and either give them a dedicated low-budget Standard Shopping campaign to force-feed them data, or accept they’re not viable and remove them from the feed.

What six months actually looks like for each approach

Pulling the threads together, here’s what each strategy realistically looks like over a six-month window on £2k/month with a ten-product personal care range:

All products in PMax from day one, hands-off. Months 1–2: heavy spending on whichever 2–3 products convert quickest, mediocre ROAS, learning phase chaos. Months 3–4: stabilises if your feed is good; some products get zero impressions. Months 5–6: ROAS settles around 2.5–4x if everything’s gone well, with most spend concentrated on a small subset Google chose, not you. Outcome depends almost entirely on feed quality and whether you’ve left it alone for the full learning period — every change resets it.

Big spend, ignore the waste, push for fame. Months 1–3: 1.5–2x ROAS, you’re losing money on a cash basis. Months 4–6: branded search starts appearing if your messaging has been consistent and your product is genuinely good, and that branded search is wildly profitable — but the trickle is small. Total: probably £4–6k of cumulative loss before any meaningful awareness compounds, on a £2k/month budget that’s most of your runway. Self-funded? Don’t.

Hero product only from month one. Months 1–2: clean data, fast learning phase exit, reasonable ROAS on one product. Months 3–6: scale that one product, ignore the other nine. Risk: you picked wrong and your real winner is invisible. Reward if you picked right: you’ve built a flagship and can layer the rest in afterwards. Worth doing only if you have genuine evidence (not instinct) that one product is the obvious lead.

Sequenced approach: feed first, all products, find hero, split. Months 1–2: minimal spend while feed is fixed and conversion tracking is verified. Months 2–4: data-gathering on full range, ~3x ROAS while learning. Months 4–6: hero pulled into dedicated campaign with tighter ROAS target, general campaign continues building the rest, account-level negative keywords protect against brand-term cannibalisation. Realistic 6-month outcome: 4–5x blended ROAS with one or two products clearly leading, a clear plan for the rest, and an account that’s set up to scale rather than restart.

The thing nobody wants to hear

The most expensive Google Ads mistakes on small budgets are almost never about the campaign setup. They’re about expectations. Founders launch with £2k/month and expect month one to look like month six. They make changes every week because performance is “bad,” which resets the learning phase and guarantees performance stays bad. They split into five campaigns because they want control, and starve every one of data. They put all ten products into PMax with a weak feed and blame Google when it spends on the wrong things.

The strategy that wins isn’t the boldest one. It’s the one that gives the algorithm enough clean data to do its job, then uses what you learn to make better decisions in month four than you could possibly have made in month one.

£2k/month is enough to build something. It’s not enough to buy fame. Spend it like the small budget it is, and it’ll grow.

Sources and further reading: