How to Set Up Your First Google Shopping Campaign — A UK Step-by-Step Guide

Prerequisites — what to have ready before you open Google Ads

Most failed Shopping launches we see at Visionary aren't failed campaigns; they're failed prerequisites. The 90-minute setup only works when the four below are already true.

Google Merchant Centre, verified and with products approved

"Verified" means business address, HMRC company number, website ownership. "Approved" is the key state — not "uploaded". An uploaded feed with 30% disapprovals is not launch-ready. UK gotcha: VAT-inclusive pricing in the feed's price attribute. See our Merchant Centre setup guide.

GA4 connected to Google Ads (and the Google Ads tag firing)

The two-tracker model: GA4 for behavioural analytics, Google Ads tag for bidding. Both must fire. De-duplicate so a single conversion isn't double-counted.

Conversion tracking — the goals you'll bid against

Three goals before campaign creation: Purchase (PRIMARY), Begin Checkout (secondary), Add to Cart (secondary). If "Purchase" isn't primary you're bidding for add-to-carts. ICO PECR requires explicit consent before non-essential tracking fires — without Consent Mode v2 you under-report conversions by ~38.4%.

A CSS partner decision (the 20% UK saving most setups skip)

UK Merchant Centre accounts must associate with a Comparison Shopping Service. Default to Google's own CSS = paying the full 20% for the lifetime of the account.

Step 1 — Link Merchant Centre to Google Ads

In Merchant Centre go to Settings → Linked accounts → Google Ads and send the link request. Then switch to Google Ads → Tools → Linked accounts and accept it. The link must show "Approved", not "Pending" — a Pending link still lets you build a campaign in the UI, but products will never serve because the feed can't be attached. Check both sides before you close the tab.

The "country of sale" trap that costs UK→IE traders three weeks

UK traders expanding into Ireland routinely mis-configure country of sale. Google treats GB and IE as two distinct target countries, and each needs to be present in the Merchant Centre feed settings and in the Google Ads campaign geography at the moment the link is created. Add IE afterwards and you'll spend two to three weeks re-verifying the feed, re-approving product-level tax and shipping rules, and waiting for the Merchant Centre re-crawl. Do it once at link time.

If you sell into Northern Ireland, remember the Northern Ireland Protocol — NI addresses use the UK feed but a subset of EU VAT rules apply on certain goods; confirm your platform's VAT logic before you go live.

Step 2 — Choose Performance Max or Standard Shopping

What Performance Max actually does (and where it stops you intervening)

PMax bids across Search, Shopping, YouTube, Display, Discover, and Gmail from a single campaign, optimising via Google's automation. Its strength is reach — a well-fed PMax campaign will serve on placements a Standard Shopping campaign never touches, and it will keep learning as your conversion data compounds. The trade-off is control: no per-channel reporting, no per-product bidding, and asset-group-level insights that are deliberately shallow. If a specific SKU is losing you money you cannot bid it down; you can only exclude it entirely or move it to a lower-priority asset group.

When Standard Shopping is still the right answer

Three cases still favour Standard: sub-£20/day budgets where PMax cannot gather enough signal to leave the learning phase; high-AOV catalogues (typically £400+ AOV, furniture and B2B) where per-SKU bidding earns back the manual overhead; and heavily seasonal categories — Christmas gifting, fireworks, garden furniture in April — where PMax's automation will systematically overshoot the peak and under-serve the shoulder weeks.

Visionary's UK ecom default — and the two exceptions

PMax for 80% of new setups. The two named exceptions are the sub-£20/day budget and the high-AOV per-SKU-control case. Everything else — apparel, beauty, homewares, supplements, mid-market furniture, food and drink — defaults to PMax with a properly-fed catalogue and a brand exclusion list. See our complete Performance Max guide for the deep dive.

Step 3 — Structure your asset groups or product groups

Asset groups for PMax — one per product theme, not one per product

The mistake we see on 60% of takeovers is a PMax campaign built with one asset group per SKU, or worse, one asset group per Shopify collection. PMax needs volume inside an asset group to learn; a 20-SKU asset group with £1,500/month behind it will out-perform 20 single-SKU asset groups with £75 each. Group by product theme — "kids' sleep", "bunk bed accessories", "mattresses over £400" — so the creative assets, audience signals, and product listings match a coherent buyer intent.

Product groups for Standard Shopping — by margin band, not by category

Standard Shopping lets you set different bids per product group. The single-biggest win most Standard Shopping accounts leave on the table is bidding by margin rather than by category. A furniture retailer bidding "beds £2.00, mattresses £1.50" is bidding on price band, not profit. Bidding "high margin £2.20, mid margin £1.20, low margin £0.40" allocates spend to the SKUs that actually pay for the campaign.

How custom labels feed both structures

Set custom_label_0 to margin tier (high/mid/low), custom_label_1 to seasonality (evergreen/seasonal/clearance), and custom_label_2 to price band. In PMax these become asset-group filters via product listing groups; in Standard Shopping they become the group tree itself. See our feed optimisation guide for the full custom-label taxonomy.

Step 4 — Set your bid strategy

tROAS vs Maximise Conversion Value at launch

Launch with Maximise Conversion Value and tROAS off for the first 14 days. This asks Google to spend your daily budget on the highest-value predicted clicks without being constrained by a return target. Once 30+ conversions have landed and cost-per-conversion has settled, switch tROAS on at a target 15–20% below your account's current ROAS — not at your aspirational number. Overshooting the tROAS target is the fastest way to strangle a healthy new campaign; the algorithm will simply reduce spend until it hits the target on a smaller pool.

The "no data yet" problem and how to handle it for the first 14 days

tROAS needs roughly 30 conversions per month per campaign to learn accurately. Below that threshold the bid signal is noise. During the no-data window use Maximise Conversion Value to buy the highest-quality clicks the budget allows, and treat that fortnight as an information-gathering exercise, not a performance test. If a client asks for ROAS numbers on day 5, the honest answer is "we're still buying the data".

The bid-strategy mistakes we see most often on UK accounts

Two mistakes: setting tROAS to a target higher than the account has ever achieved (spend collapses), and switching bid strategy mid-learning-phase (learning restarts, week 1 is wasted). Change bid strategy at most once every 21 days on a new PMax campaign.

Step 5 — Set your budget (the right way for a new campaign)

The 10-clicks-per-day rule (and why we set it higher for UK Shopping)

The industry rule of thumb is a daily budget of 10× expected CPC — enough for ten clicks a day and, over a fortnight, enough impression-to-click volume for the algorithm to learn. In UK Shopping we push that higher because the auction is denser than the US average and the true settled CPC after week 1 is usually 20–30% below the launch CPC. Aim for 12–15× expected CPC on launch, then reduce once the learning phase closes.

Translated to real budgets: £20/day for low-CPC verticals (fashion accessories, food and drink, homewares under £30 AOV); £30–£50/day for mid-CPC (apparel, beauty, supplements, mid-market homewares); £80–£100/day for high-CPC (furniture, electronics, B2B, brand-term-heavy categories).

Daily budget vs total budget — which to pick at launch

Always daily budget on a new campaign. Total (or "campaign lifetime") budget caps spend at a fixed number and breaks tROAS pacing — Google throttles delivery in the final third of the campaign window to hit the cap, which distorts the learning signal. Use Total only for genuinely time-bounded promotions (a five-day Black Friday burst, a product launch week).

Budget pacing in the first 7 days

Expect day 1 to underspend by 40–60%, day 2 by 20–30%, and days 5–7 to occasionally overshoot by up to 2× (Google is allowed to spend up to 2× daily budget on any single day, provided the monthly average holds). Don't reduce budget in response to a day-6 overspend — you'll cripple the algorithm just as it hits its stride.

Step 6 — Audience signals (Performance Max only)

What audience signals do (and what they don't do)

Audience signals are directional seeds for the PMax algorithm, not hard targeting. Adding an audience signal tells Google "buyers who look like this are likely to convert" — it accelerates the learning phase. It does not restrict targeting to only those audiences. That's a critical distinction: a beginner who adds one audience signal thinking they've narrowed the campaign is actually just biasing the seed.

The three audience signals every UK ecom PMax campaign should include

First, your own-data customer list — uploaded from your ecommerce platform (Shopify → Google & YouTube channel handles this automatically; other platforms need a manual CSV upload every 30 days). Second, the closest in-market segment Google offers for your product ("women's apparel", "home & garden furniture", "beauty & personal care"). Third, a custom-intent audience built from your top 20 competitor domain names plus 10 high-intent category search terms. These three together compress the learning phase by about a week in our observed data.

Brand exclusions — the lever 80% of new PMax setups miss

Left unchecked, PMax will bid on your own brand terms — cannibalising organic clicks, driving up cost, and taking credit for conversions that would have happened anyway. Exclude your brand and close variants via the account-level negative keyword list (Tools → Shared library → Negative keyword lists → apply to all PMax campaigns), and separately request the brand exclusion via Google Ads support if your account has access to the beta. We find this missing on 80% of audit takeovers.

Step 7 — Launch and what to expect in week 1

The learning phase, demystified

The learning phase is the first 7–14 days after launch (or after a major structural change), during which Google's algorithm gathers conversion signal and calibrates its bidding. Impressions are uneven — one day quiet, the next flooded. CPC starts high and drifts down. Conversion rate climbs from ~0.4% on day 1 to ~2.0%+ by day 7 as the algorithm identifies the winning audiences and creative combinations.

The click / cost / conversion curve we see across 50+ recent launches

The chart below is the composite curve from our 50+ UK ecom Shopping campaign launches in 2025–26. Both clicks and cost ramp roughly linearly; conversion rate lags by a day or two and then accelerates. If your day-7 conversion rate is under 1.0% on a properly-targeted campaign, the problem is almost always product-page conversion rate or feed quality, not the campaign itself.

What to fix in week 1 and what to leave alone

Fix: disapproved products in Merchant Centre, missing negative keyword list, broken conversion event firing, obvious location-targeting errors. Leave alone: bid strategy, tROAS target, asset group structure, budget (unless it's genuinely wrong by more than 3×). Structural changes reset the learning clock and turn a two-week window into a four-week one.

Common setup mistakes (the ones we keep fixing on takeovers)

• Conversion goal set as Secondary instead of Primary — bidding optimises against the wrong action.
• Currency mismatches between Merchant Centre and Google Ads — campaign won't serve.
• Location targeting set to "interested in" UK instead of "people in or regularly in" UK — wastes spend on non-UK traffic.
• ICO consent not connected to Consent Mode v2 — ~38.4% conversion under-reporting.

Pre-launch checklist (interactive — 12 items)

Week-1 expectations chart — click / cost / conversion curve

Frequently asked questions