Short answer: if Amazon, eBay and two grey-market resellers are sitting above your own listing when a customer types your name into Google, you should be bidding on it. Brand bidding for UK ecommerce isn't a vanity line item in that situation, it's the cost of keeping a customer who already decided to buy from you. But the case only holds when someone else is genuinely in the auction. If your organic result owns the top of the page and the paid slots are empty, a brand campaign is mostly a rebate you're paying Google on traffic you already had.
So the first job isn't setting up the campaign. It's finding out who's there.
Check who is actually bidding on your name
Three things worth doing before you spend anything:
- Open Auction Insights on your brand keyword (or your brand campaign, if you already run one). It lists the domains sharing the auction with you and their impression share. Amazon and eBay will usually show. So will affiliates, comparison sites, and occasionally a competitor who has decided your customers are cheap to poach.
- Search your brand name from a clean browser on mobile, three or four times across a week, at different times of day. Marketplace dynamic ads are irregular. A single incognito check tells you almost nothing.
- Look your suspected advertisers up in Google's Ads Transparency Center. It shows what a given advertiser is running, which is useful when you want to know whether a reseller is using your trademark in the headline or just bidding on it quietly.
Write down the impression share numbers. You'll want them later when someone on the board asks whether the brand budget is doing anything.
Why marketplaces appear on your name, and why they won't stop
Amazon and eBay run search advertising at a scale where individual brand terms are not decisions anyone makes. Broad match, dynamic search ads generated from catalogue pages, automated bidding against category-level ROAS targets. Your name gets picked up because a page on their site is relevant to it. There is no account manager to email, and no threshold of politeness at which they stop.
They also have an unfair structural advantage: a marketplace can bid on ten thousand brands and lose money on nine thousand of them, because the value of the click to Amazon is a Prime customer, not one bottle of serum. You are bidding for one order. You will lose a pure CPC arms race, so don't enter one.
What you can actually get removed
Google draws a line between bidding on a trademark as a keyword and using it in the ad text. Keyword bidding on someone else's brand is generally allowed. Use of the trademark in the visible headline or description is something the trademark owner can file a complaint about, with carve-outs for legitimate resellers and informational sites.
That distinction matters more than most founders realise. You will not get Amazon out of your auction. You very often can get an unauthorised seller to stop writing your brand name in their headline, which knocks their click-through rate down hard and usually makes the keyword unprofitable for them within a fortnight. Register the mark in the UK if you haven't. Then file. It's a form, not a lawsuit.
The arithmetic, on net revenue not gross
This is where most brand-bidding decisions go wrong, so here's a full worked example with UK VAT handled properly.
Say a skincare brand with a £48 average order value. Prices are VAT-inclusive, so net revenue per order is £48 ÷ 1.2 = £40. Gross margin of 62% on net gives £24.80. Take off £4.50 for pick, pack and Royal Mail 48, and you have £20.30 of contribution per order. That's the number the campaign has to beat, not the £48.
The brand campaign gets 2,100 clicks a month at an average £0.34 CPC, so £714 of spend. Conversion rate 11%, which is normal for brand traffic. That's 231 orders. Reported ROAS on net revenue is 231 × £40 ÷ £714 = 12.9. The dashboard looks fantastic.
Now the honest version. Some of those 231 people would have scrolled two centimetres and clicked your organic result for free. Suppose a holdout test shows 18% incrementality. Real incremental orders: 231 × 0.18 = 41.6. Cost per incremental order: £714 ÷ 41.6 = £17.17. Against £20.30 of contribution, that's a win of about £3 an order. Thin, and one postage price rise wipes it out.
Change one input. Add Amazon and two resellers to the auction with meaningful impression share, and a click you don't buy no longer defaults to your organic listing — it goes to a marketplace where you either lose the sale or keep it at a 15% referral fee and no customer email. Incrementality in that scenario is routinely closer to 35%. Then: 231 × 0.35 = 80.9 orders, £714 ÷ 80.9 = £8.83 per incremental order. Now you're keeping £11.47 a time and you should be raising the budget, not defending it.
Same campaign, same spend, same ROAS number in the interface. Completely different decision. The variable that moves it is who else is in the auction, which is why the diagnosis comes first.
Run the holdout properly
Nobody enjoys switching off a campaign showing 12.9 ROAS. Do it anyway, once a quarter, and structure it so the result means something.
Alternating weeks is the pragmatic version: brand campaign on for seven days, off for seven, repeated for six to eight weeks. Compare total brand-sourced sessions and orders across on-weeks and off-weeks, including organic. Not paid orders. Total. If your total barely moves when the ads are off, you were buying your own traffic.
Geo splits are cleaner statistically but harder in the UK, because London and the South East distort everything. If you go that route, pair regions by traffic volume rather than geography, and accept that you need decent monthly order counts before the result is anything other than noise. Under roughly 300 brand orders a month, alternating weeks is the more reliable read.
One trap: run the test outside your peak. A holdout across Black Friday tells you about Black Friday and nothing else.
Win the results page instead of the auction
Since you can't outbid a marketplace, out-format it. Ad Rank is bid multiplied by quality and expected impact of extensions, and on your own brand term you should be crushing everyone on the second half of that equation.
- Exact match, tight structure. Your brand, your brand plus category, your two most common misspellings, and the informational variants ("brand reviews", "brand discount code", "is brand any good"). Separate from generic search so reporting stays honest.
- Extensions that a marketplace listing cannot match. Sitelinks to bestsellers, a promotion extension for the first-order code, price extensions for your hero SKUs, and a line about free UK returns or subscription pricing. The marketplace ad is one headline and a price.
- Offer asymmetry. Bundles, refill subscriptions, gift-with-purchase, a loyalty tier. Things you can do on Shopify and cannot do inside a marketplace listing. This is the only durable advantage you have on your own brand term.
- Landing page speed. Landing page experience feeds Quality Score, and Quality Score is what lets you sit above a bigger bidder for less money. A brand-term landing page with a 3.5s LCP on 4G is costing you CPC as well as conversion. If you don't know your numbers, SwiftStore will scan the store and track the score while you fix theme and app-script weight.
Also fix the organic side of the same page. Brand SERPs are winnable with structured data, a proper about page, review markup and a clean Google Business Profile, and every organic pixel you own reduces how much you need the ad. Our SEO and content team treats brand SERP defence as a separate workstream from category ranking, because the intent and the competitors are different.
Stop Performance Max from eating your brand budget
Left alone, Performance Max will hoover up brand searches, report them as its own conversions, and make your generic acquisition look far better than it is. Two features exist for this: brand exclusion lists in PMax and brand inclusions on Search. Use them. Put your brand and misspellings into a brand list, exclude it from PMax and Demand Gen, and keep a dedicated exact-match Search campaign for the brand terms.
The reporting consequence is the point. When brand and non-brand are mixed, you cannot tell whether your prospecting works. We have seen accounts where removing brand from PMax dropped reported blended ROAS by a third and the actual revenue didn't move at all. That's not a loss. That's the first honest number the account has produced.
Distribution terms, not price policing
The instinct when a reseller undercuts you on your own brand term is to set a minimum price they must not go below. Be careful. Under UK competition law, a supplier fixing or enforcing a minimum resale price is treated as a serious restriction, and the CMA has fined suppliers for exactly that. Recommending a price is fine. Enforcing one is not. Get proper advice before anything goes into a stockist agreement.
What you can control: who you supply, on what terms, with what marketing conditions, and whether authorised stockists may bid on your trademark in paid search at all. An advertising clause in the reseller agreement is legitimate and normal. Selective distribution criteria are legitimate. Differentiated pack sizes or SKUs for marketplace channels reduce the direct price comparison without touching anyone's pricing freedom.
Don't pay twice for the same customer
If you also sell on Amazon, check whether your Sponsored Brands campaigns are bidding on your own brand term inside Amazon while your Google brand campaign pays to send that same customer to your Shopify store. Both agencies will show you a healthy ROAS. Both are counting the same person.
Defensive bidding on Amazon has a real purpose, which is keeping competitors off your product pages. It just needs to be a deliberate decision with a number attached, agreed between whoever runs your marketplace ads and whoever runs your Google account. If those two people have never spoken, that's your cheapest fix this quarter.
Where to start
Pull Auction Insights for your brand term, note the impression share for Amazon, eBay and any reseller domains, then work out your contribution per order on net-of-VAT revenue the way we did above. Those two numbers decide the budget. If Amazon's brand impression share is above 20% and your contribution per order is comfortably over £15, fund the brand campaign properly and go and win the format war on extensions and offers.
If you'd rather have someone else pull it apart first, our UK team does a free audit that covers brand SERP ownership alongside the store itself.