Standalone paid search ads (text ads triggered by keywords you choose) are a distinct format from Shopping and Performance Max campaigns (covered in Google Shopping and Performance Max for Ecommerce Sellers, Explained). Paid search gives you direct keyword-level control that Shopping/PMax largely automate away — useful for branded search defense, high-intent non-product searches (like "return policy" or comparison terms), and situations where you specifically need to control which query triggers which ad and landing page.

Match types, explained simply

Exact match shows your ad only for the specific keyword or very close variants of it — the tightest control, lowest volume, typically highest relevance. Phrase match shows your ad for searches that include your keyword's meaning within a longer phrase — broader than exact, still reasonably controlled. Broad match shows your ad for a wide range of related searches based on Google's interpretation of intent — highest volume potential, lowest control, and it requires active management (strong negative keyword lists, careful bidding) or it can burn budget on loosely relevant traffic quickly.

Quality Score and Ad Rank

Quality Score is Google's assessment (on a 1-10 scale, visible in your account) of how relevant your keyword, ad, and landing page are to a given search, based on expected click-through rate, ad relevance, and landing page experience. It doesn't directly determine your rank, but it's a major input into Ad Rank, which does — meaning a highly relevant ad with a strong Quality Score can outrank a higher-bidding but less relevant competitor, and can also mean you pay less per click for the same position. For ecommerce specifically, landing page experience often gets undervalued: sending search traffic to a generic homepage instead of the specific, relevant product or category page tends to hurt both Quality Score and conversion rate together.

Similar automated, goal-based strategies apply here as in Shopping/PMax: Maximize conversions (spend the budget to get as many conversions as possible, useful early or for smaller test budgets), Target ROAS (bid to hit a return-on-ad-spend target — the most common steady-state choice once there's enough conversion history), and Target CPA (bid to hit a target cost per acquisition — more common for lead-gen than multi-price-point ecommerce, but usable for a narrow, similarly-priced product line). Manual CPC bidding still exists and gives full control, but is increasingly uncommon for ecommerce accounts with reasonable conversion volume, since it requires constant manual adjustment that automated strategies now handle reasonably well.

Building a basic campaign structure

A simple, workable ecommerce search structure: a branded campaign (your own brand name and close variants — typically cheap, high-converting, and important to defend even though it "should" convert without ads, because competitors sometimes bid on your brand name) and one or more non-branded campaigns organized by product category or intent (e.g., a category-specific campaign, and a separate campaign for comparison/research-stage terms like "best X for Y"). Keeping branded and non-branded spend separate is important for honestly evaluating performance — branded search ROAS is almost always dramatically higher than non-branded, and blending them hides how your acquisition campaigns are really performing.

Negative keywords: the unglamorous but critical maintenance task

A negative keyword tells Google not to show your ad for a given search term. Regularly reviewing your search term report (the actual queries that triggered your ads, as opposed to the keywords you set) and adding irrelevant or low-converting terms as negatives is one of the highest-leverage, lowest-cost ongoing tasks in paid search — especially on phrase and broad match, where irrelevant matches are expected and need active pruning rather than being a sign something's broken.

Worked example: reading a search term report

Say your keyword is phrase-match "waterproof hiking boots," and your search term report shows spend against: "waterproof hiking boots men's" (converting well — consider adding as its own exact-match keyword to control bid independently), "waterproof hiking boots reviews" (getting clicks but no conversions — a research-stage query; consider a negative if it's draining budget without downstream value you can measure), and "waterproof hiking boots kids" (if you don't sell kids' sizes — a clear negative-keyword candidate). This kind of review, done regularly, is what keeps a phrase/broad-match campaign efficient over time rather than degrading.

Common mistakes

  • Sending all search traffic to the homepage instead of the specific, relevant product or category page, hurting both Quality Score and conversion rate.
  • Ignoring the search term report for weeks or months, letting broad/phrase match accumulate irrelevant, budget-draining matches.
  • Blending branded and non-branded campaigns into one, making acquisition performance look better than it actually is.
  • Switching bidding strategies too frequently, never giving any one strategy enough of a learning period to show real performance.
  • Not defending branded search at all, assuming "no one bids on my brand name" — competitors sometimes do.

Best practices

  • Send search traffic to the most specific relevant landing page available, not a generic homepage.
  • Review the search term report on a regular cadence and maintain a negative keyword list actively.
  • Keep branded and non-branded campaigns and budgets separate for honest performance evaluation.
  • Give a bidding strategy a genuine learning period (commonly at least 1-2 weeks of stable settings) before judging or changing it.

Checklist: launching a first ecommerce search campaign

  • [ ] Conversion tracking verified and firing correctly before spending real budget
  • [ ] Branded campaign set up separately from non-branded
  • [ ] Landing pages mapped to specific product/category pages, not just the homepage
  • [ ] Initial negative keyword list built from obvious irrelevant terms (competitor brand names if not intentionally targeting them, unrelated product types, "free," "jobs," etc. as applicable)
  • [ ] Bidding strategy chosen with a plan to hold it stable for an initial learning period

FAQs

Should a small store even bother with branded search ads? Usually yes if there's any risk of competitor bidding on your name, since losing that traffic to a competitor's ad is often a bigger loss than the (typically low) cost of defending it.

How much should I expect to spend before paid search "works"? There's no universal number — it depends heavily on your industry's cost-per-click, conversion rate, and average order value. A more useful frame than a dollar figure is tracking whether cost-per-acquisition is trending toward a level your margins can sustain (see Break-Even ROAS Calculator) over a reasonable testing period, rather than expecting profitability from day one.

Is broad match worth using for a smaller account? It can work well when paired with active negative-keyword management and a strong automated bidding strategy with real conversion data behind it, but it carries more budget risk for a small account without the volume to quickly identify and prune bad matches. Many smaller accounts start with phrase/exact match and expand into broad match once they have a stable negative keyword foundation.