Text version of this lessonExpand
The value of a Search campaign is control: you know which search intent you want to buy, which keyword set should trigger ads, and which page should receive the click. If your first Search campaign becomes a keyword warehouse, later clicks, search terms, and conversions will be hard to read.
Your first Search campaign is not a setup form
Many beginners open Google Ads and follow the screens: campaign, ad group, keywords, ads, and final URL. The workflow is fine, but the account can still become unreadable. One ad group gets too many terms, RSA copy stays generic, all clicks land on the home page, and after week one nobody can tell whether the keyword, ad, page, or conversion setup failed.
This lesson gives you one operating asset: a Search launch structure sheet. It answers seven questions before launch: what is the campaign goal, which intent does the ad group buy, which match types start the test, what does the RSA promise, what page proof supports the promise, which terms are blocked first, and what first-round data decides keep, pause, or expand.
Plain terms before setup
Search campaign is a Google Ads campaign triggered by user search queries. Someone searches "leak proof travel mug", then your ad can show. It is useful for validating active demand, not for educating a market with no awareness.
Ad group is the layer inside a campaign where keywords and ad copy are grouped. One ad group should carry one close search intent. If the same RSA and landing page cannot answer the whole keyword set, the terms should not sit together.
Phrase match / Exact match are keyword match types. Phrase match uses quotation marks. Exact match uses brackets. They do not guarantee only identical queries, but they keep the first sample easier to read than a broad start.
RSA means Responsive Search Ads. You enter multiple headlines and descriptions, and Google can combine them. The goal is not to fill every slot. The goal is to cover product, problem, proof, and click reason.
Negative keywords stop ads from showing on poor-fit searches. If you sell travel mugs, terms like free, DIY, repair, jobs, or used may become starter negative candidates so week-one spend does not leak into obvious low-intent searches.
Lesson output: Search launch structure sheet
Fill this sheet before launch. It is not a pretty template. It is a decision gate: each layer needs evidence before you touch the account.
| Decision layer | Write down | Healthy sign | Do not do yet |
|---|---|---|---|
| Campaign goal | Sales, leads, or visits; market, language, budget, and brand boundary | Serves one main job, such as non-brand purchase demand | Do not mix brand, competitor, broad discovery, and testing jobs |
| Ad group intent | The one need behind this search set | One RSA and one page can answer the group | Do not use the ad group as a keyword warehouse |
| Keywords and match type | Starter terms, Phrase / Exact start, held terms | Small, high-intent, explainable sample | Do not open the boundary too wide on day one |
| RSA promise | Product, problem, proof, click reason | Headlines and descriptions answer the search intent | Do not use one generic ad for every query |
| Landing page and negatives | Final URL, first-screen proof, starter negatives, pause / expand rules | The user quickly sees proof for the ad promise | Do not write a specific ad and send traffic to the home page |
Intent-to-page alignment router: prove one page can answer the group
Google Ads asks you to set a goal, keywords, ads, budget, and final URL. The account becomes readable only when search intent, RSA promise, landing page proof, and negatives stay on the same line.
| Search scenario | How to place it | What the page must support | What to read first |
|---|---|---|---|
| Product-type buyer | Can be a core first-round ad group, such as "waterproof dog seat cover" | Closest product page or tight collection page with product, specs, proof, and buy path above the fold | Whether search terms still show buying intent, and whether CTR, add_to_cart, and purchase reconcile with Shopify |
| Problem solver | Launch only if the page explains the problem and solution, such as "mug that does not leak in bag" | Use-case copy, comparison visuals, reviews, or proof, not only a product grid | Whether search terms extend the same problem and whether post-click behavior is weaker than product-type terms |
| Brand defense | Keep separate, usually in its own campaign or ad group | Brand page, home page, or closest product page that confirms the official path | Read brand traffic and conversion separately; do not use it to prove non-brand health |
| Competitor or comparison | Usually hold in round one unless you have a compliant, truthful, provable comparison page | Comparison or alternative page with reviewable proof, limits, and claims | Search terms, policy risk, page behavior, and conversion quality, not only CPC |
Starter negative candidates
Prepare obvious poor-fit candidates before launch: free, DIY, repair, jobs, used, support, login, tutorial. They are not always permanent negatives. They are questions you must answer before paying for those searches: can the page serve them, and are they worth the business value?
20oz Search structure lab: choose the query mess, then choose the launch action
To make the structure sheet practical, use a 20oz leak-proof commute tumbler as the working example. The common first-campaign mistake is not missing a Google Ads button. It is treating different problems as one problem. You see related terms and put them all in one ad group. You see cheap CPC and avoid negatives. You see brand terms convert well and assume non-brand Search is also healthy. Each mistake makes week-one data harder to learn from.
| Query mess | What it looks like | Correct launch action | Write back to the structure sheet |
|---|---|---|---|
| "leak proof travel mug", "20oz insulated tumbler", [commuter coffee tumbler] | Same product buyer intent, different angles | Split into one-intent ad groups, not mixed with gift, bulk, or broad accessory terms | Ad group: 20oz leak-proof commute tumbler; Phrase / Exact start; 20oz product page; negative candidates: free, DIY, repair, jobs, used |
| "mug that does not leak in bag", "travel mug no spills" | Problem-solving terms about leak prevention | If the page is only a product grid, fix the page before launch | Page task: add seal proof, bag-carry use case, real reviews, and return promise; buy the terms after review |
| "your brand tumbler" mixed with "leak proof travel mug" | Brand demand makes total conversion look stronger | Read brand terms separately; do not use brand ROAS to approve non-brand Search | Brand boundary: place your-brand terms in a separate campaign or ad group; keep non-brand focused on cold buyer intent |
| free tumbler, DIY travel mug, tumbler repair, jobs, used coffee mug | Clicks may be cheap, but they do not serve the purchase test | Add starter negative candidates and confirm campaign-level or ad-group-level exclusion in week one | Review rule: exclude searches with no buying intent before adding budget, even when CPC is cheap |
The point of this drill is simple: do not turn every problem into add more keywords. The first Search structure should protect readability. Product buyer terms should be bought clearly. Problem-solving terms need a page answer first. Brand terms should be read separately. Obvious low-intent terms should enter the negative candidate list. Then your first review can tell whether the issue is search-term quality, ad promise, page fit, or conversion tracking.
Official setup boundaries: completing the workflow does not prove the structure is right
The official Google Ads workflow for creating a Search campaign shows how to set goals, bidding, targeting, ad groups, keywords, and ads. That proves you can build the campaign in the interface. It does not prove the campaign deserves launch. The first Search campaign must validate whether the search intent is clear, match scope is explainable, RSA promise is specific, negative scope is safe, and the landing page can support the click.
First, match type is not simply wide versus narrow. Broad match is the default and broadest match type. Phrase / Exact can still match searches with related meaning, not only identical text. Starting with Phrase / Exact protects week-one readability, but it does not make the setup automatically safe. If you later allow broad or the broad match campaign setting, the structure sheet needs conversion, negative keyword, and page-fit evidence.
Second, RSA quality is not about filling every headline slot. Responsive Search Ads combine multiple headlines and descriptions, and asset reports can help compare assets. But the job is to write clear product, problem, proof, and click-reason promises. One asset label cannot prove a headline caused sales by itself. You still need search-term, landing-page, and conversion sample review.
Third, more negative keywords do not automatically mean safer traffic. Search campaigns can use broad, phrase, and exact negative keywords, but negative match types behave differently from positive match types. Account-level negative keywords can affect Search / Shopping inventory. Before launch, write negative candidates and scope, and protect brand terms, high-value terms, or terms you may test separately later.
Fourth, landing page experience is about relevance, usefulness, navigation, and what users expected after clicking. The first screen should support the ad final URL promise: product, specs, proof, CTA, shipping, or return evidence. But good page experience does not prove conversion. You still reconcile Shopify orders, GA4, and Ads conversions later.
| Official fact | Structure-sheet use | Cannot prove |
|---|---|---|
| Search campaign setup includes goal, bidding, targeting, ad group, keywords, and ads. | State the one job this campaign validates. | Does not prove brand, non-brand, competitor, and broad terms belong together. |
| Broad is the default and broadest match type; Phrase / Exact can still match related meaning. | Write why Phrase / Exact protects round one and when expansion is allowed. | Does not prove broad is always wrong or Phrase / Exact is always safe. |
| RSA combines headlines and descriptions; asset reports help compare assets. | Write RSA as product, problem, proof, and click-reason promises. | Does not prove every combination answers search intent. |
| Negative keywords can exclude poor-fit searches at different scopes. | Record negative candidates, scope, and terms that must not be blocked. | Does not prove more negatives are always better. |
| Landing page experience considers relevance, usefulness, navigation, and post-click expectation. | Record final URL and first-screen proof for every ad group. | Does not prove the page will convert. |
Admin evidence paths: do not only say "it is set up"
The useful proof for a first Search campaign is not a dashboard image. A reviewer should be able to follow the same admin path and find the same fields. Image captures only show that something looked complete at one moment. Field records, status records, change history, and order samples make next week's review practical: was the issue structure, page fit, search terms, or conversion trust?
| Admin path | Fields to keep | What this proves | Do not do |
|---|---|---|---|
| Campaigns > Campaign settings | campaign name, objective, conversion goal, bidding strategy, daily budget, network, location, language, start date | Proves the campaign validates one job and that budget, market, and language did not pollute the sample. | Do not keep changing goal, bidding, or location after launch; if you must, write it into change history. |
| Ad groups > Keywords | ad group name, keyword, match type, intent label, landing page URL, held terms, starter negatives | Proves the keyword set is not a warehouse and that one page can answer the shared intent. | Do not add every Keyword Planner idea; hold terms the page cannot answer. |
| Ads > Responsive search ad | headline, description, path, final URL, asset status, policy status, first-screen proof | Proves the ad promise and landing-page first screen tell the same story. | Do not claim offers, inventory, comparisons, or return promises the page cannot prove. |
| Tools > Ad Preview and Diagnosis; Campaigns > Change history | preview query, location, device, ad status, policy issue, change date, changed by, launch note | Proves the ad has diagnosable serving conditions and that launch actions have a change record. | Do not validate by Googling your own ad; use preview diagnostics and status records. |
| week-one readout: Search terms, Keywords, Ads, Landing pages, reconciled with GA4 events and Shopify Orders | search term sample, CTR, CPC, add_to_cart, purchase, order id sample, negative action, keep / pause / expand decision | Proves week one is not edited from one ROAS number; decisions come from query sample, post-click behavior, and orders. | Do not rebuild the structure on day two; wait for the minimum sample and record why to keep, pause, expand, or fix the page. |
This evidence path gives ads, page, product, and data teammates one shared record. For a 20oz leak-proof commute tumbler, Search structure is not "I like this keyword." It is: which intent the keyword came from, what the ad promised, what the final URL proves above the fold, which low-intent terms were blocked, and which search terms and order samples will decide the next move.
Control campaign-level variables first
Let the first Search campaign serve one main goal, such as non-brand search sales. Keep location, language, budget, and bidding logic simple. When budget is small, splitting into many micro-campaigns only makes every sample too thin.
If branded search already has meaningful volume, read it separately. Brand terms often convert better, but they do not prove non-brand Search is healthy. If brand and non-brand sit together, you may think the first Search campaign works when brand demand is simply rescuing the numbers.
Group ad groups by close intent, not by guesswork
To decide whether terms belong in the same ad group, ask one question: can one RSA and one landing page answer them? If yes, they can start together. If no, split them, hold them, or fix the page first.
For example, "waterproof dog seat cover", "dog hammock for car", and [dog car seat cover waterproof] can usually sit around a dog car seat cover because the page can explain waterproofing, back-seat fit, size, installation, and cleaning. "Free dog seat cover", "dog seat cover wholesale", and "dog car sickness" should not be blended into the same group because the intent and page need are different.
Do not start with too many keywords
Your first keyword set usually comes from three places: direct product-type terms, problem or use-case terms, and the value propositions already repeated on your product pages. This keeps ad copy and page proof aligned, and it makes later search-term review easier.
A safer first round
- Start with high-intent terms the site can clearly answer.
- Use Phrase / Exact first to protect readability.
- Hold obvious low-purchase, informational, free, job, and tutorial terms.
- Do not assume more keywords make scaling easier. Beginners need interpretability first.
RSA copy is a promise that the page must keep
Responsive Search Ads should cover four jobs: what the product is, what problem or use case it serves, why it is trustworthy, and why the user should click now. Trust proof can include material, reviews, delivery speed, or return policy. Click reason can include stock, shipping speed, an offer, or a trial threshold.
A user searching "waterproof dog seat cover" wants waterproofing, durability, easy cleaning, size fit, and delivery confidence. Vague brand language such as "premium lifestyle accessory" does not answer that search. The more specific the ad promise, the more the first screen must show matching proof.
Do not let the landing page break the promise
The landing page is where users arrive after they click the ad. In Google Ads, this is tied to the ad final URL. Google Ads Help describes landing page experience through relevance, usefulness, ease of navigation, and whether users see what they expected after clicking the ad.
So the first Search campaign should not send all traffic to the home page. Strong product-type terms should land on the closest product page or tight collection page. Problem-solving terms need use-case proof. Competitor or comparison terms should wait unless you have a compliant comparison page.
30-minute Search launch review
Before launch, run a simple 30-minute review. Spend the first 5 minutes on the goal: is this campaign validating non-brand buyer terms, brand defense, problem-solving terms, or competitor/comparison terms? It cannot carry every job at once. Spend the next 10 minutes on ad groups: can each keyword set be answered by one RSA and one page? If not, split it or hold it. Spend another 10 minutes on page fit and negatives: does the final URL first screen show product, specs, proof, and a buying path; have you judged candidates such as free, DIY, repair, jobs, used, support, login, and tutorial? Use the final 5 minutes to write the first-review sentence: which search terms, CTR, add_to_cart, purchase, Shopify orders, and post-click behavior will decide keep, pause, or expand.
This review does not need to be complicated. It only needs to preserve why the campaign was built this way. Without that record, the next search-term and negative review starts from guessing: why this term entered, why this page received traffic, and why this budget was spent.
Stop and go rules before launch
Pause if these are true
- An ad group cannot explain which single search intent it serves.
- Brand, non-brand, competitor, and broad terms are blended together.
- The ad is specific, but the first screen has no matching proof.
- There are no starter negatives or first-round keep, pause, and expand conditions.
Minimum acceptance checks
- Check: Each ad group serves one search intent.
- Check: Each keyword set has a landing page, match type, and negative candidates.
- Check: The launch has clear keep, pause, or expand conditions.
- Copyable lesson notes: The responsible person can review the action from the Search launch structure sheet without guessing why it was built this way.
Copyable lesson notes: make the first Search campaign reviewable
After the first Search campaign goes live, the useful output is not a dashboard image. It is a launch record that next week's reviewer can understand: why the campaign is grouped this way, why these terms are bought, why this page receives the click, which terms are held back, and how week one decides keep, pause, or expand.
Search launch structure copyable lesson notes
Current pressure: Do not build the first Search campaign as a keyword warehouse. Each ad group should serve one search intent.
First proof: Write the campaign goal, ad group intent, first Phrase / Exact keywords, RSA promise, final URL, first-screen proof, and starter negatives.
This week's action: Launch only terms that one RSA and one page can explain. If a problem-solving term lacks page proof, fix the page before buying the click.
Stop action: Do not read brand, non-brand, competitor, free, DIY, repair, jobs, used, or other mixed-intent terms inside the same group.
Review window: In week one, read search terms, CTR, add_to_cart, purchase, Shopify orders, and post-click behavior against the structure sheet.
Next route: If search terms drift, move to search terms, negatives, and match types. If clicks are good but page quality is weak, return to ad promise and landing page. If conversion trust is unstable, return to conversion tracking QA.
How this connects: after launch, watch search terms and budget guardrails
Launching a Search campaign is not the finish line. In week one, prove that Google is buying the right real queries and that low-quality terms are not consuming the budget.
- Next lesson: search terms, negatives, and match types to turn search term, matched keyword, and negative action into a reviewable record.
- Budget guardrail: budget and bidding basics to define break-even ROAS, Max CPA, and review window before changing budget.
Next step: use the sheet for search-term and negative review
After launch, do not change the account every day from one metric. Read search terms, CTR, post-click behavior, conversions, and Shopify orders against the structure sheet. If search terms drift, adjust match type and negatives. If clicks look good but post-click quality is weak, return to ad promise and landing page. If brand terms perform well, do not use them as proof that non-brand structure works.
The lesson is complete when you can close the first review in one sentence: because of this search-term and page evidence, we will keep, pause, or expand this ad group, observe until this date, and assign this person to gather the missing proof.