Text version of this lessonExpand
Google Ads is not one placement. It is an ad system shaped by search intent, product data, creative assets, conversion goals, and budget rules. The first beginner lesson is not to turn on Search, Shopping, Performance Max, Display, and Video at once. It is to choose one traffic path you can explain.
Boundary: choose by job, not by button popularity
Last reviewed: June 6, 2026. This lesson is for ecommerce beginners deciding how Search, Shopping, Performance Max, Display, and Video should be used before the first launch. Google Ads surfaces and labels can change, but the beginner rule stays stable: start with the traffic source that answers the first business question clearly.
This is not a dashboard walkthrough
- What: Google Ads has several campaign types, and each answers a different question.
- Why: If the first launch tests too many variables, you cannot tell whether keywords, page fit, feed quality, assets, tracking, or budget caused the result.
- How: Write the business question first, choose the campaign type that produces the clearest proof, and record what will wait.
First understand Google Ads: it is not one traffic pool
When someone says "run Google Ads," they may mean five different jobs: buy active search demand, show product cards, let automation find conversions across channels, run remarketing, or educate the market with video. These jobs can all live in Google Ads, but they are not the same traffic.
If you sell a 20oz leak-proof commute tumbler, a person searching "buy leak proof travel mug" is high-intent demand. A person seeing a YouTube lifestyle video is education or reach. A person comparing Shopping cards is testing product data and price fit. These need different evidence.
The 5 common campaign types and the job each one answers
| Type | Best first question | Do not use it for | First evidence |
|---|---|---|---|
| Search | Do people actively search for this product, use case, or problem? | Demand education, remarketing, or feed acceptance | Search terms, CTR, CVR, Purchase/Lead, first-screen match |
| Shopping | Can Google read your product data, price, image, and stock correctly? | Scaling before the feed is reviewed | Merchant Center diagnostics, title, image, price, stock, GTIN/MPN, page match |
| Performance Max | Can stable goals, feed, assets, and conversion signals support broader reach? | A shortcut while tracking is still unverified | Conversion quality, product grouping, asset coverage, brand and remarketing boundaries |
| Display | Reach, remarketing, and education | Proving that a product can sell from impressions or cheap clicks | Audience source, frequency, visit quality, remarketing pool size |
| Video | Product demos, use cases, and awareness | A tiny first launch that expects stable purchases immediately | Completion, engagement, search lift, downstream page quality |
Main artifact: campaign type selection map
The map is not about memorizing every campaign type. It makes the first account move readable. Ask: what proof is missing now? Do people actively search? Is the feed stable? Is conversion tracking trusted? Are assets ready? Can budget support a review window?
| Current situation | Suggested start | Can add later | Hold this round |
|---|---|---|---|
| People actively search for the product or problem | Search | Shopping / Performance Max | Display / Video as first purchase validation |
| Product data is stable | Shopping or feed-centered PMax | Search for high-intent capture, PMax for wider reach | Scaling before feed review |
| Conversions, assets, and feed are mature | Performance Max | Search for high-intent control, Shopping for product grouping checks | Unverified tracking, thin assets, short budget window |
| The job is education or remarketing | Display / Video | Search captures new demand, Shopping/PMax captures product clicks | Treating impressions, views, or cheap clicks as product proof |
Why beginners usually start with Search
Search is not always better. It is easier to read. You can see what people searched, which terms wasted spend, whether the ad promise matches the first screen, and whether Purchase / Lead tracking is trusted. For a small budget, readability matters more than broad inventory access.
Search also works as a minimum viable test. You do not need a large video library, many image formats, audience signals, or a mature feed just to learn from 10-30 high-intent queries.
Search still has launch gates
- You can write real high-intent queries, not broad interest words.
- The landing page first screen matches the ad promise, price, shipping, and proof.
- Purchase / Lead has passed conversion tracking QA.
- The budget can support a review window instead of splitting samples too thin.
Performance Max is powerful, but not a beginner shortcut
Google describes Performance Max as a goal-based campaign type that can access Google Ads inventory from one campaign. For beginners, the main lesson is not "it reaches more places." The lesson is that PMax depends on the goals, assets, audience signals, product data, and conversion signals you give it.
If Purchase value, currency, trigger timing, or deduplication has not passed QA, PMax may learn toward a bad goal. If feed title, price, stock, or GTIN/MPN is wrong, PMax may learn from weak product signals.
Before PMax, write four checks
- Conversion goal and value are trustworthy.
- Feed is no longer an unknown risk.
- Assets can support multi-surface serving.
- Brand terms, remarketing, high-margin products, and low-margin products have boundaries.
Launch risk router: First decide whether you are validating or blending variables
If the first campaign feels confusing, the problem is often not "not enough optimization." The problem is that the risk was not routed before launch. Put these four situations into the campaign type map.
| Risk scenario | Surface symptom | First action | Do not do |
|---|---|---|---|
| All types at once | Week one has data, but nobody can explain which traffic caused it | Pause the blended readout and choose one validation path | Do not compare PMax, Search, and Display ROAS as the same job |
| PMax before tracking QA | Purchase / Lead count, value, currency, or dedupe does not match | Return to conversion tracking QA first | Do not replace acceptance with "let it run a few more days" |
| Shopping with weak feed | Disapprovals, price or stock mismatch, or titles that look like internal SKU names | Sample title, image, price, stock, variants, and page match | Do not blame bidding while the feed is weak |
| Video/Display used as product validation | Impressions, views, or click cost look good, but orders and profit do not improve | Define it as education, remarketing, or asset learning | Do not write impressions or cheap clicks as product validation |
Stop / Go rules for the first campaign
| Stop | Go | Proof |
|---|---|---|
| Conversion tracking has not passed QA | Purchase/Lead, value, currency, trigger timing, and dedupe have proof | Notes or screenshots from Google Ads, GA4, and Shopify |
| Search, Shopping, PMax, and Display launch together | The first campaign answers one question | The selection map states which types wait this round |
| Shopping / PMax launches before feed review | Merchant Center diagnostics, price, stock, and variants are explainable | Feed checklist and product-page samples |
| Impressions, views, or cheap clicks are used as product proof | Purchase validation stays with Search, Shopping, or a trusted conversion path | Review sheet separates visit quality, purchase quality, and profit quality |
Handoff: bring these four lines into the next lesson
Copy this into your account plan
- This round's question: what business question should the first campaign answer?
- This round's start: which campaign type starts first, and why?
- This round's hold list: which campaign types wait, and why?
- Review rule: how long do you observe, and what signal triggers pause, review, or continue?
The next lesson covers account setup and conversion tracking. Do not enter it with only "we will run Google Ads." Bring the campaign type selection map. Then account structure, conversion goals, and budget guardrails will serve the same validation question.