This plan assumes the Google Ads account exists (or you can create one in ten minutes), the website is in Astro, and you have a few hours each day. Each step links to the lesson with the detail.
Read that lesson when you reach the step, not before.
Your plan: Starter budget
With $5,000/month or less, you get one Search campaign, one or two ad groups and one region. Daily budget is roughly monthly budget ÷ 30.4, so $3,000/month is about $100/day.
Don't split the money across regions or campaign types. Concentrated spend teaches you something in weeks; spread spend teaches you nothing for months.
Your plan: Intermediate budget
With up to $50,000/month, launch two or three campaigns in week one:
Hold competitor campaigns until week three, after you have seen your own search terms.
Your plan: Advanced budget
Above $50,000/month, the first week is usually about a clean rebuild or a new region launch inside an existing account.
The same five days apply, but you should also confirm the CRM export for offline conversions on day two, so stage two tracking can follow within a month.
Don't launch the full budget on day five; start at a third to half and ramp over two to three weeks.
Day 0: account setup (about 30 minutes, only if you have no account yet)
The detail, including which access level to give whom, is in How Google Ads Works.
Pick one product or use case with clear search demand. A good test: would someone type a problem or category phrase into Google when they need it? "Invoice approval software" works.
"AI-native workflow platform" usually doesn't.
Write down the one conversion you'll optimise for: a completed free trial sign-up (account created, not just the form started). a submitted demo request from a business email. Then write down what you'll judge results on in a month, which is further down the funnel: trials that activate (reach a key action in the product) or convert to paid. demos that are accepted by sales as qualified (SQL) and the pipeline they create.
Get two things from colleagues today, because they take time: the list of countries you don't sell to, and a CRM export of last quarter's leads with source and stage. Draft the messaging brief with your LLM.
Read: 2. Messaging · 1. Foundations: keywords
Add the Google tag to the Astro base layout, create the conversion action in Google Ads, link GA4, and make sure auto-tagging is on. Capture the GCLID and UTMs in hidden form fields so they reach your CRM with the lead.
That last step is what makes stage two possible later.
EU/UK: You need a consent banner (CMP) that sends Google Consent Mode v2 signals, with all four signals set to denied by default before the visitor chooses. Do this today, not after launch.
Without it you lose most measurement for EEA and UK visitors, and Google restricts features for non-compliant accounts. As of October 2026
US/Canada: Google doesn't require Consent Mode here, but your privacy policy must disclose advertising cookies, several US states give an opt-out right for targeted advertising, and Quebec requires opt-in consent for tracking.
If you also get Quebec or EU traffic, set up the CMP now.
Your region: Check the consent rules in the Tracking lesson for your countries.
If any of your traffic comes from the EEA, UK or Switzerland, the Consent Mode v2 setup applies to those visitors regardless of where your campaign is aimed.
Test it: submit the form yourself from an ad-style URL (add ?gclid=test123) and check the lead arrives in the CRM with the value saved.
Read: 7. Tracking, stage one
Build or adapt one landing page for the campaign. It should repeat the ad's promise in the headline, show proof from companies like the visitor's, and have one form.
Use your LLM to make the Astro change and to critique the page before it ships.
Build the keyword list: 10 to 25 keywords in one or two tight themes, using phrase and exact match. Add the starter negative keyword list from the templates page.
On a Starter budget, avoid broad match for now. It needs conversion data and budget to learn, and it will spend your month on loosely related searches.
You can test broad match in one ad group once Smart Bidding has steady conversion data, usually a month in. Start with phrase and exact.
Read: 4. Landing pages · 1. Foundations: match types · 6. Wasted spend · Negative keyword list
Write one responsive search ad per ad group with 10 to 15 headlines and 4 descriptions, plus sitelinks and callouts. Then build the campaign with these settings:
Read: 3. Creating ads · 1. Foundations: bidding · 9. Budget decisions
Run through this list before you enable the campaign. Each item has caught a real launch.
Launch. Then set a calendar reminder for the same time next week to do your first weekly review.
For the first week after launch, look at the search terms report every day or two and add negatives for anything irrelevant: job seekers, students, free downloads, unrelated industries. After that, once a week is enough.
Don't change bids, ads and keywords all at once. If you change three things, you won't know which one mattered.
How long to wait before judging results depends on volume:
Starter: give it three to four weeks, or about 300 clicks, before deciding whether the campaign works. Daily numbers will jump around; ignore them.
Intermediate: two to three weeks, and at least 30 conversions, before you judge a campaign or change the bid strategy. Smart Bidding needs about a week after any big change before results settle.
Advanced: one to two weeks for a read on lead volume, but judge quality on SQLs or pipeline at four to six weeks, because sales cycles lag. Use campaign experiments for big changes.
In month two, start stage-two tracking (sending CRM outcomes back to Google Ads). Once you have a month or more of data, look at attribution.
Tracking comes before the campaign because Google's bidding and your own decisions both depend on conversion data. Clicks without conversions recorded are spend you can't learn from.
Teams that launch first and "add tracking later" usually lose the first month's data and then argue about whether the campaign worked.
Narrow scope matters because Google Ads learns per campaign and per conversion action. Fifty conversions in one campaign are useful. Five conversions in each of ten campaigns are mostly noise.
Expanding into more keywords, regions, Performance Max or YouTube makes sense once the core campaign produces steady conversions you trust.