Most founders start paid ads the same way. They add a card, boost a few posts, set a budget that feels safe, and check the dashboard every morning. Two weeks later the spend is gone, the leads are thin, and the conclusion is “ads don’t work for us.”
Sometimes that conclusion is right. More often the test was never built to give an answer. This guide covers how to set up your first round of pay-per-click advertising so that, win or lose, you learn something you can act on.
Decide If Paid Search Even Fits Your Startup
Google Ads captures demand that already exists. Someone types a problem into a search box, and you pay to appear. That works well when people already know the category you sell in.
If you are creating a new category, nobody searches for your solution by name. You can still bid on the problem people describe, but the volume is often tiny. Before you fund search campaigns, type five phrases your ideal customer would use into Google and look at who shows up.
- If competitors and agencies fill the page, search demand exists and you can compete for it.
- If the results are blog posts and forums, demand is weak. Start with social platforms, where you create interest instead of harvesting it.
- If the results are a mix, test both with a small split of budget.
This one check saves more money than any bidding trick. Plenty of startups pour budget into search for a product nobody is looking for yet.
Set the Budget as a Learning Cost
Treat your first $5,000 as tuition, not as a customer acquisition plan. That mindset changes how you decide.
Write down three things before you spend a dollar:
- What you want to learn (which message, audience or channel produces a qualified lead).
- What result means “keep going.”
- What result means “stop.”
The third one matters most. Founders who skip it keep funding campaigns because stopping feels like admitting failure. A rule written in advance, such as “if we spend $1,500 and have zero qualified leads, we pause and change the offer,” removes the emotion.
Track the Event Closest to Revenue
Ad platforms optimize toward whatever event you tell them to count. If you track page views or button clicks, the system finds cheap clicks. If you track qualified demo requests or paid signups, it looks for people who behave like those.
Pick the event closest to revenue that still happens often enough to give the system data. Meta’s own guidance says ad sets need a minimum volume of conversion events to exit the learning phase. You can read the details in Meta’s Business Help Center.
If paid signups are too rare, use the step before them, such as a started trial. Then check in your own database how many of those people become paying customers.
Don’t Trust the Platform’s Return Number
Every ad platform grades its own homework. It reports conversions it believes it caused, and it often counts the same customer that also came from email, organic search or a direct visit.
Build one simple check outside the platform. Once a week, export new customers from your CRM or billing tool and ask each one where they first heard about you, in the signup form or in onboarding. Compare that with what the ad dashboard claims.
If the dashboard says 40 conversions and your database shows 22 new customers who mention ads, use the lower number to plan. The gap is normal. The mistake is budgeting off the larger figure.
Know Your Numbers Before You Scale
You do not need a perfect model. You need a rough answer to one question: how much can I pay to win a customer and still be fine?
Here is a hypothetical example, not a benchmark. Your product costs $60 a month and your gross margin is 70%, so each month leaves you $42. If customers stay about six months, that is roughly $252 of gross profit per customer. If you also need to recover the cost within a few months so cash stays healthy, you might cap your acquisition cost near $120.
Your own retention and margin will differ. The point is to set a ceiling from your margin and your cash, not from what other startups say they pay. If you do not have retention data yet, use a conservative guess and update it after your first cohort.
Build the Account Simply
Complicated structures eat budget before they produce data. For a first test:
- Run one or two campaigns, not eight.
- Give each campaign one clear goal and one landing page that matches the ad.
- In search, start with exact and phrase match keywords tied to buying intent. Add negative keywords on day one (words like “free,” “jobs,” “template,” “reviews”) so you do not pay for the wrong clicks.
- Separate branded searches (people typing your name) from non branded. Branded clicks are cheap and make blended numbers look better than they are.
Google documents how keyword match types work in its Google Ads Help Center.
Resist the Urge to Touch Everything Daily
Founders tend to adjust campaigns constantly. Changing budgets, audiences and creative every day makes the data impossible to read, and large edits can restart a platform’s learning period.
Pick a review rhythm and keep it. A weekly check is enough at this stage. Between reviews, look only for broken things: disapproved ads, tracking that stopped firing, or a budget that ran out in an hour. Change one major variable at a time, so you know what caused the result.
Fix the Landing Page Before the Ads
If your ads send traffic to a homepage with six menu items and three goals, the ads will look bad even when the targeting is right. Before you launch, make sure the page does four things:
- Repeats the promise from the ad in the first screen.
- Asks for one action.
- Loads fast on a phone.
- Has no navigation that sends visitors elsewhere.
Open it on your own phone, on mobile data, and try to complete the action. You will find problems quickly.
Choose Who Runs It
Many founders manage the first campaigns themselves, which is fine and often smart. It forces you to learn the basics. When spend grows, a specialist usually pays for themselves by catching waste you cannot see.
Whoever runs your account should explain their reasoning in plain language, share access to the ad account so you own your data, and be willing to show you where the money went. Practitioners such as Jorge Pérez are a good example of the profile worth looking for: someone who manages ad budgets daily and can talk through trade-offs instead of reciting jargon. Be wary of anyone who promises a specific return before looking at your product and margins.
A Simple Plan for the First 30 Days
Week one: confirm tracking, build one campaign, and launch with a modest daily budget.
Week two: review search terms or placements, add negatives, and cut anything that spent without producing a lead.
Week three: compare platform numbers with your own customer data and write down your real cost per qualified lead.
Week four: apply the stop or continue rule you wrote at the start. If it passes, raise spend in steps of about 20% and watch whether cost holds. If it fails, change the offer or the audience before you add money.
Paid ads reward founders who stay curious and keep records. Run the test with a clear question, a stop rule and honest numbers, and even a failed first month tells you what to try next.