TL;DR
Before you spend a single rupee on Google Ads, lock down four things: airtight conversion tracking, a clean account structure, a budget large enough to exit the learning phase, and the right bidding strategy for your stage. Most startups skip the tracking, scale on broad match, and burn weeks of budget on clicks that never convert. This is the pre-launch checklist we use at Varnan to stop that from happening.
Here’s the uncomfortable truth most agencies won’t tell an early-stage founder: Google Ads will happily take your money whether or not your campaigns are set up to win. The platform is built to spend your budget — it is your job to make sure that spend is measurable and pointed at the right people. For an AI or software startup, where a single qualified demo can be worth lakhs, getting the foundation wrong in the first month doesn’t just waste cash; it teaches Google’s algorithm the wrong thing about who your customer is.
At Varnan, our performance team manages Google and Meta budgets for software and AI startups, and the same preventable mistakes show up again and again. This guide is the exact pre-launch sequence we run before turning a single campaign on. Work through it in order — the order matters.
Why most startup Google Ads accounts leak money in week one
New accounts fail for boring, structural reasons — not because the ad copy wasn’t clever enough. The three most common: campaigns go live before conversion tracking is verified, budgets are spread so thin that Smart Bidding never gathers enough data to learn, and default settings (Search Partners, the Display Network, auto-applied recommendations) quietly siphon spend into low-intent placements.
The fix is to treat setup as a checklist, not an afterthought. You can recover from a weak ad. You cannot recover the rupees spent driving the algorithm toward the wrong audience for three weeks.
If you can’t measure a qualified lead, every rupee you spend on Google Ads is a guess — not an investment.
Step 1: Conversion tracking is the only thing that matters first
Nothing else on this list works without this. Google’s Smart Bidding — the system that decides who sees your ad and how much to bid — is only as good as the conversion signal you feed it. Garbage signal in, wasted budget out.
Install the Google tag and verify it before launch
Set up the Google tag (gtag.js) or deploy via Google Tag Manager, then create a conversion action in Google Ads for your real business goal — a demo booking, a signup, a qualified lead form, not a pageview. Use Google’s Tag Assistant to confirm the tag fires on the actual thank-you or success page before you spend anything. Per Google’s own conversion tracking documentation, this is the single setting that most directly affects bidding performance.
Turn on Enhanced Conversions and Consent Mode
Enhanced Conversions uses hashed first-party data (like an email a lead submits) to recover conversions that cookie-based tracking misses — increasingly important as browsers restrict third-party cookies. If you serve any users in the EEA or UK, Consent Mode v2 has been required since 2024 to keep measurement and remarketing functioning. Even outside those regions, building on first-party data now is future-proofing.
For B2B SaaS, plan for offline conversion import
This is the step most startups miss. If your real revenue event is a closed deal that happens days later in your CRM — not the form fill — a form submission alone tells Google to optimise for cheap leads, not good ones. Offline Conversion Import (via GCLID or Enhanced Conversions for Leads) sends the “this lead became a paying customer” signal back to Google so it learns to find more buyers, not more tyre-kickers.
Pro tip
Connect GA4 to Google Ads on day one, even if you optimise on in-platform conversions. The cross-tool view of bounce rate and engaged sessions is your early-warning system for landing pages that look fine but convert nobody.
Step 2: Build an account structure you won’t regret
A clean structure makes every later decision — budget shifts, pausing losers, scaling winners — fast and obvious. A messy one means you can’t tell what’s working. Keep it intent-based: group keywords by the searcher’s intent and your offer, not by some clever theme.
Use a consistent naming convention from the start, for example: Search | Brand | India, Search | Competitor | India, Search | Category-NonBrand | India. You’ll thank yourself when the account has 30 campaigns instead of three.
| Campaign type | Best for a startup when… |
|---|---|
| Search — Brand | Cheap, protects your name from competitors. Launch this first. |
| Search — Non-brand | You have clear high-intent keywords and verified tracking. Your core demand engine. |
| Performance Max | You already have conversion data to feed it. Avoid as your first campaign — it’s a black box that needs signal to behave. |
| Display / Demand Gen | Retargeting warm visitors or brand awareness — never for direct cold lead gen on a small budget. |
One more non-negotiable: in every Search campaign’s settings, turn off “Search Network partners” and the Display Network expansion until you’ve proven the core campaign. They’re on by default and they dilute your results.
Step 3: Set a budget that can actually exit the learning phase
Smart Bidding strategies enter a “learning phase” while the algorithm gathers data, and performance is unstable until they exit it. The blocker for startups is almost always budget that’s too thin to generate enough conversions to learn from.
Work backwards. If your category keywords cost roughly ₹40–₹120 per click in India (highly competitive software and AI terms run higher), and your landing page converts at, say, 4%, then one conversion costs you 25 clicks — call it ₹1,000–₹3,000. To gather the ~15+ conversions in 30 days that Google recommends before Target CPA bidding works reliably, you need a budget that can realistically reach that volume. Spreading ₹500/day across five campaigns guarantees none of them ever learn.
15+
conversions in 30 days before Target CPA bids reliably (Google guidance)
The practical move: concentrate budget on one or two campaigns until they’re producing consistent conversions, then expand. Depth beats breadth when you’re starting out.
Step 4: Pick the right bidding strategy for your stage
The biggest early error is handing the keys to an automated strategy before it has any data. Match the strategy to how much conversion history you actually have.
| Your situation | Use this bidding strategy |
|---|---|
| Brand-new account, zero conversion data | Manual CPC or Maximize Clicks (capped) to gather data cheaply |
| A few conversions, want volume | Maximize Conversions, watch CPA closely |
| 15+ conversions/month, stable cost | Target CPA — set it near your proven cost per lead |
| Tracking revenue/deal value | Target ROAS once you have enough value-based conversions |
The 6 mistakes that burn rupees fastest
- Launching before tracking is verified. Every day live without it is data you can never get back.
- Broad match with no negative keyword list. You’ll pay for “free,” “jobs,” “tutorial,” and “course” searches within hours. Start with phrase/exact and a tight negatives list.
- Leaving Search Partners and Display expansion on. Default settings, low intent, quiet drain.
- Accepting auto-applied recommendations. Turn these off in account settings — they routinely raise budgets and loosen match types without your sign-off.
- Ignoring the search terms report. This is where you find — and block — the irrelevant queries actually triggering your ads. Check it twice a week minimum.
- Starting with Performance Max. It needs a healthy conversion signal to optimise. Give it one first.
Warning
If a Google rep or the in-platform assistant pushes you to “raise the budget to capture more demand” in week one, pause. Until your conversion data is clean and your CPA is proven, more budget just buys more of whatever isn’t working yet.
Your pre-launch checklist
- Conversion action created and firing on the real success page (verified with Tag Assistant)
- Enhanced Conversions on; Consent Mode v2 live if you reach EEA/UK users
- GA4 linked; offline conversion import planned for CRM-closed deals
- Intent-based structure with a clear naming convention
- Search Partners and Display expansion switched off; auto-apply recommendations disabled
- Budget concentrated to reach 15+ conversions/month per campaign
- Bidding matched to your data stage; negative keyword list in place
Frequently Asked Questions
How much should a startup spend on Google Ads to start in 2026?
Enough for one or two campaigns to generate at least 15–30 conversions a month, not a thin spread across many. In India, that often means a focused daily budget rather than a token amount per campaign. Work backwards from your cost per click and landing page conversion rate to size it — if a conversion costs you ₹2,000, a ₹300/day budget will never produce enough data to optimise.
Do I really need conversion tracking before I launch?
Yes — it’s the one step you cannot skip. Smart Bidding optimises toward the conversions you report. Without tracking, Google optimises for clicks, you have no idea which keywords drive leads, and the budget spent before you fix it is unrecoverable. Set it up and verify it fires before turning on a single campaign.
Should an AI or software startup start with Performance Max or Search?
Start with Search. Performance Max is powerful but it’s a black box that needs an existing conversion signal to perform — feed it that signal first with a focused Search campaign. Once you have consistent conversion data, layering in Performance Max for expansion makes sense. Leading with it on a cold account usually means opaque, hard-to-diagnose spend.
Why are my ads getting clicks but no leads?
Usually one of three things: irrelevant search terms triggering your ads (check the search terms report and add negatives), a landing page mismatch between the ad promise and the page, or broad match pulling in low-intent queries. Audit the search terms report first — it’s the fastest way to find where the rupees are leaking.
How long before Google Ads starts working for a new account?
Expect a learning phase of roughly one to two weeks per Smart Bidding strategy while the algorithm gathers data, and a few weeks more to accumulate enough conversions to optimise confidently. Resist the urge to change bids, budgets, or targeting daily during this window — frequent edits reset learning and prolong the instability.
What’s the difference between Target CPA and Maximize Conversions?
Maximize Conversions spends your full budget to get as many conversions as possible without a cost ceiling — useful early to gather data. Target CPA aims for a specific cost per conversion and is the better choice once you know what a lead reliably costs you (Google recommends around 15+ conversions in the last 30 days before it bids dependably). Move from one to the other as your data matures.
Want your Google Ads set up right the first time — by a team that runs budgets for AI and software startups?
Written by the Varnan Digital performance marketing team. We build and manage paid search and social campaigns for early-stage AI and software companies, with a focus on clean tracking and measurable pipeline. Guidance here reflects Google’s official documentation and our hands-on account management — always validate settings against your own conversion data before scaling spend.