Determining Budget and Bid Strategy Without Historical Data

Launching a new Google Ads account without historical data is one of the most challenging situations in digital advertising. Without previous campaigns, conversion volumes, or click data, Google's Smart Bidding algorithm has nothing to optimise towards. Yet you still need to make a decision: what budget do you set, and which bid strategy do you choose? Making the wrong call can waste budget during the learning phase or stunt growth before it even begins. This article walks through how to establish a data-driven starting point, which bid strategies suit a cold start, and how AdBrains uses proprietary AI technology to structurally improve this critical launch period.

Why Historical Data Matters, and What to Do Without It

Google Ads operates on machine learning. Smart Bidding strategies like Target CPA and Target ROAS analyse dozens of signals simultaneously: search term, device, location, time of day, audience, and more. The more conversion data available, the more precisely the algorithm can adjust its bids. Google itself recommends at least 30 to 50 conversions per month per campaign before Smart Bidding performs optimally. Without historical data, this foundation is entirely absent.

That does not mean you have to start blind. There are multiple data points outside your own account that you can use as a starting point. The Google Keyword Planner provides search volume and CPC estimates. Industry benchmarks offer average conversion rates. Internal business data such as average order value or lead value round out the picture. By combining these external inputs, you can form a well-reasoned assumption about which budget and bid strategy are realistic for your situation.

Setting Your Budget Using External Data Points

The first question every new advertiser asks is: how much budget do I need? The answer starts not with an arbitrary number, but with a calculation based on your goals. Use the following framework as a starting point:

• Establish your target CPA or revenue goal: what can a conversion cost you? For an online course platform like HACCP-cursus.com, a realistic target CPA based on margin might be 10 to 15 euros.
• Use Keyword Planner for CPC estimates: search your core keywords and review estimated cost per click. This indicates how many clicks your budget will generate.
• Work backwards from conversion rate: if the industry benchmark for comparable services is a 2 to 4 percent conversion rate, you know how many clicks you need per conversion.
• Set a minimum budget for the learning phase: Google typically needs 2 to 4 weeks to complete the learning phase. Your budget must generate enough clicks during that period to build statistically meaningful data.
• Add a safety margin: plan 20 to 30 percent extra budget for testing periods, as actual CPC can diverge from estimates.

As a practical example, consider Clima-Active.nl, an air conditioning and heat pump installation company working on quote requests. At the start of a new Google Ads account, Clima-Active has no historical data. The Keyword Planner shows that search terms like "airco installeren" and "warmtepomp offerte" carry an average CPC of 2 to 4 euros. At an expected conversion rate of 3 percent (industry average for installation services), approximately 33 clicks are needed per quote request. To collect at least 10 requests per month to feed the learning phase, roughly 330 clicks are needed, equating to a minimum budget of 660 to 1,320 euros per month. This gives a justified starting point rather than a guess.

Manual vs. AI-Driven Campaign Launch

Comparison of approach for new campaigns without historical data

Manual management

• Budget based on gut feeling or industry averages
• Manual CPC as default bid strategy
• Weeks of waiting before first optimisations
• No systematic learning phase monitoring
• Risk of waste in the first month
• Ad Strength rarely actively monitored

AdBrains AI approachAdBrains AI

• Budget backed by Keyword Planner data and margin targets
• Smart bid strategy selection based on conversion volume forecast
• Daily learning phase monitoring via multi-agent system
• Automatic escalation on deviating CPA or ROAS
• Keyword Incubator protects production campaigns
• RSA improvement system active from day one

In e-commerce the same logic applies, but via ROAS rather than CPA. For ToetsJeKennis.nl, an online exam and course platform with an AOV of 50 euros, the minimum ROAS depends on the margin. Assuming a 60 percent margin, break-even ROAS is 1.67. A target ROAS of 300 percent then provides a healthy margin above variable costs. In budget terms: if the estimated CPC is 0.80 euros and the conversion rate is 3 percent, 100 euros generates roughly 125 clicks and 3 to 4 purchases, delivering 150 to 200 euros in revenue — in line with the ROAS target.

Choosing the Right Bid Strategy for a Cold Start

Selecting a bid strategy for a new account is one of the most consequential decisions in the launch phase. Many advertisers make the mistake of immediately choosing Target CPA or Target ROAS, while the algorithm simply does not yet have enough data to act on. This leads to an extended learning phase, erratic behaviour, and unnecessarily high cost per click.

The recommended progression for scaling bid strategies looks like this:

1. Phase 1: Maximize Clicks Use this strategy to rapidly build click data and search term data. Set a maximum CPC bid as protection against extremely high click prices. This phase typically lasts 2 to 4 weeks.
2. Phase 2: Manual CPC or Enhanced CPC Once you have insight into which search terms convert (even if conversion volumes are still low), switch to manual control with Enhanced CPC as a bridge to full automation.
3. Phase 3: Maximize Conversions When at least 15 to 20 conversions have been collected, Maximize Conversions is a safe step. The algorithm now actively optimises for conversions without holding to a strict bid target.
4. Phase 4: Target CPA or Target ROAS Only when there are 30 to 50 conversions per month is Smart Bidding with an explicit target truly optimal. Set the target loosely at first (higher than your final goal) to give the algorithm room to collect data.

This phased model applies equally to lead gen campaigns like Clima-Active.nl and to e-commerce campaigns like E-4motion.com, the webshop for new electric folding bikes. At E-4motion.com, the average order value is substantial, making it even more important not to switch bid strategies too early: one poor algorithmic decision on high-AOV products costs far more than on low-priced items.

A critical enabler across every bid strategy choice is flawless conversion tracking from day one. Without reliable conversion measurement, you feed the algorithm incorrect data. Server-side tracking via a dedicated sGTM infrastructure is the gold standard here: first-party signals are richer and more accurate, accelerating the learning phase considerably.

How AdBrains AI Automates and Accelerates This Launch Trajectory

Faster Out of the Learning Phase with AI

Average results from AI-driven campaign launches without historical data

2-3x

Faster Smart Bidding learning phase

35%

Less budget waste in weeks 1-4

89%

Of campaigns hit Target CPA within 6 weeks

daily

AI monitoring of bid strategy and learning phase

AdBrains has developed proprietary AI technology that tackles exactly this problem: how do you start a new Google Ads account in a structured, data-driven way even without historical data? Our approach combines several specialised AI modules that work together to accelerate the learning phase and minimise budget waste.

First, our system always deploys a Keyword Incubator for new campaigns. New keywords never launch directly into a production campaign. Instead, they start in a separate incubator campaign with a lower budget and broad match settings. The AI monitors daily which keywords show sufficient impressions and conversion intent, and automatically decides when a keyword is ready to be promoted to the production campaign. This prevents a new keyword from immediately consuming a large share of budget without evidence of conversion potential.

Second, our multi-agent verification system acts as a safety net for every bid strategy decision. Four independent AI agents review every proposed change before it is executed. This is especially valuable during the launch phase: if one agent wants to switch a campaign to Target CPA but insufficient conversions have accumulated, the other agents automatically block that change until the conditions are met.

Third, our system monitors bid strategy progression daily. As soon as thresholds for the next bid strategy are reached, the AI automatically proposes and executes the transition after verification. This is the strategy-switch system: campaigns are never kept in a suboptimal bid strategy longer than necessary, but are also never switched too early.

Our RSA improvement system is active from day one. Even without historical ad performance data, the AI analyses the Ad Strength of every Responsive Search Ad and automatically provides concrete improvement suggestions based on landing page content and search intent analysis. This ensures ad quality does not lag behind while the algorithm is still learning.

Finally, our server-side signal enrichment via a dedicated sGTM infrastructure enriches conversion signals with first-party data. This makes Smart Bidding more precise and faster, even during the learning phase. Advertisers using server-side tracking see significantly more conversions tracked accurately, leading directly to better bid strategy steering and a shorter path to a stable Target CPA or Target ROAS.

Search Term Mining as a Launch Accelerator

Keyword Incubator: Flow to Production

Percentage of keywords passing through each phase (PERCENT-mode)

Entered into Incubator

All new keywords start here safely

Sufficient impressions

Keywords with enough search volume proceed

First conversions achieved

Only keywords with proven intent pass

Promoted to production

Only the strongest keywords reach production

28% of incubated keywords reach the production campaign

Alongside bid strategy, monitoring search terms in the early phase is crucial. Without historical data you do not know which queries are triggering your ads. Broad match keywords, which are valuable in the early phase for rapidly building data, can also attract irrelevant traffic. This traffic costs budget and distorts the signals Smart Bidding uses to learn.

AdBrains solves this with automated search term mining. Every day our system analyses all search terms that generated clicks. Irrelevant search terms are automatically detected and added as negative keywords, keeping budget consistently focused on search terms with conversion potential. This is even more valuable in the early phase than later: every euro not wasted on irrelevant traffic can instead be deployed to reach the Smart Bidding threshold faster.

• Daily automated review of all triggered search terms
• Automatic negative keyword addition for irrelevant queries
• Identification of high-intent search term clusters for campaign expansion
• Integration with the Keyword Incubator for structured keyword promotion
• Continuous quality improvement of the data fed into Smart Bidding

For a new campaign launched for LeroyBrouwer.nl, a lead generation specialist, this means the system actively filters out irrelevant queries during the first weeks while identifying the most valuable search term segments. The result is faster data accumulation and, more importantly, higher-quality data that steers the algorithm in the right direction from the very start.

Common Mistakes During a Cold Start

Knowing what to do is important, but so is knowing what to avoid. The following mistakes are the most common and costly when starting without historical data:

• Choosing Target CPA or Target ROAS too early: the algorithm has insufficient data and bids erratically, leading to extreme cost fluctuations.
• Setting too low a budget: an insufficient daily budget dramatically extends the learning phase because not enough data is collected.
• Not setting up conversion tracking before campaigns go live: every click before correct conversion measurement is lost learning data.
• Placing all keywords in broad match without a negative keyword list: this attracts irrelevant traffic and lowers the quality of your first dataset.
• Adjusting budget mid-learning-phase: significant budget or bid changes during the learning phase partially reset the counter and extend the time to optimal performance.
• Launching too many campaigns and ad groups simultaneously: spreading conversion data too thin means no individual campaign reaches the Smart Bidding threshold.

Frequently Asked Questions

How do I set a realistic starting budget with no prior Google Ads experience?

Start with the Google Keyword Planner to find the average CPC for your most important keywords. Combine this with a realistic estimate of your conversion rate (use industry benchmarks if you have no own data) and calculate how many clicks you need for at least 20 to 30 conversions per month. This is the minimum budget for an effective learning phase. Add 20 to 30 percent on top as a safety margin for deviations in actual CPC.

Which bid strategy is safest to start with?

For a cold start, Maximize Clicks with a maximum CPC cap is the safest choice. This strategy optimises for clicks rather than conversions, allowing you to build data quickly without the algorithm waiting for conversion signals that do not yet exist. Once you have 15 to 20 conversions, switch to Maximize Conversions. Target CPA or Target ROAS should only be applied once you have 30 to 50 conversions per month.

How long does the Smart Bidding learning phase last and how can you shorten it?

The official Smart Bidding learning phase typically lasts 1 to 4 weeks, depending on conversion volume. The more conversions per week, the faster the algorithm converges. You can shorten the learning phase by setting sufficient budget (at least ten times your target CPA per day), by using broad match to build search term data quickly, by adding micro-conversions as additional optimisation signals (such as time on page or form start), and by implementing server-side tracking for more complete data collection. AI-driven systems like AdBrains monitor the learning phase daily and automatically intervene when progress stalls.

Should I choose Search or Performance Max for a cold start?

For a cold start, Search campaigns are preferable to Performance Max. Search campaigns give you more control over keywords, bid strategies, and targeting, which is essential when you want to understand the first data points. Performance Max works best when there is already conversion volume available and campaigns have proven strong signals. Once your Search campaigns perform stably and have accumulated sufficient conversion data, expanding to Performance Max is a logical next step. AdBrains monitors this trajectory automatically and detects the right moment for campaign expansion.