The most expensive mistake in Google Ads is not a high CPC. It is the wrong KPI.
If you measure an awareness campaign by ROAS, you will pause it – even though it works. If you measure a lead-gen campaign by ROAS without tracking offline conversions, you are optimizing on an estimate. If you measure an e-commerce campaign by impressions, you are burning budget.
I’ll show you here which KPI fits which campaign goal – and where the limits of each metric lie. The basic rule does not come from a blog post but from effectiveness research: Binet & Field show in their IPA study The Long and the Short of It (2013) that short-term ROI metrics systematically undervalue brand effects – while pure reach metrics say nothing about revenue. Therefore: first the goal, then the KPI.
The core problem: KPIs are not universal
Campaigns are often paused because they are measured against the wrong benchmark. A branding campaign is cancelled because the ROAS is “bad” – when reach was the goal, not direct revenue. Binet & Field quantify this as budget balance: around 60 % into long-term brand building, around 40 % into short-term activation. Anyone who measures everything by short-term return shifts this balance unnoticed towards activation – and wonders about shrinking margins.
The KPI must fit the goal. Not the other way round.
Google itself distinguishes three campaign goals that require different KPIs:
| Campaign Goal | What It Measures | Typical Use Case |
|---|---|---|
| Awareness | Visibility, reach | New brand, product launch, building local presence |
| Consideration | Interest, engagement | Traffic, video views, app installs, lead-magnet downloads |
| Conversion | Action, revenue | E-commerce sales, contact form, phone call |
The KPI Matrix: Which Metric Fits Which Goal
| Campaign Goal | Primary KPI | Secondary KPI | What It Really Measures |
|---|---|---|---|
| Awareness | CPM (Cost per Mille) | Reach, Frequency | How cheaply do I reach 1,000 people? |
| Consideration | CPC (Cost per Click) | CTR, Video View Rate | How cheaply do I get interest? |
| Conversion (E-Commerce) | ROAS (Return on Ad Spend) | CPA, Conversion Rate | How much revenue comes back per euro spent? |
| Conversion (Lead Gen) | CPA (Cost per Acquisition) | Lead Quality, Offline Conversion Rate | How much does a real customer cost? |
ROAS: When It Works, When It Lies
ROAS (Return on Ad Spend) is the standard KPI in e-commerce. But it only works if two conditions are met:
Condition 1: You have clear revenue per conversion.
An online shop with fixed prices? ROAS works. A service provider who gets enquiries and turns them into projects? ROAS only works if you know and track the average project value.
Condition 2: You track offline conversions.
When you generate leads (contact form, phone call), you only know days or weeks later whether they became customers. If you do not feed this information back into Google Ads, the system optimizes for leads – not customers. In practice, Smart Bidding then learns: “People who fill out the form are good.” Not: “People who fill out the form and buy are good.”
Contextualizing manufacturer numbers: Google’s own Economic Impact Studies cite an average of $2 revenue per $1 ad spend. This is a manufacturer number – useful as an order of magnitude, useless as a planning target. Your benchmark remains your own break-even calculation: Break-Even ROAS = 1 / profit margin. At 40 % margin the threshold is 2.5 – anything below is loss, regardless of what industry studies say.
Automation’s data hunger: Google recommends roughly 30 conversions in 30 days for Target CPA, at least around 15 for Target ROAS (search network). Below that, Smart Bidding lacks the learning basis – another reason why lead-gen accounts without offline import optimize in the wrong direction.
When ROAS makes sense:
- E-commerce with clear product prices
- Service providers with fixed package prices (e.g., “SEO analysis for €490”)
- Subscriptions with known lifetime value
When ROAS does not make sense:
- Awareness campaigns (no direct revenue)
- Lead gen without offline conversion tracking
- Long sales cycles (B2B, 6+ months to close)
Awareness KPIs: CPM, Reach, Frequency
When the goal is for people to get to know a brand, ROAS is the wrong benchmark. Instead, CPM, Reach and Frequency take centre stage.
CPM (Cost per Mille): What does it cost to reach 1,000 people?
Formula: CPM = (Ad Spend / Impressions) * 1,000
Reach: How many unique people saw the ad?
Frequency: How often did the average person see the ad?
Honest context on frequency: The “Rule of 7” – the assumption that a prospect needs around seven contacts before acting – is a widely used marketing heuristic. There is no official primary study for it; the rule is usually traced back to cinema advertising in the 1930s and treated in specialist literature as a heuristic, not a proven threshold (see GrowthMethod, n.d.). What effectiveness research actually supports: brand growth comes from reach and mental availability in the category (Ehrenberg-Bass Institute, How Brands Grow, 2010) – not from narrow frequency targeting on small groups. For practice this means: set Reach as the primary KPI, use Frequency as a control against one-off contacts – and do not hang either on ROAS.
Consideration KPIs: CPC, CTR, Engagement
When the goal is for people to engage with content (website visit, video view, download), CPC and CTR are the main KPIs.
CPC (Cost per Click): What does a click cost?
Formula: CPC = Ad Spend / Clicks
CTR (Click-Through Rate): What percentage of impressions become clicks?
Formula: CTR = (Clicks / Impressions) * 100
Benchmark for orientation: The classic WordStream benchmark dataset (thousands of US campaigns) cites an average of around 3.2 % CTR in the search network and around 0.5 % in the display network. Benchmarks are orientation values, not targets: industry, intent and ad position shift these figures significantly. [PRÜFEN: verify current LocaliQ dataset in Layer-2 review]
Why CTR matters: A high CTR signals relevance – and relevance lowers actual click prices via the Quality Score. It is a self-reinforcing cycle.
But beware: A high CTR is useless if the landing page does not deliver what the ad promises. Then visitors bounce immediately – the click was paid for anyway.
Conversion KPIs: CPA, Conversion Rate, ROAS
When the goal is a concrete action (purchase, enquiry, call), you cannot avoid CPA and Conversion Rate.
CPA (Cost per Acquisition): What does a conversion cost?
Formula: CPA = Ad Spend / Conversions
Conversion Rate (CVR): What percentage of clicks become conversions?
Formula: CVR = (Conversions / Clicks) * 100
Benchmark for orientation: WordStream cites an average of around 3.75 % CVR and around $49 cost per conversion for the search network – again as orientation, not as a target.
The Meta-KPI: Quality Score
The Quality Score (1–10) is not a business KPI, but the lever for all others. According to Google Ads Help, it consists of three components: expected click-through rate, ad relevance and landing page experience. Via Ad Rank it directly influences which CPC you pay and which position you get – at the same bid.
Keywords with Quality Score 3 or 4 that have been running for months are wasted budget: at a low score you pay more for every click than competitors with more relevant ads. These keywords should be paused or fundamentally reworked.
Lead Gen: The Offline Conversion Problem in Detail
Lead gen is the most complex campaign type because true success often only becomes visible weeks after the click.
The problem: Google Ads only sees the lead (form submission, phone call). It does not see whether the lead became a customer. In B2B decisions it is further complicated that purchasing decisions are made in groups and a large part of the buyer journey takes place without direct supplier contact (Gartner analyses on the B2B buying journey) [PRÜFEN: verify exact percentages in review]. The path click → close is therefore long – and a last-click ROAS simply does not capture it.
The solution in three steps:
- Track leads in Google Ads (form, phone, chat)
- Track customers in your CRM (which lead became a customer? which revenue?)
- Feed offline conversions back into Google Ads (via CSV/API import or Enhanced Conversions for Leads)
Only then does Smart Bidding optimize for real customers instead of form submitters. Without this step, the CPA per lead remains a nice number – and the CPA per customer an unknown quantity.
What Does This Actually Achieve?
| Campaign Goal | Right KPI | Wrong KPI | Consequence |
|---|---|---|---|
| Awareness | CPM, Reach, Frequency | ROAS, CPA | Campaign wrongly paused |
| Consideration | CPC, CTR | ROAS | Content not optimized |
| Conversion (E-Commerce) | ROAS, CPA | CTR, Impressions | Budget distributed inefficiently |
| Conversion (Lead Gen) | CPA + Offline Conversion | ROAS without OCI | Optimization for leads, not customers |
When That Is Not Enough
Knowing KPIs is the first step. Interpreting them correctly and drawing the right conclusions is the second. If you are just starting out, check the basics first: the article “Starting Google Ads? These 7 Questions Show If You’re Ready” clarifies budget, landing page and target values before the first euro.
If your data is there but the strategy is missing – or if there is no time left to wade through the report weekly – then the point has been reached where an SEA consultant makes sense. The Google Ads Account Check for €290 brings clarity: review KPIs, identify biggest budget drains, define levers to pull. How support costs are composed is shown in the open calculation.
Checklist
- Campaign goal defined (Awareness / Consideration / Conversion)
- Primary KPI chosen to match goal (see matrix above)
- Break-even ROAS calculated from own margin
- For lead gen: offline conversion tracking set up or planned
- Smart bidding data hunger checked (≥ 30 conversions/30 days for tCPA)
- Quality Score of main keywords checked (all > 6?)
- Vanity metrics (impressions, clicks) not used as primary success criterion
- Search terms report maintained regularly (negative keywords)
Sources
- Binet, L. & Field, P. (2013). The Long and the Short of It. IPA Institute of Practitioners in Advertising.
- Sharp, B. / Ehrenberg-Bass Institute (2010). How Brands Grow.
- Google Ads Help: Quality Factor & Its Components (expected CTR, ad relevance, landing page experience).
- Google Ads Help: Ad Rank & Auction Principle (Quality Score as CPC and position lever).
- Google Ads Help: Offline Conversion Import & Enhanced Conversions for Leads.
- Google Ads Help: Smart Bidding Recommendations (conversion minimum for tCPA/tROAS).
- Google Economic Impact Studies ($2 revenue per $1 ads) – manufacturer number, only as order of magnitude.
- WordStream/LocaliQ: Google Ads Industry Benchmarks (CTR ~3.2 % search, CVR ~3.75 % search, ~$49 cost/conversion) – classic dataset, verify current status in review.
- Gartner: B2B Buying Journey Research (buying groups, low direct-supplier-contact share) – verify percentages in review.
- GrowthMethod (n.d.): What is the Rule of 7? – secondary reference for marketing heuristic (no official primary source). https://growthmethod.com/what-is-the-rule-of-7/
Note: This article serves general information and does not replace individual campaign strategy. Benchmarks are orientation values, not guarantees.