Marketing KPIs are useful only when they support a decision. Impressions, clicks and form submissions describe activity. Lead quality, won customers, revenue and contribution margin show whether that activity was commercially relevant.
GA4 observes behaviour on a website or app. A CRM records the sales process. Store and finance systems confirm orders, refunds and recognised revenue. When these levels are reviewed separately, the business receives many reports but no shared picture.
This guide explains how SMEs can build a manageable KPI chain, align definitions between marketing and sales and use the evidence to make decisions about channels, budgets, websites and lead handling.
Which marketing KPIs does an SME actually need?
An SME does not need the largest possible metric set. It needs a small number of connected measures across several decision levels. A reliable chain begins with the commercial outcome and works backwards to marketing activity.
New customers, confirmed revenue, contribution margin, repeat purchases or available capacity.
Qualified leads, proposals, close rate, sales cycle and documented loss reasons.
Successful enquiries, bookings, purchases and cost per relevant conversion.
Product views, form starts, baskets, landing-page use and important events.
Impressions, reach, clicks, CPC, sessions and other early channel signals.
Practical rule: a metric becomes a KPI only when an owner knows which decision follows a material change.
The five levels are a Salestudia working model, not an official GA4 framework. Businesses may use different labels but should preserve the relationship between activity and commercial outcomes.
Distinguish a metric, KPI, target and threshold
| Term | Meaning | Example | Decision function |
|---|---|---|---|
| Metric | A measured value without automatic strategic priority. | Landing-page sessions. | Provides context but does not need to trigger an action. |
| KPI | A metric that supports an important objective or process decision. | Qualified leads per month. | Influences budget, offer, campaign or sales process. |
| Target | A justified desired value for a defined period. | Required number of new customers. | Describes the planned state, not automatically a forecast. |
| Threshold | A value that triggers a review or intervention. | An excessive share of unsuitable enquiries. | Starts analysis, limitation or a process change. |
A KPI without a target, owner and next action quickly becomes a decorative dashboard number. A reliable digital marketing strategy for SMEs therefore connects measurement with business objectives, channel roles and an actionable roadmap.
What role does GA4 perform?
GA4 is an event-based analytics platform. It can measure website and app interactions, traffic sources, user journeys, key events and ecommerce data. It is not a complete CRM or the company’s accounting system.
Technical interaction
A page view, form start, download, product view or another measured behaviour.
Question: what happened on the website or app?
Important action
An event considered particularly important to the success of the business.
Question: which measured action has greater commercial significance?
Advertising optimisation
A conversion made available to Google Ads for reporting or bidding.
Question: which outcome should advertising support?
Google explains the current Event → Key Event → Conversion relationship in its official guide to conversions and key events in Google Analytics.
Not every event should become a key event, and not every key event should become a primary advertising conversion. A scroll, video view or form start may be diagnostically useful without being equivalent to a confirmed lead or purchase.
Measure lead generation as a complete process
For service businesses and B2B, measurement does not end at `generate_lead`. A submitted form may be spam, outside the service area or unrelated to the offer. Later CRM stages are therefore required.
| Stage | Possible GA4 logic | CRM meaning | Suitable KPIs |
|---|---|---|---|
| Enquiry received | generate_lead | A contact or form was captured successfully. | Leads, cost per lead, conversion rate. |
| Lead reviewed | qualify_lead or disqualify_lead | The contact meets or fails the agreed qualification criteria. | Qualification rate, cost per qualified lead. |
| Lead worked | working_lead | Sales contacted or actively processed the lead. | Response time, contactable leads, handling rate. |
| Sale won | close_convert_lead | The lead became a customer or confirmed order. | Close rate, CAC, revenue per new customer. |
| Sale lost | close_unconvert_lead | The opportunity was lost for a documented reason. | Loss reasons, sales cycle, proposal quality. |
Google lists these lead-generation events in its guide to recommended GA4 events. The CRM stages may use different names; the important requirement is a documented mapping.
The CRM provides the quality behind the lead
GA4 can show that an enquiry occurred. The CRM should explain whether it was relevant, contactable, qualified, proposed, won or lost. Without that feedback, marketing often optimises towards the easiest rather than the most valuable contacts.
- unique lead ID;
- created date and owner;
- service, product or enquiry type;
- region and language where commercially required;
- source, medium and campaign;
- current status;
- qualification or loss reason;
- order, revenue or expected value;
- close date.
- When does a contact become a genuine lead?
- Which criteria define a qualified lead?
- When does an opportunity or proposal count?
- When is a sale considered won?
- How are spam, duplicates and test contacts handled?
- How are cancellations and refunds treated?
- Which system owns each status?
- Who may change a status later?
Marketing, sales and management must use the same definitions. Otherwise, “lead” may mean a form submission in the advertising platform, a reviewed contact in the CRM and an opportunity in the management report.
Ecommerce: purchase revenue is not automatically profit
In an online store, the technical conversion sits closer to revenue, but the commercial assessment still does not end with the `purchase` event.
Order event
A confirmed transaction with transaction ID, value, currency and item data.
Check: was a genuine order created and counted only once?
Returned value
A full or partial refund related to the original transaction.
Check: do the value, items and transaction ID match?
Business view
Confirmed revenue after cancellations and returns, together with relevant variable costs.
Check: is the order economical after media, goods and fulfilment costs?
Google Analytics uses a transaction ID to reduce duplicate purchases registered with the same identifier. The requirement is explained in the guide to minimising duplicate key events with transaction IDs.
At minimum, review purchases, confirmed revenue, refunds, return rate, average order value and contribution margin separately.
A shared identifier connects the systems
- Lead generation: an internal lead or opportunity ID.
- Ecommerce: a unique transaction or order ID.
- Campaign context: UTM parameters, click identifiers or documented source and medium fields.
- Multiple forms: a form ID or clearly defined enquiry route.
- Offline sale: a connection between the original lead and the won order.
The identifier should not contain unnecessary personal information. Access, storage, consent and transmission need to match the actual technical and legal configuration.
Why Meta, Google Ads, GA4, CRM and finance report different totals
Assigns outcomes to ad contacts using its own attribution, windows and identity logic.
Analyses events, sources, journeys and key events according to the Analytics configuration.
Records lead quality, orders, cancellations, returns and process ownership.
Assesses invoices, payments, costs, margin and relevant accounting periods.
The objective is not to force every system to show one identical total. The business should be able to explain the differences and identify which system governs each decision.
Do not confuse attribution with commercial truth
A customer journey may contain several touchpoints: social creates awareness, organic search supports evaluation and paid search leads to an enquiry. An attribution model distributes credit according to a selected method; it does not prove sole causation.
The official key-event attribution paths report shows touchpoints, path length, time to a key event and how credit is distributed.
- Platform revenue is an attribution value, not automatically accounting revenue.
- Several platforms may claim part or all of the same order.
- Direct may appear when earlier information is unavailable.
- Consent, device changes and offline stages limit observability.
- Attribution supports decisions but does not replace controlled testing.
Important formulas for lead generation and ecommerce
Shows the share of enquiries that meet the agreed criteria.
Shows how often a suitable contact becomes a customer.
Useful for the first funnel stage but says nothing about quality.
Connects acquisition cost with CRM qualification.
Requires a documented cost scope and a clear new-customer definition.
Describes attributed advertising revenue, not automatically profit.
“Attributable cost” may include media spend alone or also management, creative, software and landing-page work. The scope must be visible. Our guide to the marketing budget for SMEs in Germany explains why media, production, technology and personnel should be planned separately.
Return CRM outcomes to advertising systems
When a sale occurs only after a call, proposal or extended sales process, the initial online enquiry should not remain the only optimisation signal. Suitable CRM outcomes may be returned to an advertising system after technical and legal review.
Google describes enhanced conversions for leads as an upgraded offline-conversion import method that can connect permitted first-party data and CRM outcomes with earlier online interactions.
CRM stages need to be stable before automation. An import cannot repair inconsistent qualification and must not return tests, spam or incorrectly closed records as high-value conversions.
Assess consent and data completeness together
Measurement in Germany needs to respect the user’s choice. Consent may mean that not every contact can be observed completely. This is not a reason to bypass a refusal technically.
Google explains that Consent Mode communicates consent status to Google tags and adjusts their behaviour. The CMP, legal basis and implementation still require separate review.
- do not mix observed and modelled data without explanation;
- document consent states during tracking tests;
- do not assume CRM and finance data are automatically consent-free marketing data;
- show missing attribution as uncertainty;
- do not base decisions on apparently complete platform totals alone.
Our guide to Google Ads conversion tracking with GA4 and Consent Mode covers the technical relationship between website events, Analytics, advertising and consent in more detail.
A minimal KPI report for an SME
New customers, revenue, contribution margin, CAC, capacity and priority risks.
Spend, leads or purchases, qualified outcomes, CPQL, ROAS and channel development.
Response time, qualification, proposals, close rate, sales cycle and loss reasons.
Consent, missing IDs, duplicates, tracking changes and unresolved differences.
Use a recurring review process rather than a static dashboard
Review data quality
Check outages, duplicates, missing CRM stages, cancellations and consent changes.
Question: is the evidence reliable enough for this decision?
Locate the funnel issue
Identify whether demand, the website, qualification or sales is losing the outcome.
Question: where is the largest commercially relevant gap?
State a hypothesis
Document a possible cause, supporting evidence, limitations and alternatives.
Question: which explanation can be tested?
Assign the decision
Define the change, owner, period and the next observation point.
Question: who does what, and when will the decision be reviewed?
Operational anomalies may require weekly review. Campaign and lead quality often need a monthly cycle. Strategy, channel roles and budget assumptions should be reassessed at longer intervals or after material market changes.
Distribute responsibility across marketing, sales and the business
| Area | Business | Marketing | Sales, finance and technology |
|---|---|---|---|
| Objectives | Defines growth, capacity, margin and priorities. | Translates objectives into channel and measurement logic. | Confirms feasibility, sales process and financial definitions. |
| Events | Confirms commercial meaning. | Defines event plans, key events and campaign use. | Implements triggers, IDs, CRM stages and QA. |
| Lead quality | Defines the customer profile and exclusion criteria. | Analyses sources and campaigns. | Maintains stages, reasons and close data. |
| Revenue | Defines relevant commercial measures. | Uses attribution as a marketing perspective. | Confirms order, payment, cancellation, return and margin. |
| Decision | Prioritises budget and resources. | Recommends channel and campaign changes. | Confirms effects on sales, systems and commercial outcome. |
Common mistakes and checklist
- treating reach or clicks as a commercial result;
- counting every form as an equal lead;
- having no shared qualified-lead definition;
- using GA4 as a CRM or accounting system;
- confusing attributed platform revenue with confirmed revenue;
- ignoring cancellations and returns;
- optimising CPL while CPQL or CAC increases;
- comparing systems without shared IDs;
- adding attribution totals and presenting them as total revenue;
- updating a dashboard without recording a decision.
- the commercial objective and period are defined;
- every KPI has an owner and decision function;
- events and CRM stages are documented;
- lead, opportunity and customer definitions agree;
- transaction and lead IDs are stable;
- marketing-cost scope is explicit;
- revenue, cancellations and returns come from the system of record;
- consent and data gaps are visible;
- platform differences are explained;
- the next action and review date are assigned.
Methodological limits: what KPI relationships do not prove
- Demand, seasonality, price, competition and capacity may change simultaneously.
- Attribution distributes credit but cannot observe every touchpoint completely.
- CRM stages depend on consistent maintenance and stable definitions.
- Revenue may result from repeat purchases or earlier interactions.
- Small datasets create large percentage variations.
- More qualified leads do not guarantee a higher margin.
- A low CAC may be temporary and increase later.
A precise statement is therefore not: “The new channel caused the revenue.” A more defensible conclusion is: “During the reviewed period, qualified leads and confirmed orders associated with this channel increased; attribution, seasonality, other touchpoints and sales-process changes still need to be considered.”
Editorial note: The linked official sources were checked on 4 August 2026. Google Analytics features, advertising platforms, CRM integrations and legal requirements can change. The KPI cascade and review structure are Salestudia editorial working models, not official Google standards.
Frequently asked questions about marketing KPIs, GA4 and CRM
How many marketing KPIs should an SME use?
As few as possible and as many as the decisions require. One to three principal KPIs per decision level, supported by diagnostic metrics, will often be sufficient.
Is a form submission already a qualified lead?
No. It initially confirms only that an enquiry was received. Qualification applies agreed criteria such as need, region, budget, audience or technical fit.
Can GA4 show the company’s complete revenue?
GA4 can analyse correctly sent ecommerce revenue and other values. Accounting, payments, margin, cancellations, offline revenue and financial periods remain the responsibility of other systems.
Which matters more: CPL or cost per qualified lead?
CPL measures cost to the initial enquiry. Cost per qualified lead is more informative for many service businesses because it includes CRM quality. The metrics answer different questions.
Why does every advertising platform show different revenue?
Platforms use different attribution models, windows and identity signals. Several systems may assign the same order differently.
Must every SME integrate GA4 with a CRM?
Not immediately through automation. Lead definitions, stages and IDs should be reliable first. A regular manual reconciliation may be better than an automated but inaccurate integration.
Is ROAS sufficient for ecommerce decisions?
No. ROAS focuses on attributed revenue and media spend. Cost of goods, discounts, shipping, payment fees, returns and other marketing costs also need to be considered.
How often should KPIs be reviewed?
Technical failures and major anomalies need prompt review. Campaign and lead quality often require monthly analysis, while strategic objectives and channel roles are reviewed less frequently.
Turn marketing data into a manageable commercial system
A reliable KPI system connects GA4 events with CRM stages, sales, confirmed revenue, cost and clear ownership. Its value lies not in the dashboard but in better decisions.
Salestudia develops digital strategies for businesses in Germany, connecting objectives, channels, websites, advertising, analytics and sales data within a prioritised roadmap and transparent measurement model.
Develop your KPI and digital strategy with Salestudia →