Data Sources - Marketing Analytics

The accuracy of every marketing KPI, attribution model, and dashboard is determined upstream, at the data source layer. Most marketing analytics failures - inconsistent CAC figures, attribution discrepancies, reports that contradict each other - originate not in the analysis but in the underlying data infrastructure. Understanding how each marketing data source works, what it produces reliably, and where it introduces error is foundational to building an analytics function you can trust.

This article covers the seven categories of marketing data sources, their integration patterns, and the data quality challenges specific to each. For the KPIs these sources feed, see Marketing KPIs. For the analytical techniques applied to this data, see Marketing Techniques.

The Data Integration Challenge

The average B2B marketing stack includes fourteen to twenty-two distinct tools. Each was selected to solve a specific problem. Together, they generate data in incompatible schemas, using different identity models, at different latencies, with different attribution conventions. The result is a fragmented data landscape where the same customer interaction appears differently in every platform - if it appears in all of them at all.

The architectural response to this fragmentation is a centralized data warehouse that receives data from all marketing sources via automated pipelines. Common warehouse choices include BigQuery, Snowflake, and Redshift. Data pipeline tools (Fivetran, Airbyte, Stitch) provide pre-built connectors for most major marketing platforms. On top of the warehouse, a transformation layer (dbt is widely used) applies consistent business logic, resolves identity, and produces the clean tables that analysts and dashboards consume. Platforms such as Plotono combine the pipeline orchestration and visualization layers, allowing teams to connect data sources, apply transformations, and build dashboards within a single environment rather than stitching together separate tools for each step.

This architecture is not optional for organizations that want reliable marketing analytics. Without it, every cross-channel analysis requires manual data joining, every metric is recalculated from scratch each time, and discrepancies between reports are the rule rather than the exception.

Category 1: Web Analytics - Google Analytics 4

GA4 is the primary web analytics platform for the majority of marketing organizations. It replaced Universal Analytics in July 2023, bringing a fundamentally different event-based data model.

What GA4 Measures

GA4 tracks user behavior across websites and apps using an event-based model where every interaction (page view, scroll, click, form submission, video play) is logged as an event with associated parameters. Key dimensions include: traffic source, medium, campaign, device, geography, landing page, and user journey sequences.

Core GA4 data assets:

Sessions and engaged sessions
Users (new vs. returning, total)
Traffic source and channel attribution
Event completions and conversion events
User journey and path exploration
Funnel visualization (multi-step conversion funnels)

GA4 Integration Considerations

GA4 exports to BigQuery natively for organizations with Google Analytics 4 properties, making it one of the cleaner data source integrations available. The raw BigQuery export provides event-level data with full session and user context.

Known GA4 data quality issues:

Data sampling: GA4 uses sampling in Explorations reports at high traffic volumes. BigQuery exports are unsampled and should be the basis for any analysis requiring precision.

Cookieless limitations: As third-party cookies deprecate, GA4’s ability to track cross-device and cross-session journeys degrades. Machine learning modeled conversions partially compensate but introduce estimation uncertainty.

Session versus user scope: GA4 reports some metrics at session scope and others at user scope, and the distinction matters for conversion rate calculations. Confirm the scope of every metric you extract.

UTM parameter enforcement: GA4 data quality is directly proportional to the consistency of UTM tagging across all marketing programs. Without enforced UTM conventions, traffic sources appear as “direct” or “none,” making attribution unreliable.

UTM Parameter Standards

UTM parameters are the mechanism by which GA4 (and any analytics platform) attributes sessions to marketing campaigns. Five parameters are available:

utm_source    - The platform or publisher (google, linkedin, mailchimp)
utm_medium    - The channel category (cpc, email, organic, social)
utm_campaign  - The campaign name (q1-2024-brand-awareness)
utm_content   - The specific ad or creative variant
utm_term      - The keyword (for paid search)

Enforcing consistent values across every person and every campaign that generates links is the single highest-leverage data governance action a marketing team can take. Define allowed values for each parameter, publish them in an internal wiki, and use a UTM builder tool to enforce them. Inconsistent UTM tagging cannot be fixed retroactively.

Category 2: Marketing Automation Platforms

Marketing automation platforms (MAPs) are the systems of record for lead data, email engagement, campaign workflows, and behavioral tracking at the contact level. They bridge anonymous web activity and identified lead records.

HubSpot

HubSpot is widely adopted across SMB and mid-market B2B organizations. Its CRM is built into the platform, making it easier to close the loop between marketing activities and contact/deal data without a separate CRM integration.

Key HubSpot data assets:

Contact records with lifecycle stage and lead source
Email send, open, click, and bounce events
Form submission events with source data
Workflow enrollment and completion history
Landing page and CTA performance
Deal pipeline data (HubSpot CRM)

Integration pattern: HubSpot’s API and native warehouse connectors (available via Fivetran, Airbyte, and HubSpot’s native data export) provide access to contact, company, deal, email event, and form submission data. The native Salesforce integration, when used, requires careful deduplication logic to prevent double-counting contacts that exist in both systems.

Marketo

Marketo (Adobe Marketo Engage) is the dominant MAP in enterprise B2B environments. It offers more sophisticated lead scoring, program tracking, and revenue attribution than HubSpot but requires more technical resources to operate.

Key Marketo data assets:

Person (lead/contact) records with scoring history
Program performance (email, webinar, content, event programs)
Activity log (detailed event stream for every contact interaction)
Revenue Cycle Model stage tracking
Revenue Attribution (multi-touch, using Marketo’s built-in model)

Integration considerations: Marketo’s REST API is comprehensive but rate-limited. Extracting high-volume activity logs requires careful pagination and incremental loading. The Marketo data model distinguishes between Leads and Persons (post-merge), which affects deduplication logic in warehouse integrations.

Pardot (Salesforce Marketing Cloud Account Engagement)

Pardot is Salesforce’s native MAP, tightly integrated with Salesforce CRM. Organizations running Salesforce as their CRM of record benefit from native object synchronization without the mapping complexity of third-party integrations.

Key Pardot data assets:

Prospects with scoring and grading
Engagement history (email, form, landing page, custom redirects)
Campaign influence data (aligned to Salesforce campaign model)
Automation rule execution history

Integration notes: Pardot data is accessible via Salesforce Reports and the Pardot API. For warehouse integration, Salesforce’s native connector (available in most ETL tools) provides access to Pardot objects as Salesforce-native objects when Connected Campaigns is enabled.

Category 3: CRM Systems - Salesforce

The CRM is the system of record for pipeline, opportunity, and closed-won revenue data. Connecting marketing analytics to CRM data is the mechanism that enables closed-loop attribution - the ability to trace from marketing touchpoint to closed revenue.

Salesforce Data Assets Relevant to Marketing

Campaign and Campaign Member objects: Campaigns in Salesforce capture lead and contact associations to marketing programs. Campaign Member records track each person’s response status. Multi-touch attribution models typically operate on Campaign Member data, attributing pipeline and revenue to campaigns based on the Campaign Member created dates relative to opportunity creation and close dates.

Lead and Contact objects: Source fields (Lead Source, Campaign Source) on Lead and Contact records provide marketing channel attribution at the person level. These fields are frequently poorly maintained. Audit the population and consistency of source fields before relying on them in attribution models.

Opportunity object: Pipeline amount, close date, stage, and owner data. The Opportunity Created Date is the reference point for most marketing attribution windows - the question attribution models answer is: which marketing activities touched the account or contact before the opportunity was created?

Account object: For ABM analytics, the Account is the primary unit of analysis. Account-level attributes (industry, size, territory, tier) enable segment-level pipeline analysis.

Salesforce Data Quality Issues

Lead deduplication: Salesforce environments at scale accumulate duplicate records, particularly when multiple inbound channels create lead records simultaneously. Deduplication before warehouse loading is essential for accurate attribution.

Stage progression consistency: Opportunity stage names and their meanings vary across organizations and sometimes across sales teams within an organization. Establish and enforce a consistent stage model before building pipeline analytics.

Campaign member response status: Response statuses (Sent, Responded, Registered, Attended) must be consistently populated by marketing automation sync rules. Inconsistent response status data renders campaign influence analysis unreliable.

Category 4: Paid Media Platforms

Paid media platforms generate high-volume impression, click, and conversion data at the campaign, ad group, and creative level. Each platform has its own attribution model, which almost always conflicts with your independent analytics.

Google Ads

Google Ads is the dominant platform for paid search and display advertising. It provides campaign, ad group, keyword, ad, and extension level performance data.

Key Google Ads metrics: Impressions, clicks, CTR, average CPC, Quality Score, conversion rate, conversions, cost per conversion, ROAS, Search Impression Share.

Attribution mismatch: Google Ads reports conversions using its own attribution model (which defaults to data-driven attribution within Google’s ecosystem). These numbers will differ from GA4-reported conversions and from CRM-attributed pipeline. Document the expected discrepancy and establish which number is authoritative for which reporting purpose.

Google Ads API and BigQuery Data Transfer: Google provides native BigQuery Data Transfer Service integration for Ads data, enabling automatic daily loading of campaign performance data at multiple levels of granularity.

Meta Ads (Facebook and Instagram)

Meta Ads Manager provides campaign, ad set, and ad level performance across Facebook and Instagram placements.

Key Meta Ads metrics: Reach, impressions, frequency, CPM, CPC, CTR, link clicks, landing page views, conversions, ROAS, cost per result.

Pixel and Conversions API: The Meta Pixel tracks web events and sends them to Meta for attribution. The Conversions API (CAPI) is the server-side complement that reduces loss from browser-side tracking limitations (ad blockers, iOS privacy restrictions). Organizations relying solely on the Pixel are underreporting conversions to Meta’s algorithm, degrading campaign optimization and reporting accuracy.

Attribution windows: Meta’s default attribution window is seven-day click and one-day view. Changing the window changes reported ROAS substantially. Align your reporting window selection to your sales cycle length.

LinkedIn Ads

LinkedIn is the primary paid B2B demand generation platform for most organizations targeting professional audiences. It offers targeting by job title, seniority, company size, industry, and function.

Key LinkedIn Ads metrics: Impressions, clicks, CTR, average CPC, conversions, lead gen form completions, cost per lead, company engagement.

LinkedIn Insight Tag and CAPI: Analogous to Meta, LinkedIn’s site tracking relies on the Insight Tag. Server-side conversion tracking via CAPI is available and recommended for improving measurement accuracy.

Account engagement data: LinkedIn’s Account Engagement product provides visibility into which accounts in your target list are engaging with your ads, enabling ABM campaign analysis at the account level. This is one of the few paid platforms that provides account-level data rather than purely person-level data.

Category 5: SEO and Content Tools

SEO tools provide competitive intelligence, keyword research, backlink data, and technical site health monitoring that web analytics platforms do not supply.

Ahrefs

Ahrefs provides backlink analysis, keyword ranking tracking, content gap analysis, and site audit functionality.

Key Ahrefs data assets:

Domain Rating and URL Rating scores
Referring domain count and new/lost backlink tracking
Organic keyword ranking positions and traffic estimates
Competitor keyword and backlink gap analysis
Site audit health scores

API access: Ahrefs provides API access for Domain Rating, organic keyword data, and backlink data, enabling automated loading into a warehouse for trend tracking.

SEMrush

SEMrush offers similar capabilities to Ahrefs with additional focus on paid competitive intelligence, position tracking, and content optimization tools.

Key SEMrush data assets:

Organic and paid traffic estimates by domain
Position tracking for target keyword sets
Competitive ad research (copy, landing pages, spend estimates)
Content audit and optimization recommendations

Google Search Console

Google Search Console is the only source of actual Google search performance data (impressions, clicks, CTR, and position) for your own domain. All third-party tools estimate this data. Search Console is authoritative.

Search Console BigQuery integration: Search Console exports are available via BigQuery Data Transfer Service, enabling page-level and query-level search performance data to be loaded into a warehouse for trend analysis and content attribution.

Category 6: Email Platforms

For organizations using email as a primary demand generation or customer nurture channel, standalone email platform data complements MAP data with deliverability and engagement detail.

Mailchimp

Mailchimp is widely used for content newsletters, announcement emails, and SMB marketing programs. Its API provides send, open, click, bounce, and unsubscribe data at the subscriber and campaign level.

Klaviyo

Klaviyo is the dominant email platform in ecommerce, with native Shopify integration and flow-based automation tied to purchase behavior. Its data model connects email engagement to purchase events, enabling direct revenue attribution from email programs.

Key Klaviyo data assets:

Email flow (automation) and campaign performance
Revenue attributed to email flows and campaigns
List growth and churn rates
Predictive lifetime value and churn risk scores (Klaviyo-computed)

Social listening tools provide brand mention monitoring, sentiment analysis, and share-of-voice measurement that social platform native analytics do not offer.

Sprout Social and Hootsuite provide cross-platform social publishing scheduling, engagement metrics aggregation, and audience analytics across LinkedIn, Instagram, Facebook, Twitter/X, and other platforms.

Brandwatch and Mention provide brand mention monitoring across social platforms and the broader web, enabling share-of-voice analysis and sentiment tracking against competitors.

Identity Resolution: The Integration Problem

Every marketing data source uses a different identifier: GA4 uses client IDs and user IDs; Salesforce uses lead/contact IDs; email platforms use email addresses; paid platforms use cookie and device identifiers. Connecting these identities across systems is the core technical challenge of marketing analytics.

The standard approach:

The MAP (HubSpot, Marketo, Pardot) serves as the identity backbone, associating anonymous web visitor IDs with identified email addresses at form submission.
The MAP syncs with Salesforce, associating the identified lead or contact record with a person-level interaction history.
The warehouse receives data from both systems and joins on shared keys (email address, lead ID) to create unified customer journey records.

Email address as primary key: Email address is the most reliable cross-system join key in B2B marketing. Protect it: enforce email normalization (lowercase, trimmed) at the point of collection, deduplicate aggressively, and validate format on ingest.

This data infrastructure, when functioning correctly, enables the attribution models and analytical techniques covered in Marketing Techniques and the dashboards described in Marketing Dashboards.