1. Tracking and measurement audit
A review of every tag, event, conversion action, UTM convention and data flow from site to ad platforms to CRM. Each issue is ranked by how much it distorts the numbers you use to spend money.
We fix how your marketing data is captured, connect it across GA4, ad platforms and your CRM, and turn it into a short list of decisions every month.
Most marketing teams have more dashboards than decisions. Ad platforms each claim the same sale, GA4 shows a different number, the CRM shows a third, and leaders quietly go back to instinct. Marketing intelligence fixes the inputs first, then builds the analysis on top, so the numbers your leadership sees are numbers they are willing to use.
Your accounts · Your data · Your decisions
Marketing intelligence is the practice of collecting accurate marketing data, connecting it across platforms and turning it into decisions about where to spend, what to change and what to stop. It covers three connected jobs: capturing conversions and behavior correctly, organizing that data in analytics and warehouse tools, and analyzing it with methods that hold up when platforms disagree.
The term is sometimes used for competitor and market research. On this page it means measurement: the data that tells you which marketing works. At The Viral Ideas we deliver it through three services: Conversion Tracking, GA4 Consulting and Marketing Analytics.
| Standard marketing reporting | Marketing intelligence | |
|---|---|---|
| Starts from | Whatever the platforms report by default | Business questions written down before any tag is added |
| Data sources | Each platform read on its own | Ads, GA4, CRM and spend joined on shared IDs and definitions |
| Conversion counts | Each platform claims its own number | One reconciled count, with known gaps explained |
| Data quality | Checked when a number looks odd | Monitored daily with alerts on breaks |
| Attribution | Last click, or each platform's own model | Several methods compared, calibrated with experiments |
| Output | Dashboards with dozens of charts | A short list of decisions with the evidence behind each |
When leadership ignores the dashboard, the cause is rarely the dashboard. It is usually one of these problems underneath it. Each comes with a quick test you can run this week.
| Symptom | Likely cause | Quick test |
|---|---|---|
| Google Ads, Meta and LinkedIn together claim more conversions than you actually got | Each platform counts by its own attribution window and model | Add up platform-reported conversions for last month and compare with your CRM |
| GA4 conversions do not match form submissions in the CRM | Duplicate or missing event triggers, spam, or consent gaps | Compare one week of GA4 key events with CRM records created from the same forms |
| A large share of traffic shows as direct or unassigned | Missing or inconsistent UTM parameters, redirects stripping parameters, cross-domain gaps | Check what share of last month's sessions were direct or unassigned |
| Numbers changed overnight with no campaign change | A site release removed or duplicated a tag | Check whether anyone is alerted when key event volume drops |
| Ad platforms optimize toward cheap form fills that never buy | Only the first conversion is sent back; sales outcomes never reach the platform | Ask whether qualified leads or revenue are imported into the ad platforms |
| Year-on-year comparisons are impossible | GA4 explorations limited by data retention; no raw data export | Check the GA4 retention setting and whether a BigQuery export exists |
| Different teams report different numbers for the same metric | No shared definitions or single source for each metric | Ask two teams to define a "conversion" and compare the answers |
The same Gartner survey found that a third of respondents saw decision makers pick data to support existing opinions. That happens most when the data is easy to dispute. Clean collection and agreed definitions remove the easy excuses, which is why we fix inputs before we build reports.
We build marketing intelligence in four layers. A problem in a lower layer cannot be fixed in a higher one: no model can correct a conversion that was never recorded.
| Layer | Question it answers | What we build | Covered in detail by |
|---|---|---|---|
| 1. Collect | Did we record what actually happened? | Measurement plan, data layer, browser and server-side tags, conversion APIs, CRM and offline imports, consent integration | Conversion Tracking |
| 2. Organize | Can people find and trust the data? | GA4 property design, key events, custom dimensions, channel groups, BigQuery export, data warehouse models | GA4 Consulting |
| 3. Analyze | What is working, and what would happen if we changed spend? | Attribution comparison, marketing mix modeling, incrementality tests, forecasting, funnel and cohort analysis | Marketing Analytics |
| 4. Decide | What do we do next month? | Role-based dashboards, a monthly insight memo and a decision log | This page, Section 13 |
The tools that feed these layers, such as the CRM, marketing automation and the customer data platform, are set up through our Martech service. Marketing intelligence makes sure what flows out of them is measured correctly.
Six services, run alone or together. Most clients start with the tracking audit, because every later step depends on its findings.
A review of every tag, event, conversion action, UTM convention and data flow from site to ad platforms to CRM. Each issue is ranked by how much it distorts the numbers you use to spend money.
Browser and server-side tagging, platform conversion APIs, enhanced conversions and CRM imports that send qualified outcomes back to ad platforms. Go deeper: Conversion Tracking
Property and data stream design, key events, custom dimensions, cross-domain tracking, filters, audiences and BigQuery export, documented so your team can maintain it. Go deeper: GA4 Consulting
Attribution comparison, marketing mix modeling, incrementality experiments, forecasting and budget scenarios that show where the next dollar should go. Go deeper: Marketing Analytics
Ad spend, GA4, CRM and revenue data joined in a warehouse, modeled on shared definitions and presented in dashboards built for each audience rather than one screen for everyone.
Automated checks on event volumes, tag health, UTM compliance and match rates, with alerts to a named person when something breaks, usually after a site release.
The three services are layers of one system. Tracking decides what gets recorded, GA4 decides how it is organized, and analytics decides what it means. Buying one without the others usually moves the problem rather than solving it.
| Conversion tracking | GA4 consulting | Marketing analytics | |
|---|---|---|---|
| Main job | Record every conversion once, accurately, and send it where it is needed | Structure behavior data so it can be explored, trusted and exported | Turn joined data into spend and channel decisions |
| Typical work | Tag audits, data layer, server-side tagging, conversion APIs, offline and CRM imports | Property design, key events, dimensions, cross-domain, filters, BigQuery export, training | Attribution comparison, marketing mix modeling, experiments, forecasting, dashboards |
| Signs you need it | Platforms over-report; ad bidding optimizes toward junk leads | Reports nobody trusts; direct or unassigned traffic is high; no history beyond a few months | Budget decided by habit; no view of what is incremental |
| Main output | A verified conversion map and working integrations | A documented GA4 property and warehouse feed | A model, a test plan and a monthly decision memo |
| Depends on | A measurement plan | Correct tracking | Correct tracking and organized data |
We map every conversion that matters, from form fills to closed deals, and make sure each is recorded once and sent to GA4, the ad platforms and the CRM with the right value. Go deeper: conversion tracking services
We design or repair your GA4 property so the data is clean, consistent and exportable, then train your team to use explorations and reports without guessing. Go deeper: GA4 consulting services
We combine attribution, modeling and controlled experiments to estimate what each channel actually adds, and turn that into a budget recommendation. Go deeper: marketing analytics services
Most tracking setups grow one tag at a time, which is why they break. We write the plan first: the decisions you need to make, the metrics that inform them, and the exact data needed. This is an example for a B2B company with a sales team.
| Business question | Metric | Data needed | Source of truth |
|---|---|---|---|
| Which channels bring buyers, not just leads? | Qualified leads and won revenue by first source | UTM and click IDs stored on the lead; CRM stage and deal amount | CRM |
| Is the website converting the traffic we pay for? | Visit-to-lead rate by landing page and channel | Sessions, landing page, form submit event with form ID | GA4 |
| Are ad platforms optimizing toward the right people? | Share of platform conversions that become qualified | Click IDs or hashed emails matched to CRM outcomes | CRM, imported to ad platforms |
| Which content moves buyers toward a demo? | Demo requests from sessions that viewed key content | Content view events with content group; demo event | GA4 and BigQuery |
| Would more spend in a channel add revenue? | Incremental revenue per extra dollar | Weekly spend, revenue and controls over at least two years | Warehouse and model |
| Event name | Fires when | Key parameters |
|---|---|---|
| generate_lead | A lead form is successfully submitted, not when the button is clicked | form_id, form_name, lead_type, page_location |
| book_demo | A meeting is confirmed in the scheduling tool | meeting_type, owner_region |
| view_pricing | The pricing page loads | plan_viewed, billing_period |
| start_trial | A trial account is created | plan, signup_method |
| download_content | A gated asset download completes | content_id, content_type, topic |
The rules behind the table matter more than the names: lowercase with underscores, one event per real user action, parameters instead of new event names for variations, and every event documented with an owner.
Browser-based tracking alone now misses conversions and shortens the memory of returning visitors. These are the changes we design around.
| Change | Effect on your data | How we respond |
|---|---|---|
| Safari caps cookies set by page scripts at seven days | Returning visitors after a week look new; attribution windows shorten | Server-set first-party cookies through server-side tagging on your own subdomain |
| Consent choices and Global Privacy Control signals | Some visitors cannot be tracked with cookies or shared with ad platforms | Consent-aware tags, consent mode, and modeled conversions where the platform supports it |
| Ad blockers and privacy browser extensions | Client-side tags never fire for part of your audience | Server-side collection and conversion APIs as a second path |
| Google retired most Privacy Sandbox APIs in 2025 and kept third-party cookies under user choice in Chrome | Plans built on those APIs, such as the Attribution Reporting API, have no future | First-party data and platform conversion APIs, not browser APIs that may disappear |
| Long B2B sales cycles | Deals close after platform attribution windows end | Click IDs and hashed emails stored in the CRM; outcomes imported weeks later |
| Several people and devices in one purchase | One deal looks like many unrelated visitors | Account-level reporting in the CRM and warehouse, not only user-level reports |
A durable setup uses more than one path, with deduplication so each conversion is counted once.
| Method | How it works | Strengths | Limits | Best used for |
|---|---|---|---|---|
| Browser tags | Tags in the page send events from the visitor's browser | Quick to deploy; rich page context | Blocked or shortened by browsers and extensions | Behavior and on-site events |
| Server-side tagging | A tagging server on your own domain receives events and forwards them (Google's overview) | More control over what is shared; first-party context | Hosting cost and setup effort | Core analytics and ad conversions |
| Platform conversion APIs | Your server sends events directly to platforms, such as Meta's Conversions API | Works when browser tags fail; can include CRM data | Needs careful deduplication with browser events | Leads and purchases for ad optimization |
| CRM and offline imports | Outcomes such as qualified leads or won deals are uploaded later, for example with Google's enhanced conversions for leads | Teaches bidding which leads become revenue | Delayed; needs click IDs or hashed contact data stored at capture | B2B lead quality and revenue |
For B2B advertisers the last row matters most. If ad platforms only see form fills, they find more people who fill forms. If they see which leads became customers, they find more customers. Campaign management itself runs through our Performance Marketing team; we make sure the signals it relies on are accurate, including for Google Ads and Meta Ads.
Many GA4 problems are not visible in reports. They sit in settings that were left at defaults or set once and forgotten.
| Setting | What goes wrong | What we do |
|---|---|---|
| Data retention | Explorations lose history fast if retention is left short. Standard properties allow 2 or 14 months (Google) | Set the maximum your plan allows and export raw data to BigQuery for longer history |
| BigQuery export | Not switched on, so raw data is never kept. Standard properties have a daily export limit of 1 million events (Google) | Enable export early, size it for your volume and model the tables for reporting |
| Key events | Too many events marked as key events, or the wrong trigger, inflating conversions | Only events tied to the measurement plan are key events, each with a documented trigger |
| Cross-domain and referrals | Visits through booking, payment or app subdomains split into new sessions | Cross-domain measurement and unwanted-referral rules for every tool in the journey |
| Internal and test traffic | Staff, agency and QA visits inflate engagement and conversions | Internal traffic rules and a separate test property or stream |
| Custom channel groups | Paid social, partner and email traffic land in the wrong channel | Channel rules written to match your UTM standard |
| Personal data | Email addresses or names leak into page URLs or event parameters | Data layer and URL checks; redaction before data reaches GA4 |
Every measurement method answers a different question and has a blind spot. We compare them side by side instead of crowning one.
| Method | Answers | Strength | Blind spot | Best for |
|---|---|---|---|---|
| Ad platform reporting | What did this platform's ads touch? | Fast and detailed | Each platform claims credit; totals overlap | In-platform optimization |
| GA4 attribution | Which tracked touchpoints led to a conversion? | One view across channels on your site | Misses untracked and offline touches; limited by consent | Channel and campaign comparisons |
| Marketing mix modeling | How much did each channel add, controlling for other factors? | Uses aggregate data; no user tracking needed | Needs long history; weaker for small or new channels | Annual and quarterly budget planning |
| Incrementality experiments | What changed because of this spend? | Closest to cause and effect | Takes time and some spend held back | Testing big bets and calibrating models |
| Self-reported attribution | Where do buyers say they heard of you? | Captures word of mouth, podcasts and communities | Memory is imperfect; open text needs coding | Reading what software cannot see |
Open-source tools have made modeling more accessible. Google's Meridian is an open-source marketing mix model that can be calibrated with experiment results, which is the approach we prefer: model the whole budget, then test the assumptions that move the most money.
For B2B pipeline reporting, such as sourced and influenced pipeline, we work to the definitions set on our Revenue Marketing page so both teams read the same numbers.
Most tracking breaks during routine site releases, and most teams find out weeks later. We set checks that find it within a day.
| Check | How often | Alert when |
|---|---|---|
| Key event volume against its recent baseline | Daily | Volume moves well outside its normal range for that weekday |
| Tag firing on critical pages and forms | Every release, plus weekly | A required tag is missing or fires twice |
| UTM compliance on new campaigns | Weekly | Links use values outside the approved list |
| Direct and unassigned traffic share | Weekly | Share jumps without a known cause |
| Match rate for CRM or offline imports | Each upload | Matched or accepted records fall sharply |
| Consent rate by region and device | Weekly | Consent rate changes after a banner or site change |
| Duplicate conversions | Daily | Browser and server events for the same conversion are both counted |
| Personal data in URLs or parameters | Weekly | An email or phone pattern appears in collected data |
One dashboard for everyone serves nobody. We build each view for the people who use it and the decision they make.
| Audience | Cadence | Questions it answers | Format |
|---|---|---|---|
| Leadership | Monthly | Is marketing spend producing growth? Where should next quarter's budget move? | One-page memo with three to five decisions |
| Head of marketing | Weekly | Which programs are ahead or behind plan, and why? | Dashboard with targets and notes |
| Channel owners | Daily | Which campaigns, audiences or pages need changes today? | Channel dashboards with alerts |
| Finance | Quarterly | What did each channel return, and how confident are we? | Model results with ranges, not single numbers |
Every monthly memo ends with decisions, owners and the metric that will show whether each one worked. The next memo opens by reviewing them. Over a year this log becomes the clearest record of what marketing has learned.
Most companies we audit sit at stage one or two. Find yours, and the next step follows from it.
| Stage | What it looks like | Next step |
|---|---|---|
| 1. Platform reports | Each ad platform and GA4 read separately; numbers rarely match | Audit tracking and write a measurement plan |
| 2. Tracked and tagged | Key events defined and tags maintained; no CRM outcomes in ad platforms | Add server-side collection and offline or CRM imports |
| 3. Joined and governed | Spend, GA4 and CRM data joined in a warehouse; checks and definitions in place | Compare attribution methods and run the first incrementality test |
| 4. Modeled and tested | Budget planned with a calibrated model and a standing test calendar | Automate scenario planning and expand tests to new channels |
| Weeks | Focus | What happens |
|---|---|---|
| 1 to 2 | Audit | Tags, events, conversion actions, UTMs, consent setup and data flows reviewed. GA4, ad platform and CRM numbers reconciled for one month. |
| 3 to 4 | Plan | Measurement plan, event naming standard and data ownership agreed with marketing, sales operations and IT. |
| 5 to 6 | Fix and build | Tracking repaired, server-side collection and conversion APIs set up where justified, GA4 settings corrected, BigQuery export live. |
| 7 to 8 | Validate and report | Numbers reconciled again against the baseline, monitoring switched on, first dashboards and the first monthly memo delivered. |
Modeling and experiments start after this, once the data has been clean for long enough to trust.
Tracking decisions are now privacy decisions. We build measurement so that it keeps working when people say no.
| Area | What we do |
|---|---|
| Consent signals | Tags read the visitor's consent state before they fire. Where Google tags are used, we implement consent mode so tags adjust to each choice and Google can model some of the gaps. |
| California visitors | The California Attorney General states that a user-enabled Global Privacy Control must be honored as a valid request to opt out of the sale or sharing of personal information. We configure tags and ad platform data sharing to respect it. |
| Data minimization | Only the parameters in the measurement plan are collected. Names, emails and other personal data never go into analytics tools. |
| Hashing and contracts | Contact data sent to ad platforms is hashed first and sent only where consent and your agreements with each platform allow it. |
| Tag inventory | Every third-party tag listed with its owner, purpose and the data it receives. Unknown tags are removed. |
| Sensitive industries | Healthcare, financial and other regulated businesses get a tag-by-tag review with counsel before launch. |
This is how we build measurement, not legal advice. Your counsel should confirm which federal and state rules apply to your business.
| Role | The question they bring | What changes |
|---|---|---|
| CEO or founder | Which marketing actually grows the business? | A monthly memo with decisions, ranges and the evidence behind them |
| CMO or head of marketing | Why does every meeting start with arguing about the numbers? | One reconciled set of numbers and agreed definitions |
| Paid media lead | Why does bidding chase leads that never buy? | Qualified-lead and revenue signals flowing back into the ad platforms |
| Web or product team | Did our release break tracking? | Release checks and alerts within a day |
| Finance | What is the return on each channel, really? | Model-based estimates with confidence ranges, tested by experiments |
| A good fit when | Not yet a fit when |
|---|---|
| You spend meaningfully across two or more paid channels | You run almost no paid or tracked marketing yet |
| You have a CRM or order system that records outcomes | Sales outcomes are not recorded anywhere |
| Leaders want to change budgets based on evidence | Budgets are fixed regardless of what the data says |
| Someone internally can own the definitions after launch | Nobody will act on the monthly findings |
** SECTION 18 WORKING WITH US**
| Model | Suited to | What is included |
|---|---|---|
| Tracking audit | Teams that suspect their numbers are wrong | Audit, one-month reconciliation and a ranked fix list |
| Measurement build | Teams ready to fix tracking and reporting | Plan, tracking fixes, server-side setup, GA4 configuration, warehouse feed and dashboards |
| Ongoing intelligence | Teams that want analysis every month | Monitoring, monthly memo, modeling and an experiment calendar |
Fees are a flat project or monthly amount based on scope, not a percentage of ad spend. Cloud hosting for server-side tagging and the warehouse is billed by the provider to your account. You receive a fixed quote after the audit.
| Measurement plan | A document linking business questions to metrics, events and data sources. |
|---|---|
| Data layer | A structured object on a website that passes information to tags in a consistent format. |
| Key event | A GA4 event marked as important to the business, such as a lead or purchase. |
| UTM parameters | Tags added to links that record the source, medium and campaign of a visit. |
| Server-side tagging | Collecting events on a server you control and forwarding them to tools, instead of sending them straight from the browser. |
| Conversion API | A direct server-to-platform connection for sending conversion events to an ad platform. |
| Enhanced conversions | Google Ads features that use hashed, consented first-party data to match conversions to ad interactions. |
| Offline conversion import | Uploading outcomes that happen after the click, such as a qualified lead or a signed deal, to an ad platform. |
| Consent mode | A Google feature that adjusts how its tags behave based on each visitor's consent choices. |
| Global Privacy Control | A browser signal that tells sites the user opts out of the sale or sharing of their personal information. |
| Marketing mix modeling | A statistical method that estimates the effect of each channel from aggregate data over time. |
| Incrementality | The extra result caused by marketing, beyond what would have happened without it. |
| Holdout group | A group deliberately not exposed to a campaign, used to measure incrementality. |
Marketing intelligence is the practice of collecting accurate marketing data, connecting it across platforms and turning it into decisions about where to spend, what to change and what to stop. It covers accurate tracking, organized data in tools such as GA4 and a warehouse, and analysis that holds up when platforms disagree.
No. Competitive intelligence studies rivals and markets. Marketing intelligence, as we use the term, measures your own marketing: what worked, what did not and where to spend next.
They count differently. Each ad platform uses its own attribution window and credits itself, while GA4 applies one model across channels and only sees what its tags record. Some gap is normal; a large or changing gap usually means a tracking problem.
Not always. It helps when browsers and blockers cut browser-based tracking, when you want more control over what reaches third parties, or when you send many conversions to ad platforms. It adds hosting cost and setup work, so we recommend it only when the audit shows the gain.
Store click IDs and consented contact data when a lead is captured, then upload the outcome when the lead qualifies or buys, through offline imports, enhanced conversions or conversion APIs. Our Conversion Tracking page explains the setup.
Set retention to the longest your plan allows and export raw data to BigQuery. Retention limits explorations, not standard aggregated reports, but you lose detailed history without an export.
A statistical model that estimates each channel's effect from aggregate spend and outcome data over time. It suits companies with consistent spend across several channels and a couple of years of history. Smaller programs usually get more from experiments first.
By comparing a group exposed to marketing with a similar group that is not, for example by region or audience holdout, and measuring the difference in outcomes. The result calibrates attribution and modeling.
Tags respect each visitor's consent state, we implement consent mode for Google tags, honor Global Privacy Control signals for California visitors, keep personal data out of analytics tools, and hash any contact data sent to ad platforms.
Tracking fixes usually take effect within weeks. Most teams have a reconciled baseline and monitoring by the end of the first two months. Models need more time because they learn from clean history.
No. We fill the gaps your team cannot cover, such as tracking engineering or modeling, and document everything so your team can run it.
Read access to GA4, your tag manager, ad accounts and CRM reports, plus a one-hour session with whoever owns the website releases.
Yes. Calls are held between 8 a.m. and 12 p.m. US Eastern time, Monday to Friday, and contracts are invoiced in US dollars. Our team is headquartered in India.
Tell us what you're working toward. We'll help identify the best first step, the systems involved, and the outcomes worth measuring.