Count every conversion. Exactly once.
Browsers drop cookies, consent banners block events and every platform claims the same sale. We rebuild your measurement so the conversions you optimise against are ones you can actually defend in a meeting.
- Google Analytics 4
- Google Ads
- Meta Ads
- Microsoft Clarity
Events reaching the platform, not the browser bin.
Sample dashboard shown for illustration
Tracking that survives the real internet
A pixel pasted into a theme file was enough in 2018. It isn't now. We rebuild measurement in four layers — plan, infrastructure, platform connections and ongoing QA — so the data holds up when browsers, consent rules or your site change.
Measurement plan
Before any code: which events matter, what parameters they carry, what each one is worth and who uses it. Naming conventions written down so the setup survives your next hire.
Server-side infrastructure
A server-side GTM container and first-party data layer so events are sent from your infrastructure — past ad blockers, short cookie lifetimes and browser restrictions.
Platform connections
Conversions API, enhanced conversions and offline imports wired into each ad platform, deduplicated against each other so one sale isn't counted five times.
QA & governance
Automated tag monitoring, alerts when a conversion goes quiet, and a monthly reconciliation against your CRM or order data — so breakage surfaces in days, not quarters.
A tracking audit: what fires, what double-fires, and what silently stopped months ago.
A documented setup — event spec, container notes and a reconciliation sheet.
Every container, property, pixel and API connection, under your own accounts.
Standard tracking is quietly failing
Nothing throws an error when tracking degrades. Conversions just drift down, cost per acquisition drifts up, and the algorithm optimises against a partial picture. Here's what breaks, and what we put in its place.
- Browser restrictions — Safari and Firefox block third-party cookies by default and cap first-party ones, so returning visitors look like new ones.
- Ad blockers and network filters stop client-side pixels from ever loading, silently removing a share of every audience.
- Consent banners block tags before they fire. With roughly a third of visitors accepting cookies, a setup without Consent Mode simply loses the rest instead of modelling them.
- Offline and delayed conversions — phone calls, quotes, showroom visits and deals that close weeks later never make it back to the platform.
- Double counting — the same purchase logged as a pixel event, a CAPI event and a GA4 import, inflating reported return across every channel.
- Silent breakage — a theme update or a renamed thank-you page kills a conversion action, and nobody notices until the month-end report.
- Server-side event collection through your own subdomain, so events originate from your infrastructure rather than the visitor's browser.
- Hashed first-party matching — email, phone and order identifiers passed securely to raise match rates without exposing customer data.
- Consent Mode v2 and regional rules configured deliberately, so you stay compliant and still recover modelled conversions.
- Offline conversion imports from your CRM, so bidding optimises toward qualified leads and closed revenue, not raw form fills.
- One event ID everywhere — deduplication rules across pixel, API and import so each conversion is counted exactly once.
- Monitoring and monthly reconciliation against source-of-truth order or CRM data, with alerts when a stream goes quiet.
How much of your traffic the platform can actually identify and learn from.
Share of events correctly matched between pixel and server, not counted twice.
Gap between platform-reported conversions and your own record of them.
How long broken tracking runs before an alert reaches a named owner.
One thing worth knowing: Chrome never completed its third-party cookie phase-out — it moved to a user-choice model instead. That's often mistaken for a reprieve. Safari and Firefox still block by default, consent rates keep falling, and Consent Mode v2 has been mandatory for EEA and UK advertisers since March 2024. The signal loss arrived anyway; it just arrived gradually enough that most teams never saw a drop.
What accurate tracking gives you back
Measurement isn't the deliverable — it's what makes every other decision defensible. Here's what changes once the numbers are trustworthy.
A return figure you can defend
Once each sale is counted once and reconciled against your own orders, ROAS stops being three platforms' optimistic guesses added together. You get one number that survives a finance review.
- E-commerce
- Lead generation
- Retail
Revenue and value, not just counts
Passing order value, margin tier or lead quality back to the platform is what makes value-based bidding possible at all. Without it, the algorithm chases the cheapest conversion rather than the most profitable one.
- Value-based bidding
- Shopping
- Performance Max
Attribution across the whole journey
Calls, quotes, showroom visits and deals that close months later traced back to the campaign that started them — so long-cycle channels stop looking like failures on a same-session report.
- B2B
- Automotive
- Home services
Session recordings that explain the gap
Analytics tells you people abandoned checkout. Clarity recordings and heatmaps show you the field that wouldn't submit on mobile. Every recurring friction pattern becomes a documented test.
- Checkout
- Forms
- Landing pages
Traffic you can actually read
Clean channel grouping, working cross-domain tracking and referral exclusions, so payment gateways stop stealing credit and "direct / none" stops being your biggest source.
- GA4
- Channel reporting
- Attribution
Breakage caught in days, not quarters
Automated tag monitoring with alerts to a named owner. The alternative is discovering at month end that three weeks of spend optimised against nothing.
- Monitoring
- Alerting
- Reconciliation
Every platform measures it differently
Each one uses its own identity signals, attribution windows and counting rules. We set them up individually, then reconcile them against a single source of truth.
Google Analytics 4
GA4 is where behaviour, sources and conversions meet. Most properties are running the default install, which measures pageviews well and business outcomes poorly. We rebuild it around the events your business actually reports on.
- Event taxonomy and data layer — a documented set of events and parameters, with the full e-commerce or lead sequence implemented properly rather than auto-collected.
- Key events and custom definitions — conversions marked deliberately, with custom dimensions and metrics for the segments you actually manage against.
- Server-side GTM and Measurement Protocol for back-end events like refunds, cancellations and CRM stage changes.
- Identity and cross-domain — user-ID stitching, cross-domain linking and referral exclusions so payment gateways stop stealing credit.
- Consent Mode v2, data retention and regional settings configured to your legal position, not left on defaults.
- Reporting layer — explorations, funnels and path analysis, plus BigQuery export where you need raw, unsampled data.
- Data layer
- Key events
- BigQuery export
- DebugView QA
Google Ads
Smart bidding optimises toward whatever you tell it to value. Feed it every form fill and it will find you cheap, unqualified ones. We make the conversion signal reflect real business outcomes.
- Conversion action audit — duplicates removed, primary versus secondary set correctly, and counting, windows and attribution chosen per action rather than left default.
- Enhanced conversions for web and for leads — hashed first-party data and click-ID capture stored in your CRM to recover conversions the browser drops.
- Offline conversion import from CRM stages, so bidding learns from qualified leads, quotes and closed deals instead of raw enquiries.
- Value assignment — static values by lead type or dynamic revenue passed back, which is what makes value-based bidding possible at all.
- Call tracking — calls from ads, call extensions and website calls, counted only past a meaningful duration threshold.
- Deduplication and diagnostics — no double counting against GA4 imports, plus regular checks on conversion lag, tag health and unattributed conversions.
- Enhanced conversions
- Offline import
- Value-based bidding
- Call tracking
Meta Ads
A pixel on its own now misses a meaningful share of conversions, and the ones it does catch are often unmatched. We run pixel and Conversions API together, deduplicated, and work the match quality score like a performance metric.
- Dual-send architecture — browser pixel and server-side Conversions API firing the same events with a shared event ID so nothing is counted twice.
- Event Match Quality work — hashed email, phone, name, location, click and browser identifiers and a stable external ID passed with every event.
- Event configuration — standard and custom events with value, currency, content IDs and catalog parameters so dynamic ads and value optimisation actually function.
- Aggregated Event Measurement — domain verification and event priority ordering set to your real funnel, not whatever was configured at launch.
- Offline and CRM conversions uploaded for lead-gen accounts so Meta optimises toward qualified pipeline rather than form submissions.
- Validation — deduplication monitoring, diagnostics review and attribution-window comparisons, backed by holdout tests where budget allows.
- Conversions API
- Event dedupe
- Match quality
- Offline events
- em
- ph
- fbp / fbc
- external_id
- value
- currency
LinkedIn Ads
LinkedIn drives expensive clicks toward deals that close months later. Judging it on same-session conversions makes it look terrible and buries the channel that filled your pipeline. We connect it to what happens after the lead.
- Insight Tag plus Conversions API — server-side event sending so conversions aren't lost to blockers, with deduplication against the tag.
- Conversion rules built properly — URL and event-based definitions, with click and view attribution windows set to your real sales cycle instead of the default.
- Lead Gen Form capture synced straight into your CRM with campaign and creative context attached to every record.
- Offline conversions from CRM — qualified lead, opportunity and closed-won stages pushed back so LinkedIn is measured on pipeline value.
- Value modelling — weighted values by deal size or segment, so a small enquiry and an enterprise opportunity aren't scored identically.
- Account-level reporting — company engagement and matched audience performance reviewed alongside lead volume, not in place of it.
- Insight Tag
- Conversions API
- CRM offline stages
- Pipeline value
Microsoft Clarity
Clarity doesn't count conversions — it explains the ones you didn't get. When GA4 shows a drop-off at checkout, this is where you watch the sessions that failed and find the button nobody could tap.
- Recordings and heatmaps configured across key templates — product, checkout, landing page and form — rather than sitewide noise.
- Custom tags and identifiers passed in, so you can filter sessions by campaign, device, lead status or test variant instead of scrubbing at random.
- Friction signals monitored — rage clicks, dead clicks, excessive scrolling and quick backs, reviewed as a weekly list rather than a curiosity.
- Paired with GA4 through the native integration, so a funnel drop-off in analytics leads directly to the recordings behind it.
- Privacy masking for form fields and personal data, configured before recording starts, not after a complaint.
- Feeding the test roadmap — every recurring friction pattern becomes a documented hypothesis for the CRO backlog.
- Session recordings
- Heatmaps
- Custom tags
- GA4 integration
- Heatmap
- Scroll depth
- Recordings
Four stages, fully documented
Tracking work goes wrong when it lives in one person's head. Everything we build is specified, tested and written down before it's handed over.
Audit
We map every tag, event and conversion action currently running, then compare what the platforms report against your CRM or order data to size the gap.
Measurement plan
A written spec: events, parameters, values, conversion definitions and attribution settings per platform — signed off before any implementation starts.
Build & deploy
Data layer, containers, server-side endpoint and platform APIs implemented in a staged rollout, tested in preview and validated in each platform's own diagnostics.
Monitor
Automated tag monitoring, alerts to a named owner and a monthly reconciliation against source-of-truth data, so drift is caught while it's still small.
Two to five weeks from audit to validated deployment, depending on stack complexity.
Your developers or ours — we write the spec either way and QA the result.
Containers, properties, API connections and full documentation, in your own accounts.
Conversion tracking, answered plainly
The questions we get most often on a first call. If yours isn't here, ask us directly — no discovery gauntlet first.
Ask us directly →Server-side tracking sends conversion events from your own server through a subdomain you control, instead of from the visitor's browser. It matters because browser-based pixels are increasingly blocked or shortened — Safari and Firefox restrict third-party cookies by default and cap first-party cookie lifetimes, and ad blockers stop pixels loading at all. You need it if a meaningful share of your traffic comes from those browsers, if you run paid media at volume, or if conversions happen offline. A small site running one channel usually doesn't.
Because they each count differently, and none of them are wrong on their own terms. Google Ads credits the click that started the journey and reports the conversion back on that click's date. Meta uses its own attribution windows and view-through credit. GA4 attributes to the last non-direct session. Add the same purchase arriving as a pixel event, a Conversions API event and a GA4 import, and you get a total larger than your bank account. The fix is deduplication with a shared event ID plus a monthly reconciliation against your order or CRM data — not picking whichever platform you like best.
Yes — Google abandoned forced deprecation in July 2024 and moved to a user-choice model instead. That's widely read as a reprieve, and it isn't one. Safari and Firefox have blocked third-party cookies by default for years, which already covers a large share of traffic. Consent rates are low, ad blockers are common, and first-party cookie lifetimes are still capped on Safari. The signal loss happened gradually rather than on a deadline, which is exactly why most teams never noticed a drop in their reports.
It has been mandatory since March 2024 for advertisers serving users in the European Economic Area and the UK — without it, Google restricts remarketing and personalised advertising features. If you only serve Canada or the US it isn't legally required by Google, but advanced Consent Mode still recovers modelled conversions from users who decline, and US state privacy laws are moving in the same direction. Worth noting: advanced mode's conversion modelling needs a volume threshold before it kicks in, so smaller sites won't see the same recovery.
Offline conversion import. The click ID is captured when the visitor lands, stored against their record in your CRM, and pushed back to the platform when the lead reaches a stage that matters — qualified, quoted, closed-won. Bidding then optimises toward outcomes rather than form fills. This is the single highest-impact change for automotive, B2B, home services and anything with a sales cycle longer than a session, and it's the reason those channels often look broken on default reporting.
Two to five weeks from audit to validated deployment, depending on stack complexity. Campaigns keep running throughout — we deploy in a staged rollout with the old setup live until the new events are validated in each platform's own diagnostics. Expect reported conversion numbers to change once it's live, usually downward at first as duplicates are removed. That drop is the correction, not a loss, and we document the before-and-after so nobody mistakes one for the other.
GA4 tells you what happened; Clarity shows you why. When analytics reports a drop-off between checkout steps, Clarity's session recordings and heatmaps let you watch the sessions that failed — the field that wouldn't submit on mobile, the button below the fold, the rage clicks on something that isn't a link. It's free, it integrates natively with GA4, and it turns a number you can't act on into a specific fix. We configure it on key templates with privacy masking rather than recording everything.
Ready to measure what actually happened?
Partner with Digital Jockey and find out what your tracking is missing. We'll audit what's running now, reconcile it against your own numbers and show you the gaps — before you commit to anything.
- Free tracking audit
- Documented handover
- Your accounts stay yours