Customer Impact
Tryve
← All results
Tryve Tracking & conversion measurement

Conversion numbers Tryve can trust again

Spam, internal test traffic and mislabelled ad clicks were counted in the conversion numbers. We rebuilt the measurement from GA4 to Google Ads, so every channel gets its own leads back and Smart Bidding learns from real enquiries only.

Result Tracking reliable again
Client
Tryve
Sector
B2B work management
Role
Tracking, GA4 & Google Ads
Result
Reliable conversion data
The problem

The conversion numbers included noise: bots, internal tests and ad clicks credited to the wrong channel. Budget decisions rested on polluted data.

What we did

An audit of GA4 and Google Ads, then a clean-up at the source: filters, conversions on a real click, a redirect that keeps the gclid, one UTM standard and a tight set of primary conversions.

The result

Every conversion in the reports is a real enquiry, credited to the right channel. Google Ads now learns from leads only.

The challenge

If you allocate budget on polluted numbers, you allocate it wrong

Tryve steers its marketing on conversions: forms, phone clicks and email clicks. When we lined up every GA4 conversion against its source and medium, we found four kinds of noise: referral spam and bots from search partner domains, conversions from our own Tag Assistant tests, Meta ad clicks that ended up in Organic Social or Unassigned, and a group with no channel at all.

Two causes sat in the measurement itself. The phone and email conversions were set to fire on a page view instead of a click, so a crawler opening the contact page counted as a lead. And the redirect from www to the main version of the homepage dropped the query string: the Google Ads click ID (gclid) and the UTM parameters disappeared before the page loaded. On top of that, old conversion actions in Google Ads were running alongside their successors.

What we built

The tracking architecture, step by step

01

An audit of every conversion

We lined up the full GA4 history against source, medium and conversion action, and Google Ads against GA4. That showed which conversions were noise and where the measurement leaked. Our method is in the conversion tracking audit checklist.

02

Stopping noise at the source

Internal traffic defined by IP and the GA4 data filter set to Active, search partner and spam domains added to the list of unwanted referrals, and a reusable exclusion segment that also cleans the history in reports, without deleting any data.

03

Conversions on a real click

Phone and email conversions now only fire on a click on a tel: or mailto: link, through a click trigger in Google Tag Manager. A bot loading a page no longer creates a conversion.

04

The gclid survives the redirect

Together with the web builder, one canonical redirect rule for the homepage and every path: permanent (301) and with the query string included. The Final URLs in Google Ads point to the canonical version, so ad clicks do not even need the redirect.

05

One UTM standard for Meta

All Meta ads get the same URL parameters with utm_medium=paid_social, set at account and campaign level. A custom channel group also moves the historical Meta conversions under Paid Social.

06

Only real leads as primary conversions

Form, phone click and email click are the primary conversions in Google Ads. Old duplicate actions are archived, soft signals are secondary, and the outdated key events in GA4 are unmarked so nothing counts twice.

Why this way

The choices behind the setup

Every choice follows the official documentation. Google Ads explicitly asks that, when a site uses redirects, the gclid is passed to the final landing page. Without the click ID a conversion cannot be tied to the ad and Smart Bidding loses signal. That is why we fixed the problem at server level, not in the tag.

In GA4, a data filter for internal traffic does nothing while it is in the Testing state: only when it is Active does Analytics apply the filter to incoming data. Because that exclusion is permanent, we paired the filter with an exclusion segment for the history: nothing is deleted, the reports are clean.

For channel attribution, the UTM medium is decisive. Google recommends one unique utm_medium per channel, such as paid_social. A home-made value is not recognised as paid and fragments the data.

And in Google Ads only primary conversion actions are used for bidding, secondary actions are for observation only. Leave soft signals or duplicate actions as primary and you teach the algorithm to optimise for noise.

What changed

From measuring on hope to measuring by design

Phone and email conversion Before: fires on a page view After: fires on a real click only
gclid on an ad click to the homepage Before: lost in the redirect After: preserved, 301 with query string
Meta ad conversions Before: scattered across Organic Social and Unassigned After: under Paid Social, history included
Internal test traffic and spam Before: counted as conversions After: excluded through filters and a segment
Primary conversions in Google Ads Before: old and new actions mixed After: form, phone and email only
The result

Numbers you can steer on again

Real leads

every conversion in the reports is a real enquiry

Right channel

Google Ads, Meta and organic each get their own leads

Clean signal

Smart Bidding learns from form, phone and email only

The difference shows in the decisions. Anyone shifting budget between Google, Meta and organic now looks at leads instead of noise. And because the causes were fixed at the source, the data stays clean without monthly corrections. Curious which errors most often creep into GA4? Read the tracking errors that quietly make your GA4 data useless.

Frequently asked questions

What marketers ask us about this setup

How do you know your conversion numbers are polluted? +

Line up your conversions against source and medium. Referrals from domains unrelated to your business, conversions from Tag Assistant and a large group with no channel are the classic signs.

Why do bots sometimes count as leads? +

Because the conversion fires at the wrong moment. If a phone or email conversion already counts when the page loads, every crawler becomes a lead. Only let it fire on a click on the link.

Can a redirect break your Google Ads measurement? +

Yes. If a redirect does not carry the query string, the gclid disappears before the landing page loads. The click then counts as direct or organic traffic and the conversion is not tied to the ad.

Why did GA4 not see the Meta ads as paid? +

GA4 reads the UTM medium. A home-made value is not recognised as paid, paid_social is. One fixed URL template in Meta and a custom channel group for the history solve it.

Should you throw away old, polluted data? +

No. An exclusion segment cleans the history in your reports without deleting anything. From then on, filters stop the noise at the source.

Further reading

Get started on your own tracking

Read more

Tryve: the lead generation system behind the growth

→

Do you doubt your own conversion numbers?

We audit your GA4, Google Tag Manager and Google Ads, show where the measurement leaks and how to fix it at the source. Contact Tanguy.