Data Without Truth: How Ad Platforms Distort Reality

Before starting Kickbite, while I was still validating the idea, I conducted two investigations with more than 100 observations.

First, I studied the main source of data used by CMOs, Google Analytics.

Second, I analyzed the main source of data used by specialists, the data from each ad platform.

The findings were brutal.

Google Analytics’ Last Click and Data-Driven attribution models always favor Search.

Last Click is simple to understand. Search, especially Brand Search, appears throughout the user journey but dominates near the conversion point. Users who already intend to buy are in the decision stage, and Last Click captures this moment.

The Data-Driven model was supposed to distribute credit more fairly across touchpoints. In practice, it did the opposite. It gave even more weight to Google channels than Last Click did.

After reading Google’s documentation carefully, the reason became clear.

One of the key variables in their model includes clicks and video engagements “on your Search (including Shopping), YouTube, Display, and Demand Gen ads in Google Ads.”

These engagement signals apply only to Google’s own channels. Interactions from Meta, TikTok, influencers, or any external sources are ignored.

The result is a dataset that contains complete information for Google’s channels and incomplete information for everyone else. The outcome is biased from the start.

And realistically, would Google ever offer a free analytics tool that challenges its own ad business?

When we analyzed ad platform data, the situation was even worse.

Two main things became clear.

Full Credit for Every Conversion

Meta takes full credit for every conversion it touches. It does not matter if there were ten other touchpoints before or five more significant ones after. Meta claims the entire conversion, ignoring the combined effect of the full journey.

If conversions truly depended only on Meta, users would convert after a single ad. In reality, more than 90 percent of buyers require multiple exposures before making a purchase. When there is a gap between touchpoints, most will forget the first brand interaction altogether.

Weak Signals Count as Success

Meta’s attribution model accepts almost any engagement as proof of success. A like, comment, or simple reaction is enough for the platform to claim credit. Their click model is not built on real intent but on maximizing reported impact.

TikTok, Pinterest, Google Ads, and others use similar methods. Their business model depends on continuous advertiser spending, so showing weak results is not an option.

If these platforms revealed how little impact many campaigns actually have, ad budgets would shrink overnight.

In reality, a significant part of marketing spend goes to:

Platforms know this, and it is one of the reasons they promote complex marketing mix models that hide the weaknesses of their own attribution.

This situation was not acceptable. In 2018, we decided to build our own tracking and attribution system.

Our goal was to create a deterministic model, fully independent and unbiased.

In our system, each touchpoint is verified as a real session connected to a real user and a real conversion. There are no assumptions or inferred links, only validated data.

When customers compare Kickbite’s results with data from Meta, Google Ads, or TikTok, they often notice that Kickbite reports fewer conversions.

This difference is not a problem. It reflects how attribution is calculated.

Ad platforms are built to prove their value. Their attribution models count every possible interaction that could be connected to a conversion, even weak ones. The same conversion can appear multiple times across different platforms.

Kickbite works differently. Our attribution is deterministic and based on verified user sessions. Each conversion is counted once, only when we have complete evidence that a touchpoint directly led to the purchase. This eliminates duplicates and inflated reporting.

Sometimes, Kickbite reports more conversions than ad platforms.

This happens when platform tracking is incomplete or broken.

Attribution accuracy depends entirely on the quality of the tracking system. When tracking fails, attribution fails too.

Since the iOS 14 privacy updates, most browsers delete cookies after 24 or 72 hours. As a result, many touchpoints disappear from standard tracking systems, breaking the conversion chain.

Kickbite was designed to avoid this problem. Our system uses four independent tracking layers to keep attribution accurate even when cookies or scripts fail.

1. Local Storage Tracker

Kickbite stores a local tracker in the user’s browser cache. It remains active even when cookies expire and can only be deleted manually by the user.

2. External IDs and Backend Integration

We combine identifiers from marketing platforms and backend systems.

Platform IDs help with cross-device tracking, since users are often logged into their accounts on multiple devices. Backend identifiers provide the highest quality data for validation.

3. IP and Fingerprint Verification

We use IP addresses and browser fingerprints to confirm user identity and maintain session continuity across touchpoints.

4. Product Behavior Validation

Our system verifies attribution by checking product behavior.

If a user bought products X, Y, and Z, the journey must include relevant touchpoints that make sense for that purchase. This ensures our tracking remains fully deterministic and consistent.

Kickbite reports conversions based on verified user actions, not assumptions.

You may see fewer conversions because we remove duplicates and false positives.

You may see more when our tracking detects conversions that other systems miss.

The result is data you can trust, attribution that is closer to reality, and decisions based on more trust.
​
​