What good is a dashboard full of green arrows if half your conversions never happened and the other half can’t be legally recorded? That’s the trap many teams fall into when they pick a conversion tracking tool based on screenshots and pricing pages. The features that decide whether your numbers survive an audit, a cookie change, or a quiet update from a browser vendor rarely appear in a sales demo.
This guide walks through the features worth comparing, why each one matters, and where teams usually get it wrong.
Why the Feature List Matters More Than the Logo
Most conversion tracking tools can count a button click. The differences show up in harder places: how they handle consent, whether they can collect events server-side, which attribution models they support, and how much raw data you can export before it disappears.
A practical way to think about it: a tool is only as good as its weakest link in the chain from user action to reported number. If any link breaks, every report downstream is wrong. Comparing features is really about finding the weakest links before you sign a contract.
It also helps to treat tracking as part of a wider stack decision. Teams that already compare tools carefully in other areas, such as when they compare AI writing tools before subscribing, tend to apply the same discipline to analytics. The same logic applies here: test the edges, not the highlight reel.
Consent Handling and Privacy Controls
![]()
Under the EU ePrivacy Directive, storing or accessing information on a user’s device generally requires prior consent. The GDPR, in force since 25 May 2018, adds lawful-basis and data-minimisation duties on top. A tool that fires tracking pixels before consent is not a minor configuration issue; it’s a compliance risk.
When comparing tools, ask these questions:
- Does the tool support consent mode so tags adjust behaviour based on the visitor’s choice?
- Can it run cookieless or aggregated measurement for visitors who decline?
- Does it document how consent state is stored and passed to each destination?
- Can you prove, months later, what fired and when?
Google Consent Mode v2, rolled out from late 2023 and widely reported as mandatory for advertisers using Google services in the EEA from March 2024, is a concrete differentiator. A tool that supports it can use modelling to fill gaps left by declined consent. A tool that doesn’t will simply lose those conversions.
One common mistake: teams enable a consent banner but never test what happens after a visitor clicks “reject.” In practice, the tags often still fire. Test the reject path first, not the accept path.
Server-Side Tracking and First-Party Endpoints
Server-side tagging moves event collection off the browser and onto a server you or your vendor control. This matters because ad blockers and third-party cookie restrictions hit browser-side pixels hardest.
Compare tools on whether they offer:
| Feature | Why it matters |
|---|---|
| Server-side container support | Events survive ad blockers and browser limits |
| First-party endpoint | Requests look like your own domain, not a vendor’s |
| Conversions API integrations | Ad platforms receive events reliably |
| Event enrichment on the server | You can add CRM or order data before sending |
The trade-off is complexity. Server-side setups need hosting, maintenance, and someone who understands the data layer. Smaller teams sometimes adopt server-side tracking too early and end up with a half-configured container that nobody monitors. If you can’t staff it, a managed first-party endpoint from your vendor may be the better compromise.
Attribution Models and Reporting Flexibility
Attribution models produce different numbers for the same conversions. Last-click credits the final touch. First-click credits the first. Linear splits credit evenly. Time-decay favours recent touches. Data-driven models use statistical methods to assign credit.
Two tools can report wildly different ROI for the same campaign simply because one defaults to last-click and the other to data-driven. When comparing, check:
- Which models are supported
- Whether the model is configurable per report
- Whether historical data can be re-attributed when you switch models
- How the tool documents its modelling assumptions
A practical tip: pick one model as your source of truth, document why, and stop letting different teams pull reports on different settings. Most “the numbers don’t match” arguments are attribution-model arguments in disguise.
Cross-Domain and Cross-Device Tracking
Tracking a visitor across your marketing site, your checkout subdomain, and your app requires explicit configuration. Tools vary in whether they handle identity resolution natively or rely on your own customer ID.
Look for linker parameters for cross-domain measurement, user-ID support, and a clear explanation of what happens when a visitor switches devices. If your tool relies on an identity graph you don’t control, ask what happens when that graph changes. The answer is often “your conversion numbers move.”
Offline Conversion Imports
Many businesses convert offline: a sales call closes a deal, a customer visits a store, a call centre books an appointment. Offline conversion imports bring that data back into your tracking so campaigns optimise on real outcomes.
Compare tools on supported import formats (CSV, API), match keys (hashed email, click ID, order ID), and refresh cadence. Daily imports are fine for B2B pipelines; weekly imports can starve ad platforms of the signals they need to optimise.
The most common mistake here is a broken match key. If your CRM stores emails in a different normalisation than your tracking tool expects, imports silently fail and you’ll blame the platform. Run a small test batch first and verify the matches land.
Fraud, Bot Filtering, and Data Quality
Invalid traffic inflates conversion counts and wastes budget. Compare tools on whether they offer invalid-traffic detection, bot filtering, and viewability standards. The Media Rating Council publishes invalid-traffic guidelines, and the current edition is worth checking before you rely on a vendor’s claims.
Data quality also covers simpler things: duplicate event suppression, session stitching, and timezone handling. A tool that counts the same purchase twice because the confirmation page refreshed will make every report wrong in a way that’s hard to spot.
Privacy Thresholds and What You Lose at Low Volume
Some tools suppress or aggregate data below minimum thresholds to protect privacy. This is a good thing in principle, but it means small segments become invisible. If you run niche campaigns with a handful of conversions per week, thresholding can wipe out the very data you need.
Compare tools on what they suppress, at what volume, and whether you can see a “data withheld” signal rather than a silent zero. A silent zero is the worst outcome: it looks like a campaign failed when the truth is that the tool chose not to report it.
Data Ownership, Export, and Retention
Ask three blunt questions before signing:
- Can I export raw events, and to where (data warehouse, object storage, CSV)?
- How long is data retained, and what happens at the limit?
- Will the vendor sign a data processing agreement, and as what role?
Tools that lock raw events behind an API rate limit or a paid tier are a long-term risk. If you ever want to switch vendors, migrate to a warehouse, or run your own models, raw-event access is the difference between a smooth move and starting from zero.
This is the same principle behind good tool comparisons elsewhere in your stack. Teams that compare AI research tools before subscribing usually check export options and retention terms first, because those are the terms that are hardest to change later.
Integration Depth and the Rest of Your Stack
Conversion tracking rarely lives alone. It feeds ad platforms, CRMs, email tools, and dashboards. Compare tools on native integrations, webhook support, and how much custom code each connection requires.
If your team already runs a board-based project tool, the analogy is familiar: a tool that integrates cleanly saves hours, while one that needs a middleware layer adds a permanent maintenance task. The same reasoning that leads people to compare board tools and choose one applies to tracking: fewer moving parts, fewer silent failures.
For e-commerce teams specifically, the integration question often comes down to whether the tracking tool talks directly to the storefront and the ad platforms, or whether everything routes through a tag manager. Both work; the tag manager route is more flexible and more fragile.
A Practical Comparison Checklist
Here’s a shortlist to run against any two or three tools:
| Criterion | What to verify |
|---|---|
| Consent support | Consent mode, cookieless measurement, reject-path testing |
| Collection method | Server-side option, first-party endpoint |
| Attribution | Models supported, configurability, re-attribution |
| Identity | Cross-domain linkers, user-ID, cross-device handling |
| Offline imports | Formats, match keys, refresh cadence |
| Data quality | Bot filtering, duplicate suppression, timezone handling |
| Thresholds | What gets suppressed and at what volume |
| Export and retention | Raw-event access, retention limits, DPA |
| Integrations | Native connectors, webhooks, maintenance burden |
Score each tool honestly. A tool that wins on eight criteria but fails on consent or export is not a winner; it’s a liability with good marketing.
Common Mistakes That Skew the Comparison
Three patterns show up again and again.
First, teams compare tools using demo data. Demo dashboards are clean; real data is messy. Ask for a trial with your own traffic.
Second, teams ignore the reject path. Test what fires when consent is declined and when it is granted.
Third, teams assume third-party cookie deprecation is settled. Google announced in 2020 that it would phase out third-party cookies in Chrome, then in July 2024 said it would not deprecate them outright but would offer user choice. Plans have shifted repeatedly. Treat any claim about cookies as dated and check the announcement date before you build a strategy on it.
Conclusion
The right conversion tracking tool isn’t the one with the prettiest dashboard. It’s the one whose weakest link holds under real conditions: consent declined, ad blockers on, imports running daily, and a finance team asking why the numbers moved.
Compare on consent handling, server-side collection, attribution flexibility, offline imports, data quality, thresholding, export rights, and integration depth. Test each with your own traffic. Then pick the tool you can maintain, not the one that looks best in a demo.
Frequently Asked Questions
What features should I compare first in a conversion tracking tool?
Start with consent handling and collection method. If a tool can’t respect consent choices or collect events server-side, nothing else matters much. Attribution models, offline imports, and export rights come next.
Is server-side tracking always better than browser-side tracking?
No. Server-side tracking is more resilient to ad blockers and cookie restrictions, but it needs hosting, maintenance, and monitoring. Small teams without the resources may do better with a managed first-party endpoint.
How do I know if my tracking tool handles consent correctly?
Test the reject path. Decline consent in a real browser, then check which tags fire. If tracking pixels still send data after a rejection, the tool or its configuration is not compliant.
Why do two tools report different conversion numbers for the same campaign?
Usually attribution. Last-click, first-click, linear, time-decay, and data-driven models assign credit differently. Check which model each tool uses and whether the model is configurable before assuming one tool is wrong.
What are offline conversion imports and why do they matter?
They bring offline outcomes, such as closed deals or in-store purchases, back into your tracking so campaigns optimise on real results. Compare tools on import formats, match keys, and how often imports refresh.
How long should I keep conversion data, and can I export it?
Retention depends on your tool and your legal obligations. The key question is whether you can export raw events to your own warehouse. If you can’t, switching vendors later means losing history.


