Paid Media Attribution in 2025: What’s Broken, What Works, and What to Do Now

If you’re feeling lost when it comes to paid media attribution, you’re not alone. With third-party cookies crumbling, platform data becoming increasingly siloed, and more users opting out of tracking, it’s harder than ever to connect the dots across the customer journey.

In this blog, we’ll break down:

• What attribution models are no longer fit for purpose
• What’s actually working in 2025
• How to build a modern attribution stack that doesn’t lie to you
• What attribution really looks like across key paid channels

Attribution Models That No Longer Work (Or Never Did)

If you’ve been on the marketing scene for a while you might be familiar with the more common attribution models, such as last and first click. Despite their continued appearance in dashboards, these models are outdated, inaccurate, and sometimes just plain broken:

• Last click: Still clinging on in many reports but can be dangerously misleading in multi-channel journeys.
• First click: Slightly better for demand generation, still gives an incomplete picture.
• Position-based / time-decay: Theoretically better, but rarely implemented well.
• Platform attribution: Google claims credit, Meta claims credit and the truth disappears in the crossfire

Why Paid Media Attribution Is So Hard Now

Attribution has never been simple, but several forces are making it even tougher:

• The loss of third-party cookies: Over the last few years, we’ve seen multiple browsers lose tracking of third-party cookies. Safari and Firefox have moved on, Chrome is dragging its feet, but still looks like it’s on the horizon.
• Walled gardens: Google, Meta, TikTok, and LinkedIn are all very cagey about sharing data, but all of them want to say they’re the main reason behind conversions coming through. The result Is on-platform tracking often over exaggerates performance from that platform (for obvious reasons).
• Cross-device chaos: The user journey is increasingly complex. For example, one user could use five devices. This means difficulty with tracking and no consistent trail.
• Dark social & direct traffic: There are also a number of other aspects of the user journey which are difficult, if not impossible to properly track. For example, word-of-mouth, WhatsApp messages, LinkedIn DMs are all invisible but influential
• Offline conversions: These are often under-tracked, especially in B2B and high-ticket purchases.

What Does Work in 2025 (Even If It’s Not Perfect)

So we’ve gone through a lot of the issues with attribution models at the moment, but lets talk about the things that can help you get closer to the truth:

• First-party data & UTM hygiene: Not the most exciting of points, but ensuring you have a consistent naming convention with campaigns, that you’re using UTMs consistently, and incorporating first party data should be the first point of call if you want to get things joined up.
• Conversion APIs (Meta, Google, TikTok): Every platform has their own version of server-side tracking. Getting is set up means less reliance on Cookie tracking, and greater accuracy of results.
• Self-reported attribution: Running a post-sale customer survey, or building in a part of your conversion path to ask people questions like, “How did you hear about us?” sounds simple, but can give a clue into their journey to convert.
• Multi-touch models (GA4, Looker): Use UTMs and conversion data to build a multi-touch model that assigns value to different stages of the funnel.
  • Ensure it is bespoke to your business, and look at drop off of performance metrics (such as clicks/sessions on the platform vs. GA to work out volume of lost data)
  • This can help you to build an attribution model that better reflects your user journey and assigns value to engagement across you’re your marketing channels.

How to Build an Attribution Stack That Doesn’t Lie to You

To help you put together an attribution process that better reflects your customer’s journey, and gives some much-needed clarity on where conversions are coming from, follow these tips:

1. Clean up your UTMs
   Work on a UTM structure and make sure it’s consistent across all your platforms. Sloppy tagging leads to messy data.
2. Connect your data sources
   Link your CRM, ad platforms, analytics tools, and offline data. The more data and the better data you can connect up with your platforms, the better your campaigns will perform and the more accurate your results.
3. Implement server-side tagging
   In order to bypass the slow death of Cookie tracking, use Google Tag via GTM Server or similar for better control.
4. Add qualitative data
   Add an extra step into your customer journey where you can, for example, layer in surveys, interviews, and sales feedback. This won’t be perfect, but it will give you a good finger in the air assessment of where leads are coming from.
5. Build dashboards that reflect reality
   It can be tempting to pull all your results by platform with the highest attribution possible (because it makes the numbers look great), but realism should be the priority. Move beyond vanity metrics and focus on useful, honest reporting.

Attribution Red Flags to Watch Out For

To finish up, here are a few red flags to watch out for when it comes to viewing your campaign performance:

🚩 ROAS over 15? Probably a lie (not always, but if something looks too good to be true, it probably is).
🚩 Spike in direct traffic after a campaign? That’s assisted, not organic.
🚩 Platform credit doesn’t match CRM? Trust the CRM.
🚩 “Nothing is working”? You’re likely under-reporting, not under-performing.

Final Thoughts: Get Comfortable With Uncertainty

Attribution in 2025 is murky, fragmented, and far from perfect. And that’s okay.

Your job isn’t to chase perfect tracking. It’s to blend data, context, and instinct to make smart decisions. Prioritise profitable growth, not pixel-perfect precision.

Need help building an attribution model you can actually trust?
Get in touch with us — we’ll help you cut through the noise and build a stack that reflects reality.