The standard advice on cookieless attribution follows a predictable script; starts with the death of third-party cookies, lists a stack of replacement technologies, and closes with an implementation checklist. We've been in the industry long before AI and generative LLMs were mainstay, and we can safely say that while the technologies change (server-side tracking, universal IDs, CDPs, Privacy Sandbox, conversion APIs), the premise never does.
That premise is: the goal of attribution is to trace an individual person from ad click to purchase, and cookies were the tool that made it work. Now that cookies are going away, we need new tools to do the same job.
We see it differently. Not because the tools are wrong (server-side tracking and clean first-party data are table stakes), but because the assumption underneath them is. Per-touchpoint attribution was already broken before cookies died. Rebuilding the same model with different plumbing gives you the same broken measurement in a more expensive wrapper.
At Brighter Click, we run paid media campaigns across Meta, Google, TikTok, YouTube, and Pinterest for brands in fintech, SaaS, ecommerce, healthcare, and beauty. We stopped treating platform-reported ROAS as the north-star metric before the cookieless conversation started. Not because we saw the privacy wave coming, but because the numbers were already lying to us.
This article explains what we use instead, why it works, and what the standard cookieless attribution playbook misses.
The Standard Playbook and What It Misses
Many agencies, analytics platforms, and SaaS vendors will explain that cookieless attribution means swapping in new tracking infrastructure where cookies used to be. The playbook typically looks like this: migrate to first-party data collection, implement server-side tracking, adopt universal IDs or identity graphs, connect conversion APIs like Meta CAPI and Google Enhanced Conversions, and layer it all together into a "cascade" approach.
This is solid operational hygiene. Every paid media team should be doing it. But it answers the wrong question. It asks "how do we keep assigning conversion credit to individual touchpoints?" when the real question is "were those individual touchpoints ever telling us the truth?"
Very few voices in this conversation suggest that aggregate measurement (MER, blended CAC) should be the primary answer. Marketing mix modeling and incrementality testing get an occasional mention. But the idea that per-touchpoint attribution itself is the problem, not just the cookies that powered it, is largely absent from the conversation.
The industry is rebuilding the same house on the same cracked foundation.
Platform Attribution Was Already Broken
Here is what we see when we take over accounts from other agencies or in-house teams: the platform-reported numbers look stable, sometimes even good. But three things are quietly happening underneath.
Frequency creeps up while unique reach flattens. The platform keeps showing ads to the same people instead of finding new ones. Conversions look steady because the warm pool keeps converting. But the pool is shrinking, and the account will not survive an attempt to scale.
New customer share of conversions declines month over month. Most teams never split new versus returning customers inside their blended ROAS. They watch the topline number stay flat and assume everything is fine. They never see that existing customers, people who would have purchased anyway, are propping up the metric. Platform ROAS looks healthy while actual acquisition performance deteriorates.
The CPM-to-CPMr gap widens while CPA stays flat. CPM (cost per thousand impressions) climbs. CPMr (cost per thousand reached, or unique impressions) climbs faster. That gap means the platform is serving more impressions to fewer people. CPA stays flat only because those fewer people are the easiest to convert: your warmest audience. The moment you try to push beyond that pool, CPA spikes and the account looks like it broke overnight. It did not break overnight. It was propped up.
None of these problems are caused by cookies disappearing. They existed when cookies worked fine. They are caused by treating platform-reported ROAS as the source of truth instead of as one input among several.
And here is the part that matters for the cookieless conversation: rebuilding per-touchpoint tracking with server-side tracking and universal IDs does not fix any of these three problems. You get a higher match rate. You capture more conversions. And the platform still flatters you with numbers that do not reflect genuine acquisition health.
MER: The Metric That Does Not Need Cookies
MER stands for Marketing Efficiency Ratio. The formula is simple:
MER = Total Revenue / Total Marketing Spend
That is it. No attribution window. No multi-touch model. No identity resolution. No cookies of any kind.
MER answers one question: for every dollar we spend on marketing, how many dollars of revenue come in? It measures the whole system, not individual channels. And that is exactly why it works in a cookieless world: it never depended on cookies in the first place.
When MER is your north star, you are not asking "which ad drove this sale?" You are asking "is our total marketing investment producing revenue efficiently?" That question is answerable with data you already have: total revenue from your accounting system, total marketing spend from your finance team.
MER works where per-touchpoint attribution fails because a modern purchase journey does not follow a traceable line. A person sees a Meta ad, searches the brand on Google, reads a review, comes back through a retargeting ad on a different device, and converts three weeks later through a direct visit. Attempting to assign fractional credit to each touchpoint was always a polite fiction. Cookies made the fiction feel precise. Losing cookies made it obvious.
MER skips the fiction entirely and measures the outcome.
Blended CAC: MER's Companion
MER tells you how efficiently your marketing produces revenue. Blended CAC tells you how efficiently it produces customers.
Blended CAC = Total Marketing Spend / Total New Customers Acquired
Again, no cookies. No attribution model. Just two numbers your finance team already tracks.
Blended CAC is especially useful for subscription and high-LTV businesses (SaaS companies, fintech platforms, beverage brands with subscription models) where a customer's first purchase is not the point. The question is whether you acquired that customer at a cost that makes economic sense across their full lifetime value.
When we audit client accounts, we look at blended CAC relative to LTV by customer cohort. An account might show strong platform ROAS while blended CAC is actually rising because the platform is counting existing customers as new conversions and the real acquisition cost is climbing.
The Split Nobody Makes
This is the single most common gap we see in paid media accounts: teams never separate new customer performance from returning customer performance inside their platform reporting.
A media buyer watches blended ROAS sit at 4x and reports that the account is healthy. But if you break that 4x apart, returning customers might be converting at 8x while new customer ROAS is at 1.5x. The returning customers would have purchased anyway: they already know the brand, and they are already on the email list. The platform is claiming credit for revenue it did not generate.
This is not a cookie problem. This is a measurement discipline problem. And it is not solved by switching from browser cookies to server-side tracking or universal IDs. Those technologies improve the match rate for per-touchpoint tracking. They do not force anyone to separate new from returning.
At Brighter Click, we split new versus returning in every account we manage. It is part of the audit we run before we touch a campaign. If the previous team never made this split, the ROAS they reported was meaningless. Any decisions based on it were built on noise.
The Halo Effect: Why Per-Channel Credit Does Not Work
Here is another reason per-touchpoint attribution fails, cookies or not.
When we scale spend on Meta for an ecommerce or DTC brand, we monitor what happens on Google brand search, on Amazon, and on TikTok. In most cases, increased Meta spend drives measurable lifts across all of these channels. Meta acts as a demand driver. Google brand captures the demand. Amazon and TikTok benefit from the awareness.
This is the halo effect. No per-touchpoint attribution model handles it well. A last-click model credits Google brand for the conversion. A first-click model credits Meta. A multi-touch model splits credit in a ratio that no one can verify. And all three models make the Meta spend look less efficient than it actually is, because they do not account for the revenue it drives on other platforms.
MER handles this automatically. If you increase Meta spend by $20,000 and total revenue across all channels increases by $80,000, MER reflects the real efficiency of that decision. You do not need to solve the credit-assignment problem because you are measuring the system, not the parts.
When Per-Touchpoint Tracking Still Matters
We are not saying tracking is useless. Server-side tracking, first-party data, and clean conversion APIs matter, but as operational signals, not as the north-star metric.
Per-touchpoint data is useful for:
- Creative testing. You need to know which ad variants drive engagement and which do not. Platform data, even imperfect, tells you whether Hook A outperforms Hook B within the same campaign.
- Channel-level directional signals. If you increase TikTok spend by 40% and see no movement in MER, that tells you TikTok is not contributing. You still need channel-level data to make allocation decisions; you just do not use it as the source of truth for total performance.
- Audience diagnostics. Frequency, reach, CPMr, and audience overlap data help you spot the warning signs. These are health indicators for the account, not proof of attribution accuracy.
The difference is hierarchy. MER and blended CAC sit at the top. They are the metrics you hold the team accountable to and report to leadership. Channel-level platform data sits underneath. It informs tactical decisions (budget shifts, creative refreshes, audience adjustments), but it does not define success.
What We Do Differently
At Brighter Click, our measurement approach starts with three layers, in this order:
Layer 1: Business-level metrics (MER and blended CAC). Total revenue divided by total marketing spend. Total spend divided by new customers acquired. These are the metrics that tell you whether marketing is working. They do not require cookies, identity resolution, or attribution models. They require access to your revenue data and your finance records.
Layer 2: New versus returning customer split. Inside every platform account, we separate new customer conversions from returning customer conversions. This single step eliminates the most common source of false confidence in paid media reporting. If you do nothing else from this article, do this.
Layer 3: Diagnostic health indicators. Frequency versus unique reach trends, CPM-to-CPMr gap, new customer share of conversions over time. These signals tell you whether the account is genuinely healthy or whether platform ROAS is being propped up by a shrinking warm pool.
We do use server-side tracking and conversion APIs. We implement Meta CAPI with deduplication and high match quality on every account. But we use these tools to improve the quality of our diagnostic signals, not as a substitute for measuring what matters at the business level.
The cookieless future does not require a new tracking infrastructure. It requires a willingness to measure marketing the way the CFO already measures every other department: total cost in, total result out. MER and blended CAC do that. They always have.
The only thing cookies gave us was the illusion that we did not need them.
