Cross-channel attribution: one answer for Google, Meta and LinkedIn

Add up what Meta, LinkedIn and Google Ads each report as conversions from the same campaign, and the total commonly runs two to three times higher than your actual revenue, the same double-counting Kleene.ai's 2026 analysis found across Meta and Google campaigns. Each platform is grading its own homework: Meta counts every click and view inside its own attribution window, LinkedIn counts a different set of touchpoints, Google assumes it was the first thing the customer saw, and none of them reconcile against what actually closed. Where should the next dollar go when three honest-looking numbers cannot agree with each other or with your CRM?

This is the Monday morning of every B2B marketing director running ads on three platforms. You open three tabs, export three reports, paste them into a spreadsheet, and spend an hour reconciling what each platform calls a conversion. By the time you have an answer, the question has moved. Your budget is already overspent on the channel that over-counts.

Cross-channel attribution is one consistent method for deciding where the next dollar goes, applied across every platform the same way, every week. No method solves cross-device and dark-social tracking completely, but a consistent method beats reconciling three platforms by hand every Monday. We wrote separately about the AI-powered version of this attribution work if you want the automated route.

So what does the Monday routine look like after it is wired up?

One dashboard. One set of rules applied to every platform. One number that tells you which channel drove pipeline last week, driving the budget shift instead of whichever platform's native dashboard happens to be the most optimistic that week. The team spends less time arguing about data and more time deciding what to do next.

Attribution here means applying one consistent methodology so week-over-week comparisons actually mean something. When a platform reports more conversions than your CRM shows, the fix is a system that resolves the difference and tells you which number to trust.

Here is what the framework looks like in practice.

How the cross-channel attribution framework works

The ingest stage

Every platform publishes data through APIs. Google Ads has the Google Ads API. Meta has the Marketing API. LinkedIn has the Marketing Analytics API. Each of them will give you campaign-level spend, impressions, clicks, and conversions. The trick is not getting the data. The trick is pulling it into one place with the same date range, the same attribution window, and the same conversion definitions.

Most platforms let you set a custom attribution window. Pick one window, seven-day click or one-day view is standard for B2B, and apply it across every platform. Google, Meta, and LinkedIn all support this. The numbers will still differ because each platform uses a different model to assign credit, but at least the time frame matches.

The reconcile stage

This is where the real work lives. When Google says 120 conversions and your CRM shows 45, neither number is wrong. They are measuring different things. Google counts every ad click that led to a website visit within the attribution window. Your CRM counts only the leads that submitted a form or booked a call. We covered the full attribution pipeline setup in an earlier post. The gap is signal loss. People who clicked, visited, and left without converting never reach the CRM.

The reconciliation method is straightforward. Take your CRM conversions as the ground truth. Those are real people who took a real action. Then map each CRM conversion back to the ad platform that drove it using UTM parameters or a click ID. The platform that gets the match gets the credit. The conversions that do not match anything are upper-funnel assisted touches, useful context for understanding reach. They should not move the budget number on their own.

The cost of leaving this gap open is measurable. Digital Applied's 2026 analysis puts the median dark-funnel gap, pipeline that arrives with no attributable touchpoint at all, at 38 percent. That is budget decisions being made on the majority of pipeline you can see, while more than a third is invisible to whichever attribution method you are running.

The attribute stage

Once every CRM conversion has a platform source, you can answer the real question: which platform drives the most pipeline per dollar spent. That per-dollar figure is what should move your budget, a different number from cost per click or cost per impression.

Meta's 2026 measurement changes make this stage more important than ever. On 12 January 2026, Meta deprecated its 7-day and 28-day view-through attribution windows. DOJO AI's analysis found reported conversions dropped 15 to 30 percent overnight for advertisers who had relied on those windows, purely because the measurement changed, with nothing different in the campaigns themselves. If you were comparing January to December without knowing this, you would have cut budget from a channel that was performing exactly as well as before.

A practical application of this: if Meta delivers a cost per lead of AUD 45 and LinkedIn delivers AUD 120, the obvious move is to shift budget to Meta. But if Meta's leads convert to pipeline at half the rate of LinkedIn's, the cost per dollar of pipeline might be the same. The framework catches this. It tracks the full path from spend to lead to opportunity, not just the cost per click.

The learn stage

An AI attribution system does not just report what happened. It ingests data from every platform simultaneously, applies a consistent attribution model, uses probabilistic matching to fill signal-loss gaps, and learns channel weighting over time. The first week of data gives you a baseline. Week four shows you which channel mix performs best. By week eight, the system is adjusting recommendations based on actual pipeline outcomes, not platform-reported vanity metrics. Every week of consistent data makes the next week's recommendation better than a spreadsheet rebuilt from scratch could manage, and that compounding effect is what separates a system from a report.

A manual spreadsheet tells you what happened last month. An AI system tells you what is happening today and what to shift tomorrow, faster than a person reconciling three tabs by hand ever could.

Three things you already have, none of them a new platform

You do not need a new platform or a data team to start. You need three things that already exist in your business.

One: pick an attribution window and apply it everywhere. Seven-day click, one-day view is the B2B standard. Set it in Google Ads, Meta Ads Manager, and LinkedIn Campaign Manager. Write it down so your team knows the rule. Every platform gets the same window.

Two: standardise your UTM parameters. Every campaign, every ad set, every ad needs consistent UTM tags. Source should be the platform (google, meta, linkedin). Medium should be cpc. Campaign name should match across platforms when they promote the same offer. This is tedious to set up but impossible to fix retroactively. Do it once.

Three: calculate what inconsistency is costing you. Take your total monthly ad spend across all three platforms. Multiply it by the percentage of leads you cannot confidently attribute to a source. Gartner's 2026 CMO Spend Survey found budgets sitting at 7.8 percent of company revenue, and most B2B teams estimate 20 to 30 percent of leads fall into the unattributable bucket. That is the budget going to guesswork. If your monthly spend is AUD 50,000 and 25 percent is unallocated, you are losing AUD 12,500 a month to data fragmentation. An attribution system that recovers even half of that pays for itself in the first month.

Frequently asked questions