GA4 attribution: data-driven or last click for channel comparisons?

Most setups I've read about end up with two numbers that never match. GA4 defaults to data-driven attribution, so paid search gets credit for assists that last click would hand to a branded search or a direct visit a few days later. Google Ads reports its own version, so the same conversion can look like a winner in one tool and a loser in the other.

My view is to pick one model per decision and stop arguing about the other. For bidding and budget calls in Google Ads, use the model that matches how Smart Bidding learns. For comparing Meta, email and search side by side, last click is harder to argue with, because every platform can reproduce it from its own data.

How do you reconcile the gap when a client asks why two dashboards disagree?

Agreed on splitting the decision: bidding runs on the model Smart Bidding learns from, and cross-channel comparison runs on last click. On the "why do the dashboards disagree" question, I'd look at something less glamorous than the model first, which is how the Google Ads conversion action counts, "every" or "one" per click. A lead form that fires twice, or a thank-you page that reloads, shows up as two conversions in Google Ads and one in GA4. That alone can look like an attribution problem for weeks. Once the counts match, the gap between models is usually much smaller than it looks.

Agreed on the count issue, that one bites people constantly. Date basis is the other thing to check, because it creates a gap even when the counts are clean. Google Ads files a conversion under the day it happened, while the click that started it may be a week older. GA4 can attribute the same sale to the session date. So a dental clinic pulling "last 14 days" on Monday and again on Friday can get different totals from the same tool, and the most recent days always look weaker in Google Ads because conversions are still arriving. If one side counts view-through and the other doesn't, the numbers drift even further.

The fix is boring: pick one date basis, pull both tools on the same closed date range, and only compare periods that are a few days old. Then argue about the model.

GA4 dates a conversion by the day the event fired, not the session date. The session only decides which channel gets the credit. So a sale on Friday from a click on Sunday lands on Friday in both tools, but GA4 still credits the Sunday source.

The bigger source of mismatch is view-through, which Google Ads can count and GA4 doesn't. If you're comparing the two for a given week, check whether view-through is included in the Google Ads conversion action before you blame the attribution model.

Same goes for lookback windows. Google Ads sets a click-through conversion window on each conversion action, and GA4 has its own lookback under Attribution settings. If the two differ, a Sunday click that converts on day 40 can count in one tool and vanish from the other, even with clean counts and matching dates. So before arguing about the attribution model, pull both windows for the conversion in question and see whether the mismatch is just the lookback.