Marketing attribution beyond GA4
In GA4, three in four orders look like a single click. Ad views, early touches and cross-device paths never make the report, so the channels that create demand look like they don’t work. Kickbite records the whole journey and gives every channel the credit it actually earned.
Sunweb, summer holidays market, February to July. Display ROAS: same campaigns, same spend, Kickbite vs. GA4.
The uncomfortable question
Why measurement breaks
ITP, cookie deprecation and tightening consent remove a little more of the path every year. Deterministic data keeps shrinking, and with it the share of conversions a click can still explain.
Directional. Reflects ITP, cookie deprecation and consent rates across European DTC accounts.
Why it costs you growth
Channels that create demand look expensive, so they get cut first.
Channels that harvest demand look efficient, so they get more.
The overlap between them never appears on any dashboard.
Rising CAC
CFO scrutiny
Every CFO, every quarter.
What would fixing the budget error be worth to you?
Calculate your ROIWe have many touchpoints in our marketing mix. Now we have visibility into the impact of our marketing investment.
The attribution landscape
Simple, deterministic and reported by every platform. But it cannot see the touches that created the demand, or any assist in between. Multi-touch attribution credits the whole observed path, and moves credit in both directions.
Illustrative journey mix for a European DTC account. Shares sum to 100% in both models.
Granular and cross channel. It answers what to do on Monday.
Needs user level signal, which is eroding. Walled gardens stay partly closed.
You cannot value a touchpoint you never recorded. Map the journey before you pick a model.
Map the journey first
If that is your data, every model you put on top will agree the last channel did the work. Not because it did, but because it was the only one recorded.
Share of all paid orders, for a European DTC brand.
Days from first click to purchase. Same orders, clicks only on both sides. The longer the journey, the more of it GA4 cuts off.
The biggest struggle was that our online prospecting channels were a bit of a black box, because of GA4’s attribution model and the very low results we saw in it.
See what crediting the whole journey is worth at your spend.
Calculate your ROIJourney coverage comes first
Kickbite rebuilds the journey layer by layer, and only models where nothing can be observed.
Illustrative coverage for a European DTC account. Only the last layer is modelled.
Everything observable stays observed. Only consent opt-outs are modelled.
Measured conversions are never zeroed and reallocated by a regression.
If you cannot open one order and see why it was credited, that is an opinion.
The closest thing to a holy grail
One weighting, chosen by a human, applied identically to every brand and every path. Cheap and transparent. Wrong in a specific, repeatable direction.
Weights are learned from click paths, at channel level only. Better than a rule. Still not your funnel, and still click first.
Fuzzy logic reads how far along a buyer was at each touch, learning from the ad level interactions.
Attribution, put to work
Attribution is the only method that sees conversions at ad level, every day. MMM and incrementality answer different questions at a different altitude.
A human decision, taken on a cadence.
Read credit at ad level
Not channel totals. The ad, the creative, the placement.
Separate initialisers from closers
Fund what starts journeys. Cap what only shows up at the end.
Reset windows per category
Your own path length, not a platform default.
Case study · Sunweb
At Sunweb it takes 29 days to cover 80% of booking journeys. GA4 gave nearly all the credit to the last stop, so Meta and Display ran as fixed awareness budgets nobody could defend. Kickbite replaced GA4 as the attribution source and matched Sunweb’s backend bookings 100%.
Total bookings stayed the same. Only the credit moved.
Higher Display prospecting ROAS in Kickbite than GA4 showed.
Same campaigns, same spend. The only thing that changed was the measurement.
ROAS gap inside a single budget line, between consideration and awareness campaigns.
It took 23% of the prospecting budget and produced 43% of the revenue.
Nine creatives produced 43% of revenue from 17% of spend. Seven produced no revenue at all.
Of the paid budget, February to July, as both got ROAS targets.
Beaten every month, by 26.5% on average.
First three months vs. last three.
Over six months. Beat target in five of six.
“Display and Meta used to be awareness budgets nobody could defend, now they carry targets like every other channel.”
In their own words
Find out what a three-month test would pay back for you.
Calculate your ROIWhat good looks like
MTA, MMM and Geolift all make sense. They answer different questions at different altitudes. Decide up front which one owns which decision.
Open a single order and see which touchpoints were recorded, which were views, and why the credit landed where it did.
Within three to six months, you should be able to point to an uplift in acquisition efficiency and say where it came from.
ROI calculator
Plug in your own numbers to see what a three-month Kickbite test pays back if better attribution lifts your ROAS at the same spend.
Estimate only. The +5% ROAS uplift is an assumption, not a Kickbite guarantee.
We’ll show you how Kickbite would attribute an order journey like yours, and which campaigns are really winning you new customers.