Attribution
Five models, one set of conversions, no single truth.
Every attribution model is a point of view. Quantiv computes five of them over the same deduplicated conversions so you can see where they agree, where they diverge, and what the difference says about your spend.
Attribution · Meta
$164,200 · last 30 days
| Model | Conversions | CPA | Reads as |
|---|---|---|---|
| First touch | 2,451 | $67 | Scale |
| Linear | 2,219 | $74 | Scale |
| Position based | 2,105 | $78 | Hold |
| Time decay | 2,027 | $81 | Hold |
| Last touch | 1,866 | $88 | Pause |
Same spend, same conversions, five ways of dividing the credit. The $21 spread between first and last touch is the prospecting question - not a number to average away.
Illustrative data
Models
Each model answers a different question.
First touch
Which channels introduce customers to the brand.
Last touch
Which channels close the sale.
Linear
Equal credit to every touch along the way.
Position based
Most credit to the first and last touches, the rest shared between them.
Time decay
More credit to touches closer to the conversion.
Read together
When the models disagree about a channel, that channel's role in the journey is the open question.
Touchpoints
Built on a record of every touch, per person.
Quantiv keeps a touchpoint table: who, when, through which channel, from which campaign, with which click identifiers. Every model reads from the same table.
Identity stitching joins anonymous visits to the person who later signs in or buys, so a journey that started on a phone and ended on a laptop is still one journey.
- Person, timestamp, channel, campaign, UTMs, and click IDs on every touch
- Anonymous and known identities joined through identify and alias
- Conversions deduplicated across server-side and pixel sources
Attribution · Meta
$164,200 · last 30 days
| Model | Conversions | CPA | Reads as |
|---|---|---|---|
| First touch | 2,451 | $67 | Scale |
| Linear | 2,219 | $74 | Scale |
| Position based | 2,105 | $78 | Hold |
| Time decay | 2,027 | $81 | Hold |
| Last touch | 1,866 | $88 | Pause |
Same spend, same conversions, five ways of dividing the credit. The $21 spread between first and last touch is the prospecting question - not a number to average away.
Illustrative data
Attribution is not a scoreboard. It is a set of perspectives, and the disagreement between them is where budget decisions get made.
From insight to action
The media buying agent reads the same table.
Budget proposals from the media buying agent cite the models they rely on and show where those models disagree. You approve the move with the reasoning in front of you.
Proposal · media buying agent
awaiting review
Meta · Prospecting · Skincare US
Increase daily budget $2,400 → $3,000
First-touch and linear both put this set under blended target while last touch puts it over. The spread is prospecting credit, and the set has held CPA for 11 days.
- First touch CPA
- $67 → $71 projected
- Last touch CPA
- $88 → $92 projected
Within the $3,500/day cap · 1 of 2 daily changes used
proposal #4182 · agent · awaiting review · nothing applied
Illustrative data
Questions
Common questions.
Which model should we use?
Use them together. Most teams anchor on one model for reporting and use the others to test how sensitive a decision is to the choice of model.
How is this different from the numbers in Ads Manager?
Ad platforms attribute conversions to themselves using their own windows and rules. Quantiv attributes across channels from your own event data, after deduplication.
See it on your own numbers.
A two-week pilot runs your conversions through Quantiv alongside what you report today. The difference between the two is the business case.