Signal Sparrow
Platform — Attribution models

Every attribution model, side by side.

There's no single "correct" way to credit a sale — so Signal Sparrow computes first-click, last-click and every multi-touch model at once, honestly labeled, and lets you reattribute your entire history the instant you switch. No black-box score, no argument about whose number is right.

Start freeSee pricing

Compare models side by sideReattribute history instantlyReconciled to Stripe

Attribution models, defined

How credit for a conversion gets divided.

An attribution model is the rule that decides which marketing touches get credit for a conversion — and how much. Single-touch models hand 100% of the credit to one interaction (the first click or the last). Multi-touch models split credit across every touch in the journey, weighted by position or recency.

The model you choose changes which channels look like heroes and which look like waste — which is exactly why you should see several at once. Signal Sparrow applies each model to the same revenue attribution data, so the comparison is apples-to-apples and every model still reconciles to Stripe.

Journey credit split · $149/mo conversionMetaLinkedInGoogle
First-Click
Meta (100%)
100%——
Last-Click
Google (100%)
——100%
Position-Based (40/20/40)
Meta (40%) · LinkedIn (20%) · Google (40%)
40%20%40%
Six ways to credit a journey

Pick the lens that fits how you sell.

Every model runs on the same Stripe-reconciled data, so switching lenses never changes your total revenue — only how it's distributed.

  • 01

    First-click

    100% to the first interaction

    Best for: Valuing demand gen and top-of-funnel discovery

    100%——
  • 02

    Last-click

    100% to the final interaction

    Best for: Bottom-funnel, closing channels, quick cycles

    ——100%
  • 03

    Linear

    Equal credit to every touch

    Best for: Full-journey visibility with no bias

    33%33%33%
  • 04

    Time-decay

    More credit to recent touches

    Best for: Longer sales cycles where recency matters

    20%30%50%
  • 05

    Position-based (U-shaped)

    40% first, 40% last, 20% middle

    Best for: Rewarding both discovery and conversion

    40%20%40%
  • 06

    Custom / W-shaped

    Your own weights per stage

    Best for: Teams with a defined funnel and clear stages

    30%40%30%
Why one model isn't enough

The model you pick quietly decides your budget.

Commit to a single model and you bias every decision. Trust the ad platforms' models and you get numbers that never agree.

Last-clickMetaLinkedInGoogle
——100%
Demand creatorsZero credit
Problem 01

Last-click over-credits the closer

Default last-click attribution hands all the credit to brand search or direct — starving the top-of-funnel channels that created the demand in the first place.

Meta+1
Google+1
GA4+1
Claimed 3You had 1
Problem 02

Every platform grades its own homework

Meta, Google and GA4 each apply their own model and window, so they collectively claim more conversions than you actually had. The totals can't all be right.

Attribution scoreProprietary
How was credit assigned?Hidden
Problem 03

Blended scores you can't audit

Some tools hide the model behind a proprietary "score." If you can't see how credit was assigned, you can't defend the decision it drives.

How it works

Capture the whole journey, then choose your lens.

    • MetaMobile · day 0
    • LinkedInDesktop · day 6
    • GoogleBrand search · day 14
    Converted in Stripe
    Step 01

    Capture every touch

    First-party, server-side tracking records the full sequence of clicks — across devices and weeks — that led to each Stripe conversion.

  2. First-click
    Last-click
    Linear
    Time-decay
    Position-based
    Step 02

    Apply every model

    We compute first-click, last-click, linear, time-decay and position-based credit for the same journeys simultaneously, all reconciled to MRR.

  3. Model
    Last-click
    Lookback
    30 days
    History reattributed
    Step 03

    Reattribute instantly

    Switch the model or the lookback window and your entire history recomputes on the spot — no re-tagging, no waiting, no data project.

The same journey, five ways

Watch the credit move as the model changes.

Take one customer who touched Meta, then LinkedIn, then converted on Google brand search. Here's how a single $149/mo conversion's credit is distributed under each model — same journey, very different stories. This is why you compare before you decide, then pair the result with CAC payback to see which model reflects reality.

ModelMetaLinkedInGoogle
First-click100%0%0%
Last-click0%0%100%
Linear33%33%33%
Time-decay20%30%50%
Position-based40%20%40%

Credit for one conversion. Under last-click, Meta looks worthless; under first-click, it looks essential. The truth is in the comparison.

What multi-model visibility unlocks

Change the model, keep the truth.

Models computed on every journey
5+
First-clickPosition-based
Maximum lookback window
90 days
7 days90 days
Of revenue reconciled, every model
100%
Attributed + direct/unknown= Stripe

Reattribution recomputes credit across your history in place; totals always reconcile to Stripe MRR regardless of model.

Reconciled to Stripe

Different models, one honest total.

Switching models redistributes credit — it never invents revenue. Under every model, attributed plus direct/unknown equals your Stripe net MRR for the period, to the cent. So you can debate which channel deserves the credit without ever questioning whether the total is real.

See revenue attribution
Reconciled to Stripe91.4% match
Attributed
$42,180
Direct/Unknown
+$6,920

Net MRR in Stripe

=$49,100

Difference $0.00
Your model, your window

Set the lookback, and rewrite history safely.

Attribution isn't just the model — it's the window. A 7-day window and a 90-day window tell very different stories for trial-led SaaS. Signal Sparrow lets you set the lookback per report and reattribute your entire history against it instantly, so you can pressure-test assumptions without losing the underlying data.

  • Adjustable lookback

    7, 30, 90 days or custom lookback periods.

  • Instant recompute

    History reattributes in place immediately.

  • Non-destructive

    Raw touches are never overwritten in the DB.

  • Compare models

    View two side by side before committing.

Attribution window controlActive settings
Active model
Position-based
Lookback window
90 days
Reattribution
Applied · reconciled
30-minute setup

Every model, running in 30 minutes.

Add the first-party snippet and connect Stripe and your ad accounts over OAuth — Signal Sparrow captures full journeys and computes every attribution model automatically, no tagging plan or data team required. Switch models and windows freely from day one.

Start freeSee pricing
Setup0 of 3
index.htmlPaste once
<script async
  src="https://data.yourdomain.com/sp.js">
</script>
2Stripe & ad accountsConnect
3Attribution modelsComputing
First conversion, attributed to its campaign

Questions, answered.

Everything teams ask before switching their attribution to Stripe truth. Still curious? Talk to us.

A multi-touch attribution model splits credit for a conversion across every marketing interaction in the customer's journey, rather than giving 100% to a single touch. Common multi-touch models include linear, time-decay and position-based (U-shaped).

First-click gives all the credit to the interaction that introduced the customer; last-click gives it all to the final interaction before conversion. First-click favors demand-gen channels, last-click favors closing channels — which is why looking at both matters.

There's no universal answer — it depends on your funnel. Short cycles often lean last-click; longer, multi-touch journeys suit time-decay or position-based. Signal Sparrow computes several at once so you can compare rather than guess, and switch without losing data.

It redistributes credit across your existing journeys — instantly and non-destructively. Your raw touch data and your total revenue never change; only how credit is assigned between channels does. You can switch back at any time.

The attribution window is the lookback period in which touches can earn credit for a conversion. For trial-led SaaS where paying conversions happen weeks later, a too-short window credits no one. You can set the window per report and reattribute instantly.

Yes. Every model runs on the same Stripe-reconciled data, so attributed plus direct/unknown always equals your net MRR for the period, to the cent — no matter which model you're viewing.

Ready to see which ads actually pay?

Start free, connect Stripe in 30 minutes, and get attribution that reconciles to the cent.

Start freeSee pricing
  • Free to start — no credit card
  • Live data in 30 minutes