First touch vs last touch attribution for SaaS revenue
Compare first touch and last touch attribution with a SaaS payment example, understand their blind spots, and choose a rule you can explain.
By Graytower team
TL;DR
First touch credits a customer's earliest known acquisition source; last touch credits the latest eligible source before conversion. They answer different questions about the same payment, and neither proves causation. Pick a rule that matches the decision you are making, state the observation window, and show unattributed revenue separately.
How can one payment tell two stories?
Imagine someone discovers your SaaS through an organic search result on Monday, reads a comparison page through an email link on Thursday, and pays $49 after a direct visit on Friday. If all visits belong to the same known visitor, first touch points to organic search. A simple last touch rule points to direct. A last-non-direct rule might point to email.
The $49 payment does not become $147 because three visits existed. Attribution rules distribute or assign credit for one payment; they do not create revenue. They also do not tell you which visit was necessary for the purchase.
| Question | Useful view | Main limitation |
|---|---|---|
| Where did known customers first arrive? | First-known source | Earlier untracked exposure is invisible |
| What preceded known purchases? | Last eligible touch | Recent navigation can overshadow discovery |
| Did a change cause more purchases? | Controlled experiment | Requires a sound test design and enough data |
What does first touch measure well?
First touch is useful when the decision concerns acquisition: which observed channels introduce people who eventually pay? It also resists a later direct visit taking credit away from a known earlier source. Graytower uses the visitor's first-known source for acquisition reporting.
The qualifier “known” is essential. A visitor might have heard about you in a podcast or on a colleague's screen before the first tracked visit. A first-touch chart starts at the first observation your system can link, not the first moment of awareness.
When is last touch useful?
Last touch can help study the final route to checkout, such as whether a pricing email was followed by purchases. But it is highly sensitive to implementation details. Some systems give direct traffic the last touch; others ignore direct when an earlier non-direct source exists. Cross-device purchases or missing sessions can break the chain entirely.
Do not compare first-touch and last-touch dashboards as though they were the same metric. Document which interactions count, the lookback window, direct-traffic handling, identity rules, and whether refunds change the reported amount.
How should a small SaaS team use attribution?
Choose a single primary acquisition rule and make it legible in the UI. Keep total, attributed, and unattributed revenue visible by currency. Investigate a few real payment journeys before relying on aggregate channel rankings. If the visitor-to-payment link is weak, fix unattributed revenue before debating models.
Then use experiments, customer research, and funnel data to test a proposed marketing change. Attribution identifies an observed association. It is a starting point for a decision, not a causal verdict. See SaaS revenue attribution for the underlying payment and visitor links.
Frequently asked questions
What is the difference between first touch and last touch attribution?
First touch assigns credit to the earliest known acquisition source. Last touch assigns it to the latest eligible source before a conversion. Both are rules for describing observed journeys.
Which attribution model is best for SaaS?
Use first touch to study acquisition and last touch to study the final path to purchase, where reliable journey data exists. Neither is universally best; the useful choice depends on the decision and available evidence.
Does direct traffic overwrite an earlier source?
Not under a first-known-source rule. A later direct visit does not replace an earlier recorded source. Last-touch implementations vary in how they treat direct visits, so document the rule explicitly.