OwnlyticsFirst touch vs last touch attribution for solo founders
First touch attribution credits a sale to the visitor's very first visit, so it shows what introduced people to you. Last touch credits the visit right before they bought, so it shows what closed the sale. For most solo founders, the practical answer is to pick one as your default, read the other as a cross-check, and not agonize over it.
- First visit: a blog postx.comFirst touch gets the credit under this model
- Second visit: pricing pagenewsletterLast non-direct touch gets the credit under this model
- Paid $49The same sale, two different answers
Fig.The same sale credited two ways. Run both models on the same month and compare. Sample data, not a real site.
What each model does
Suppose a buyer discovers you through a social post on Monday, reads a blog article from search on Thursday, then types your address and buys on Saturday.
- First touch gives the sale to Monday's social post.
- Last touch gives it to Saturday's visit, which is Direct (typed or bookmark).
- Last non-direct gives it to Thursday's search visit, the most recent one with a real source.
Same buyer, same money, three different answers. None is wrong. They answer different questions.
Ownlytics offers all three per website in settings: last non-direct (the default), first touch and last touch. The sources and attribution docs describe each.
Example (made-up numbers for illustration)
Here is one month for an invented product. Twelve sales, $60 each, $720 total.
| Source | First touch sales | Last touch sales | Last non-direct sales |
|---|---|---|---|
| Social | 6 | 1 | 2 |
| Search | 4 | 3 | 7 |
| Direct | 0 | 7 | 0 |
| 2 | 1 | 3 |
First touch says social is your best source. Last touch says Direct is, which is not actionable because nobody can "do more Direct". Last non-direct says search, because it ignores the empty final visits.
The useful reading comes from comparing columns. Social has six first touches but only two last-non-direct credits. That suggests social introduces buyers and search or email finishes the job. If you cut social because "it doesn't close", you may starve the top of the funnel that search later harvests.
Why last touch often favors Direct
Buyers often come back by typing your name or using a bookmark, especially after they've decided. Pure last touch hands them all to Direct, which hides the work that earned their trust.
This is why last non-direct is a sensible default for many small sites. It still credits the last visit with a real source. A checkout return, where the buyer comes back from the payment page, is never credited for a sale when the buyer had an earlier visit, so the payment step itself doesn't steal credit.
We unpack what hides inside Direct in direct traffic explained.
Which one should you use?
| Your situation | Start with | Why |
|---|---|---|
| Few channels, short buying time | Last non-direct | Closest to "what pushed them over" |
| You're deciding where to publish more | First touch | Shows what brings new people in |
| You run paid campaigns toward a checkout | Last touch, then compare | Shows the final nudge, but check Direct |
| Buyers take weeks to decide | Compare first and last | The two will disagree and both matter |
How long your buyers take matters a lot. If most buy in the same visit, the models barely differ. If they take weeks, the gap grows. Ownlytics measures time to buy and visits before buying, and shows which source found each customer versus which one closed the sale, in the path to purchase docs. Check that before debating models.
Fig.Revenue by post under last touch. Sample data, not a real site.
Pitfalls
- Switching mid-stream. Changing the model changes history for the comparison. Pick one for at least a quarter.
- Reading tiny samples. With a dozen sales, a single buyer changes a source's rank. Look at patterns, not single places.
- Ignoring unattributed sales. If many sales have no visitor attached, none of the models is reliable. Fix the link between visits and payments first. The checklist in how to know which blog post made a sale is a good start.
- Treating the model as truth. Attribution assigns credit by a rule. It doesn't prove causation. A buyer who arrived via search may have been persuaded by a podcast you never tracked.
- Devices and cleared cookies. If someone first visits on a phone and buys on a laptop, the two visits may not join. That undercounts first touch in particular.
Do campaign tags change this?
Yes, and in your favor. When you control a link, tag it. In Ownlytics, utm_source and ?ref= win over the referrer, and each post you name appears as its own campaign even where the referrer is hidden. That gives both models cleaner input. Our guide to UTM parameters covers the naming habits.
When this doesn't matter
If you have one channel, or you sell too few units to compare anything, model choice is a distraction. Spend that time getting purchases tied to visitors at all. Likewise, if your sales happen through a sales call or an invoice with no checkout link, no web attribution model will see the real path.
Use last non-direct to see what closes, first touch to see what introduces, and trust a source only when both models agree or you can explain why they don't.
Next step
You can look at how a sample dashboard separates these views in the live demo.
FAQ
Is there a best attribution model?
No. Each answers a different question. Choose by the decision you need to make, and compare against a second model before big changes.
Does changing the model alter my revenue totals?
No. Total revenue stays the same. Only the source each sale is credited to changes.
What about multi-touch models that split credit?
They spread credit across several visits. For a solo founder with limited data, the extra complexity rarely adds clarity. The three single-credit models in the attribution docs are enough to start.
How does a subscription renewal get credited?
In Ownlytics, renewals inherit the visitor of the first payment, so a customer's source stays the same over time.