First-Touch vs Last-Touch: Which One Changes Your Decisions
Use first-touch when you need to know what fills the top of the funnel. Use last-touch when you need to know what closes deals. Most businesses need both running at the same time, not one instead of the other. The mistake is picking a religion instead of picking a tool for the question in front of you.
The short answer
If your sales cycle is under 14 days, weight toward last-touch. If it's over 60 days, weight toward first-touch, or you'll starve your top of funnel. Here's why. On the DTC brand I scaled from $100K to $3M+ in monthly revenue, the buying cycle averaged 4 days. Last-touch told us almost everything we needed. Retargeting and branded search got the credit, and rightly so, because that's what closed the sale. But when I ran marketing on a $2.2B infrastructure project, the "sales cycle" was closer to 18 months of stakeholder meetings, RFPs, and political maneuvering. If I'd used last-touch there, I would have credited a single press release or a final site visit for work that actually started with two years of trade press placement and relationship building. First-touch was the only model that matched reality.
What changes the timing
- Cycle length. Short cycles compress the buyer journey into 1-3 touches, so last-touch is close enough to the truth. Long cycles have 10, 20, sometimes 50 touches, and last-touch throws away most of the story.
- Deal size. A $40 impulse purchase and a $400,000 enterprise contract are not attributed the same way, because the second one involves committees, not a single click.
- Channel mix. If you run heavy top-of-funnel spend (content, PR, sponsorships) alongside heavy bottom-of-funnel spend (retargeting, branded search, email), single-touch models will always favor bottom of funnel and slowly starve the top.
- Who's asking. A CFO wants to know what to cut. A CMO wants to know what to build. Those are different questions and sometimes need different models even for the same data.
Signs you are overdue for a change
- Your paid social and content budgets keep getting cut because they "never show ROI," while retargeting keeps getting more budget every quarter.
- Your last-touch report says branded search is your best channel, but branded search volume has nowhere to come from if nothing else builds awareness.
- Leadership keeps asking "where did this customer really come from" and the honest answer is a shrug, because your CRM only logged the final form fill.
- You've killed a channel, seen no dip in results for 60 days, then watched pipeline collapse 90 days later.
- Sales says leads "come out of nowhere" ready to buy, and marketing can't explain what warmed them up.
What happens if you wait too long
You defund the channels that create demand and overfund the channels that harvest it. This is the single most expensive mistake in B2B and considered-purchase marketing, and it's slow enough that nobody notices until the pipeline runs dry.
Here's the math on why it hurts. Say you spend $50,000 a month: $30,000 on retargeting and branded search, $20,000 on content, SEO, and organic social. Last-touch attribution credits the $30,000 with 80% of conversions. Leadership reallocates: cut content to $5,000, push retargeting to $45,000. Conversions hold steady for a month, maybe two, because retargeting is working off the awareness pool you already built.
Then the pool drains. Branded search volume drops because nothing is creating new brand awareness. Retargeting has fewer people to retarget. Three to four months later, total conversions fall 20-30%, and now you're trying to rebuild a content engine that takes 6-9 months to produce results, while the board is asking why marketing "stopped working." I've watched this exact sequence happen twice at two different companies. Both times, the fix took a year. The damage from waiting always costs more than the report you were trying to save.
The most common mistake
The most common mistake is treating attribution as a philosophical stance instead of a practical tool matched to a question. People argue "first-touch is more honest" or "last-touch is what actually converts" as if one is universally true. Neither is. First-touch overstates the value of anything that happens to be first, including channels the customer forgot about entirely. Last-touch overstates the value of anything that happens right before conversion, even if that touch was just picking up a coupon code from a channel that did zero persuading. The fix isn't a better argument, it's running both models side by side and looking at where they disagree. The disagreement is where the real insight lives.
Practical takeaway: pull your last 90 days of conversions and run them through both models. Wherever a channel's share of credit differs by more than 20 points between first-touch and last-touch, that's a channel you're probably misjudging right now. Go look at it before you cut its budget or double it.