There is no perfect attribution model. The job is picking a model whose shape matches your sales motion, applying it consistently, and resisting the temptation to switch models every time a channel underperforms.
Attribution is the practice of assigning credit for revenue to the marketing and sales touchpoints a buyer interacts with on the way to a purchase. Common models include first touch, last touch, linear, time-decay, U-shaped, W-shaped, and data-driven. Each distributes credit differently. Attribution lets teams decide which channels, campaigns, and plays actually generate pipeline, instead of relying on last-touch guesswork.
Attribution answers one practical question: which channel do we spend the next dollar on? The five points below cover the model landscape, what each one is good for, and where the common traps live.
A buyer sees a LinkedIn ad, reads a blog, attends a webinar, downloads a case study, and books a demo before buying. Attribution decides how much revenue credit each of those five touchpoints receives. The decision drives next quarter budget allocation across every channel.
First touch gives 100 percent of credit to the first known interaction. Useful for understanding what generates net-new awareness. Underweights middle-funnel content and over-rewards channels that happen to run attribution first (often paid search for branded queries, which is really demand capture).
Last touch gives 100 percent of credit to the final touchpoint before conversion. Easy to implement, which is why most ad platforms default to it. Over-rewards bottom-funnel channels (branded search, retargeting, direct) and makes awareness channels look worthless even when they are the actual drivers.
Linear splits credit evenly across every touchpoint. Time-decay weights more recent touches heavier. Both acknowledge that multiple touchpoints contribute, which fits reality better than first or last touch alone. Simpler to explain than U-shaped or W-shaped and often good enough for mid-market revenue teams.
U-shaped weights first and last touch 40 percent each and splits 20 percent across the middle. W-shaped adds a third anchor at the mid-funnel lead-conversion event. Both reward the moments that actually matter in a B2B motion: awareness, qualification, and close. The downside is more config and more arguments.
Data-driven attribution uses machine learning to assign credit based on which touch patterns correlate with conversion in your actual data. More accurate in principle, less debatable in practice. Requires meaningful volume (hundreds of conversions per model period). Common at enterprise scale, overkill for small revenue teams.
Attribution model choice should match your sales motion, your buying committee size, and your reporting volume. The notes below are the common fit patterns.
For transactional SMB motions with a 7 to 30 day cycle and a single buyer, last touch is usually directional enough. The buyer path is short, the number of touches is small, and the last-touch channel is often the real driver. The attribution investment is better spent elsewhere.
For 90 to 270 day B2B cycles with multiple stakeholders, U-shaped or W-shaped attribution is the common pick. These models give enough credit to awareness and middle-funnel touches to keep channel investments honest while still rewarding bottom-funnel close plays.
Businesses closing hundreds of deals per month with meaningful channel diversity can benefit from data-driven attribution. The underlying ML approach needs volume to find signal. Below roughly 100 to 200 conversions per model period, data-driven attribution over-fits noise and makes rules-based models more reliable.
Businesses with both SMB and enterprise motions often run different attribution models by segment. Last touch for SMB, W-shaped for mid-market, data-driven for enterprise. The reporting layer rolls up correctly only if the segment field is clean on every opportunity.
Attribution reports are easy to generate and easy to mislead with. Four traps catch most teams, and spotting them in your own numbers is often more valuable than switching models.
Last-touch models typically credit branded search queries (search for your company name) as the converting channel. These searches are usually demand capture, not demand generation. The upstream channel that caused someone to search your brand is the actual driver, and attribution models that credit branded search as origin get budget decisions wrong.
Direct traffic (someone typing the URL) is almost always a downstream effect of something that happened earlier. Attribution models that reward direct traffic as the acquisition channel are typically crediting a prior touchpoint whose cookie expired or was blocked. Treat direct traffic as a reporting artifact, not a channel.
Significant portions of modern B2B buying happen in private Slack conversations, podcasts, in-person events, LinkedIn DMs, and word-of-mouth. Attribution tools cannot see most of these. The channels most responsible for pipeline often look invisible in attribution reports. Self-reported attribution on inbound forms partially closes the gap.
Model-shopping (switching models until the preferred channel looks good) destroys the credibility of attribution reporting internally. Pick a model, run it for 2 to 4 quarters, and change only when the business motion itself has shifted. Consistency beats precision for most decision-making.
Attribution models that only cover marketing touches ignore the sales activity that moves deals forward. In B2B, the SDR, the AE, and the executive sponsor are often as important as any marketing touch. Attribution that stops at marketing hides the real revenue motion. The best models blend marketing and sales touches on a common timeline.
Many attribution reports treat small sample sizes with high confidence. A channel with 12 conversions and 11 wins looks like a 92 percent win rate until you see the overlap with other channels. Report attribution with confidence intervals or do not report conversion rates at low sample sizes.
Attribution is 10 percent model choice and 90 percent plumbing. The plumbing is what actually determines whether your reports mean anything.
List every channel that generates a known touch: paid search, paid social, organic search, direct, email, webinar, event, outbound SDR, inbound form, partner referral, review site. Each channel needs a UTM or source definition. Without a touchpoint catalog, the attribution report mixes apples and oranges.
Every outbound link from marketing (ads, emails, QR codes, partner co-marketing) should carry a UTM. A UTM template with required source, medium, campaign reduces drift. Attribution breaks silently when half the links lack UTMs and the data defaults to direct.
Attribution needs timeline data on the contact (what did this person see) and the account (what did anyone at this company see). Account-level attribution matters in B2B because the person booking the demo is often not the person who saw the ads. The CRM should stamp both.
The inbound demo form should include a short "how did you hear about us" dropdown. Self-reported attribution closes the dark-social gap and often reveals channels that trackable attribution misses entirely. Compare self-reported to trackable to understand your actual blind spots.
The useful attribution report shows revenue by channel by segment by motion (new logo versus expansion). Rolled-up channel reporting hides where a channel wins in one segment and loses in another. In Strkr, deal attribution fields can be custom per segment so reporting rolls up correctly.
Attribution reports are directional, not operational. Weekly reviews over-weight noise. Monthly or quarterly cadence is the right rhythm, with ad hoc deep dives when a channel changes materially. Give the data time to accumulate before making budget decisions.
Strkr tracks every marketing and sales touchpoint on the contact and account timeline. Deals carry source and campaign fields natively. Pricing is $29 Starter, $59 Pro, $119 Scale per seat per month. Strkr Marketing is in private beta.
First-touch assigns 100 percent of revenue credit to the first known interaction a buyer had with your brand. Last-touch assigns 100 percent to the final interaction before conversion. First-touch over-rewards awareness channels. Last-touch over-rewards bottom-funnel capture channels like branded search and direct. Most sophisticated teams use multi-touch models to split credit across the whole journey.
Multi-touch attribution assigns revenue credit across every known touchpoint a buyer interacted with, rather than giving all credit to one touch. Common multi-touch models are linear (even split), time-decay (recent touches weighted more), U-shaped (first and last weighted 40 percent each), and W-shaped (first, lead-conversion, and last weighted 30 percent each).
Data-driven attribution uses machine learning to assign credit to touchpoints based on which patterns actually correlate with conversion in your historical data. More accurate in principle, more complex in practice. Requires meaningful conversion volume (usually hundreds per model period). Common at enterprise scale, often overkill at mid-market or SMB where rules-based models work fine.
Marketing attribution covers touches owned by marketing: ads, content, email, webinars, events. Sales attribution covers touches owned by sales: SDR outreach, discovery calls, demos, executive engagements. In B2B both matter. The best attribution models blend them on a single contact and account timeline so no team is credited more than the real contribution.
B2B buying cycles are long, involve 6 to 11 stakeholders, and include many offline or dark-social touches (podcasts, Slack, LinkedIn DMs, word-of-mouth) that no tracking tool can see. The person booking the demo is often not the person who saw the ads. Self-reported attribution on inbound forms partially closes the gap but no model captures everything.
For a new team under 100 conversions per month, last-touch is usually directional enough for budget decisions, and the time is better spent cleaning UTMs. Between 100 and 500 conversions, U-shaped or W-shaped is the common upgrade. Above 500, data-driven becomes viable. Simplicity beats precision until the volume supports more sophisticated models.
In Strkr, Marketing (currently in private beta) carries UTMs and campaigns on contacts and accounts. The CRM stores touchpoints on the activity timeline. Deals can be tagged with first touch, last touch, and conversion touch for custom reporting. Strkr AI assists on summarizing deal touchpoint history. Multi-touch attribution reporting ships in Marketing GA.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
