time decay attribution modeltime decay attribution formulatime decay half-life
Time Decay Attribution Model: What It Is and When to Use It in B2B SaaS
How the time decay attribution model is calculated, how to pick a half-life, when it beats last-touch or linear, and when a hybrid fits a long B2B SaaS cycle.
Request a quote and our experts will contact you within 24 hours with tailored solutions and pricing.
Last updated: Nov 6, 2026
Time Decay Attribution Model: What It Is and When to Use It in B2B SaaS
By the RevOps for B2B SaaS Editorial Team | Last reviewed: November 4, 2026 (newest source cited: August 2026)
The time decay attribution model is a multi-touch method that gives each marketing touchpoint credit based on how close it occurred to the conversion, so recent touches earn more and earlier touches earn less. For B2B SaaS, that sounds right. Deals take months, and the touches nearest the close feel decisive.
The catch is what the model hides. Dreamdata's 2026 benchmark puts the average B2B journey at 272 days (PPC Land, reporting Dreamdata, 2026). Over that span, a recency weight can quietly erase the touches that created demand. This guide covers the formula, a worked example, how to choose a half-life, and when a hybrid fits better. For the wider model landscape, see the attribution models pillar.
Key Takeaways
The time decay attribution model weights each touchpoint by recency using the formula 2 raised to negative t over the half-life, then normalizes weights to 100%.
The half-life setting decides the result: a 7-day half-life behaves almost like last-touch on a long B2B SaaS cycle, while 30 to 60 days spreads credit more evenly.
Time decay beats last-touch on multi-touch deals, but it undervalues early demand creation and cannot credit touches that were never tracked.
Google Analytics 4 retired time decay in November 2023, so it must be rebuilt in a warehouse, a CRM or an attribution platform.
A hybrid works best for long cycles: time decay for in-funnel optimization, first-touch or W-shaped for sourcing, and budget floors for demand creation.
What is the time decay attribution model and when should it be used?
The time decay attribution model is a multi-touch method that gives more credit to touchpoints closer to the conversion. It suits B2B SaaS, where Dreamdata's 2026 benchmark measured an average journey of 272 days (PPC Land, reporting Dreamdata, 2026). It shrinks credit for early demand creation, so a hybrid often fits better.
A row of round markers growing brighter toward a flag at the right end of a timeline
Marketing Ops and revenue operations (RevOps) teams usually reach for time decay when last-touch attribution feels too blunt. Last-touch attribution is a method that gives 100% of the credit to the final touchpoint before conversion. Time decay keeps the logic that late touches matter but pays earlier touches something too.
When it is a good fit
Time decay works well when three conditions hold:
The sales cycle spans weeks or months, with several tracked touches per deal.
Most touches sit in a CRM or warehouse with dates attached.
The question is which in-funnel campaigns and programs closed pipeline, not which channel created demand.
When it is a poor fit
It fits badly when the main decision is how much to spend on early demand creation. It also fits badly when deals have few tracked touches, because a weighting scheme cannot spread credit it never sees.
How is the time decay attribution formula calculated?
The time decay formula gives each touchpoint a raw weight of 2 raised to the power of negative t divided by the half-life, where t is the days between the touch and the conversion. Weights are then normalized to total 100% and each touch receives that share of the deal amount. Many tools default to a seven-day half-life (Demandspring, 2026).
A halving curve drawn on graph paper beside a pencil and calculator
The half-life is the number of days over which a touchpoint's credit falls by half. A touch on the conversion day has a raw weight of 1. A touch one half-life earlier has 0.5, and a touch two half-lives earlier has 0.25.
A worked example at deal level
The numbers below are illustrative arithmetic, not research findings. Take one hypothetical $100,000 closed-won deal with four touches, 90, 60, 30 and 7 days before close. Each raw weight is 2^(-t/half-life), rounded to three decimals.
Touch (days before close)
7-day half-life
30-day half-life
60-day half-life
90 days (first touch)
0.000
0.125
0.354
60 days
0.003
0.250
0.500
30 days
0.050
0.500
0.707
7 days (last touch)
0.500
0.851
0.922
Sum of raw weights
0.553
1.726
2.483
To get a touch's credit, divide its raw weight by the column sum and multiply by $100,000. With a 7-day half-life, the last touch gets 0.500 / 0.553, about 0.904 of the deal, or roughly $90,400. That is last-touch with a small tail. With a 60-day half-life, the first touch gets 0.354 / 2.483, about 0.142 of the deal, or roughly $14,200. Same journey, very different story.
At opportunity level, the same weights work like this:
Pull every touch dated before the opportunity was created or closed.
Compute t for each touch against the close date.
Compute the raw weight, sum the weights, and divide each by the sum.
Multiply each share by the closed-won amount and roll up by campaign.
What half-life should a B2B SaaS team use?
A B2B SaaS team should set the time decay half-life from its own sales cycle rather than defaulting to seven days. Published practitioner guidance disagrees: Trackingplan recommends 14 to 30 days for B2B (Trackingplan, 2026), while Demandspring suggests 30 to 60 days for a six-month cycle (Demandspring, 2026).
Three hourglasses of different sizes standing in a row on a wooden shelf
Both are practitioner opinions, not studies. Trackingplan matches the half-life to the median time to conversion, and recommends 3 to 5 days for e-commerce. It also warns that a 14-day half-life still compresses too much credit into the last two weeks of a 60-day cycle. Demandspring warns that a mismatched half-life produces last-touch results in disguise, so treat that as a risk, not a certainty.
Source
Suggested half-life
Context
Type
Trackingplan (2026)
14 to 30 days
B2B
Vendor guidance
Demandspring (2026)
30 to 60 days
Six-month B2B cycle
Practitioner guidance
Trackingplan (2026)
3 to 5 days
E-commerce
Vendor guidance
How to derive the half-life from your own data
Export first-touch date, opportunity-created date and closed-won date for each deal.
Calculate days from first touch to opportunity and from first touch to closed-won.
Take the median for each, split by segment such as deal size or product.
Pick a half-life as a fraction of that median. As an illustration only, a hypothetical 180-day median at one sixth to one third gives 30 to 60 days, which matches Demandspring's range. That ratio is an editorial inference, not published guidance.
Run the model at two or three half-lives and compare the campaign ranking.
Your own median should override any published range.
Why the setting needs a yearly recheck
Dreamdata's 2026 report measured 272 days, 88 touchpoints and 10 stakeholders, up from 211 days, 76 touchpoints and 6.8 stakeholders a year earlier (PPC Land, reporting Dreamdata, 2026), as the chart shows; Dreamdata is an attribution vendor, so treat this as vendor research.
Grouped Bar chart: B2B buyer journey complexity: Dreamdata 2025 vs 2026 reports. Highest: Touchpoints per journey at 76.
When does time decay outperform last-touch or linear attribution?
Time decay attribution outperforms last-touch when a deal has many touches over weeks or months, because it rewards late touches but still credits earlier nurturing. It beats linear attribution, which splits credit equally, when recent touches truly drive the close. Last-touch stays reasonable for short cycles: 53.5% of marketers in a 2026 Ruler Analytics survey called it somewhat effective (Ruler Analytics, 2026).
The survey did not show its sample size, so read it as a sign that last-touch is not useless, not as proof. The practical contrast looks like this:
Situation
Better fit
Reason
Short cycle, one buyer, few touches
Last-touch
Little credit left to spread
Multi-month cycle, several tracked touches
Time decay
Rewards nurture and closing touches
Every touch seen as equally important
Linear
Neutral baseline, no recency bias
Many buyers, untracked early research
Hybrid
Neither model sees the early touches
The last-touch breakdown covers where that model fails. Linear is also a useful check: run it beside time decay, and a large gap between the two shows how much the recency weighting is moving credit.
What we found: In GrowthSpree's June 2026 audit of 104 Google Ads accounts, 91% had no Google click ID (GCLID) reaching the customer relationship management (CRM) system, and 89% ran default attribution. After a 90-day attribution fix, the share of closed-won deals traced to a campaign went from 0% to 67%.
Attribution only works on touches the CRM can see. Fix click ID capture before tuning a half-life.
Does time decay attribution undervalue top-of-funnel marketing?
Yes, time decay attribution undervalues early demand creation, because its weights shrink with every day before the conversion. Early brand touches may also never be tracked. In 6sense's 2025 research, the winning vendor was still on the Day One shortlist 95% of the time (6sense, 2025), and no recency weighting can credit that.
A faint lantern at the start of a long dim path with brighter lanterns near the end
Kerry Cunningham, Head of Research at 6sense, put the stakes this way: "You must become the preferred vendor during Selection." (Bizibl, 2025)
Ty Heath of the LinkedIn B2B Institute makes the same point about memory: "If you're not on their mind, you're not on the list." (Closed Loop, 2025) Those early impressions rarely show up in a CRM touch table.
Buyers are also moving further from CRM-logged touches. Gartner's May 2026 release (645 buyers) reports 67% of B2B buyers prefer a rep-free experience and 70% prefer fully digital self-service (Gartner, 2026). More of the journey sits outside the data.
The fix is to stop treating the time decay output as the whole story. Add a self-reported "how did you hear about us" field. Hold budget floors for early programs. Never cut top-of-funnel spend because it scores low under a recency model.
Does GA4 still offer time decay attribution, and how do you build it?
No, Google Analytics 4 (GA4) no longer offers time decay attribution. As of August 2026 it provides three models: data-driven, paid and organic last click, and Google paid channels last click (Mar-Sci, 2026). Time decay was retired in November 2023, so teams must build it in a warehouse such as BigQuery, in a CRM or in an attribution platform.
A steel ruler and a drawing compass resting together on a workbench
Some older guides still describe time decay as a GA4 setting. That is out of date. This account of the retirement rests on one secondary source, Mar-Sci, which also lists first click, linear and position-based among the retired models. Confirm it against Google's own documentation before changing a reporting setup.
Where to build it
What you need
Best for
Warehouse (BigQuery)
Touch table with dates, SQL for the formula
Teams with analytics engineering
CRM reporting
Touch records tied to opportunities
Teams with clean opportunity stages
Attribution platform
Connected ad, web and CRM data
Teams without warehouse skills
Build steps in a CRM or BigQuery
Load touches (date, campaign, contact or account) into a table, from the GA4 BigQuery export or CRM activity records.
Join each touch to its opportunity and the closed-won date.
Compute t, the raw weight at your chosen half-life, and each touch's normalized share.
Multiply shares by the closed-won amount and roll up by campaign.
Run linear or W-shaped credit on the same deals. W-shaped attribution is a multi-touch method that gives the largest shares of credit to first touch, lead creation and opportunity creation. If the models agree on the top campaigns, confidence is high. If they disagree, test with holdouts, incrementality tests or self-reported answers.
When is time decay the wrong choice, and what replaces it?
Time decay attribution is the wrong primary model in three cases: early-touch spend is large, most deals lack a clean opportunity stage, or buying groups are large. Dreamdata's 2026 benchmark counts 10 stakeholders per deal (PPC Land, reporting Dreamdata, 2026). A hybrid usually replaces it, pairing time decay for in-funnel optimization with first-touch or W-shaped credit and budget floors for demand creation.
Hybrid attribution is the practice of running two or more attribution views together and using each for the decision it suits. Use these decision rules:
Signal
Lean toward
Why
Cycle of a few weeks, one buyer
Last-touch or time decay
Few touches to weigh
Cycle of several months
Time decay with a long half-life, plus first-touch
Recency alone hides sourcing
Many deals lack an opportunity stage
Linear or first-touch for now
Time decay needs dated stage data
Large share of spend on early programs
Hybrid with budget floors
Protects demand creation
Ten-person buying groups
Account-level credit
Contact-level touches miss the group
Limits to state plainly
This is not a fit for teams with no CRM touch data, for very short cycles, or for anyone who needs to prove brand impact. Time decay also treats the journey as one person's, while buying groups are large. Credit at the account or buying-group level where the data allows.
What is the formula for the time decay attribution model?
The time decay attribution formula sets each touchpoint's raw weight to 2 raised to the power of negative t divided by the half-life, where t is the number of days between the touch and the conversion. Weights are then normalized to total 100%, and each touch receives its share of the deal amount.
Does GA4 still offer time decay attribution?
No, Google Analytics 4 no longer offers time decay attribution. As of August 2026 it provides three models: data-driven, paid and organic last click, and Google paid channels last click. Time decay was retired in November 2023, so teams must rebuild it from the BigQuery export, in a CRM or in an attribution platform.
How is time decay different from last-touch attribution?
Time decay attribution shares credit across every touchpoint, weighted by how recently each occurred, while last-touch attribution gives 100% of the credit to the final touch before conversion. A time decay model with a very short half-life behaves almost like last-touch, so the half-life setting decides how different the two results are.
When should I use time decay attribution instead of linear?
Use time decay attribution instead of linear attribution when recent touches plausibly drive the close and the cycle is long enough for timing to matter. Linear attribution splits credit equally across all touches, which suits teams that see every touch as equally important or want a neutral baseline to compare against.
Time decay vs W-shaped attribution: which is better for B2B?
Time decay attribution and W-shaped attribution suit different decisions. Time decay suits in-funnel campaign optimization because it rewards recent touches, while W-shaped attribution credits first touch, lead creation and opportunity creation, which makes it better for sourcing and budget decisions. Many B2B SaaS teams run both and compare the results.
Can time decay attribution be combined with first-touch or multi-touch models?
Yes, time decay attribution can be combined with first-touch or other multi-touch models in a hybrid. A common design uses time decay to optimize campaigns inside the funnel and first-touch or W-shaped credit to track sourcing, with budget floors that protect early demand-creation programs from being cut.
Conclusion
The time decay attribution model is a useful lens for long B2B SaaS cycles, as long as it is not the only lens. Keep these points:
Use the formula 2 raised to negative t over the half-life, normalize to 100%, and apply it to the closed-won amount.
Set the half-life from your own median cycle, and recheck it yearly as journeys lengthen.
Choose time decay over last-touch when deals have many dated, tracked touches.
Rebuild it outside GA4, since Google retired it in November 2023.
Pair it with first-touch or W-shaped credit, a self-reported field and budget floors for early demand.
If your team is deciding which model fits its stage and data, bring the CRM and reporting question to a free consultation: Meet with an expert.
Written by
Ishan Manchanda
Ishan Manchanda is co-founder of GrowthSpree, a B2B SaaS marketing consultancy founded in 2017 that has managed $60M+ in ad spend across 300+ B2B SaaS companies. He focuses on demand generation, go-to-market strategy and pipeline growth, and has been featured in Silicon Valley Times, SaaS Mag and the Unbounce podcast.
Ishan ManchandaNov 6, 2026
