---
title: "How to Calculate True CAC When Your Attribution Model Changes Quarterly"
canonical_url: "https://hammadshk.com/blog/how-to-calculate-true-cac-when-your-attribution-model-changes-quarterly"
type: "blog_post"
summary: "Attribution model shifts break CAC tracking mid-year. Learn how to recalculate true customer acquisition cost across model changes, avoid comparison traps, and maintain consistent cost-per-lead benchmarks."
published_at: "2026-11-21"
updated_at: "2026-11-21"
author: "Hammad Sheikh"
category: "Tracking & Attribution"
---

# How to Calculate True CAC When Your Attribution Model Changes Quarterly

> Attribution model shifts break CAC tracking mid-year. Learn how to recalculate true customer acquisition cost across model changes, avoid comparison traps, and maintain consistent cost-per-lead benchmarks.

- Category: Tracking & Attribution

## Article

Why [attribution model changes](/blog/when-manual-attribution-works-better-than-algorithmic-models) Break CAC
-------------------------------------------------------------------------------------------------------------

When you switch from last-click to multi-touch attribution (or vice versa), the credit assigned to each channel shifts instantly. A lead that last-click attributed to Paid Search might split 40/60 between Paid Search and Email under a time-decay model. Your CAC per channel jumps. Your blended CAC jumps. But your actual spend and conversion count did not change.

This creates a false signal: you think your acquisition efficiency improved or worsened when the measurement changed, not the performance. Teams often respond by cutting or doubling down on channels based on the wrong premise.

The Core Problem: Apples-to-Apples Comparison
---------------------------------------------

True CAC = Total spend on channel / Attributed conversions from that channel. The denominator changes when your model changes. If you compare Q2 CAC (last-click) to Q3 CAC (time-decay), you are comparing different metrics with the same name.

You need a single denominator: conversions that would have been counted the same way under both models. The simplest approach is to pick one model as your baseline and restate all prior periods under that model.

Step 1: Lock Your Baseline Attribution Model
--------------------------------------------

Before recalculating anything, choose one model as the source of truth for the year. Last-click is easiest to audit (every conversion has one owner). Multi-touch is more accurate but harder to reconcile retroactively.

Document this decision in a shared sheet or wiki. All CAC benchmarks, board reports, and channel budgets tie to this baseline. When you change the model again next quarter, you will restate all prior periods against the new baseline.

Step 2: Pull Raw Conversion Data (Not Attributed Counts)
--------------------------------------------------------

Export the unattributed conversion list from your analytics platform (Google Analytics 4, Mixpanel, or your CRM). Include: conversion date, user ID, session ID, all touchpoints in the customer journey, and the channel of each touchpoint.

Do not export the platform's pre-calculated "Conversions by Channel" report. That report already applies the model. You need the raw journey data so you can apply your baseline model yourself.

Step 3: Reattribute All Conversions to Your Baseline Model
----------------------------------------------------------

For each conversion, apply your baseline model's logic to the touchpoint sequence. If your baseline is last-click, assign 100% credit to the final channel before conversion. If it is first-click, assign 100% to the first channel. If it is time-decay, weight each touchpoint by a decay factor (e.g., 40% to last, 30% to second-last, 20% to third-last, 10% to earlier).

This is manual in a spreadsheet or scripted in Python/SQL depending on data volume. Most teams with <5,000 monthly conversions can do this in a pivot table. Larger volumes need a script in your data warehouse or a custom query in your analytics tool.

Step 4: Calculate Attributed Conversions by Channel and Period
--------------------------------------------------------------

Sum the reattributed conversions by channel and by quarter or month. You now have a single, consistent conversion count for each channel across all periods, even though your platform's native report changed models mid-year.

Create a table: Channel, Q1 Spend, Q1 Conversions (reattributed), Q2 Spend, Q2 Conversions (reattributed), Q3 Spend, Q3 Conversions (reattributed). This is your source of truth.

Step 5: Calculate CAC Per Channel and Period
--------------------------------------------

Divide spend by reattributed conversions for each channel-period pair. CAC = Spend / Conversions (reattributed).

Example: Paid Search spent $10,000 in Q2 and generated 50 conversions under your baseline model (even though the platform's native report said 55 under last-click). CAC = $10,000 / 50 = $200 per conversion. This is directly comparable to Q1 and Q3 CAC for Paid Search.

Step 6: Separate Model-Driven Changes from Performance Changes
--------------------------------------------------------------

Now compare CAC across periods. If Q2 CAC is $200 and Q3 CAC is $220, ask: Did the channel perform worse, or did the [attribution model](/blog/how-to-calculate-true-roi-on-linkedin-ads-for-long-sales-cycles) shift the conversion count?

Pull the platform's native Q3 report (which uses your new model) and compare the conversion count to your reattributed count. If the platform says 48 conversions but your reattribution says 50, the model shift absorbed 2 conversions. The CAC difference is partly model noise.

Document this split: "Q3 CAC rose $20 due to model change (2 conversions shifted to other channels) and $X due to actual performance decline (fewer conversions per dollar)." This tells you whether to optimize the channel or accept the new baseline.

Common Reattribution Pitfalls
-----------------------------

Do not assume your platform's API or export includes all touchpoints. Google Analytics 4 caps the journey at 25 events; older platforms may cap lower. If a user had 30 interactions before converting, you are missing 5 or more. Document this ceiling and note that your reattribution is approximate.

Do not reattribute offline conversions (phone calls, in-store visits) if you do not have the full journey data. Stick to digital channels only and track offline separately.

Do not mix conversion types (demo request vs. purchase) unless they have the same journey length and channel mix. A demo request may come from a single ad click; a purchase may involve a 10-touch journey. Reattribute each type separately.

When Reattribution Is Not Enough
--------------------------------

If you change models every quarter, reattribution becomes a recurring task. After 2–3 cycles, consider adopting a stable, multi-touch model (time-decay or position-based) and committing to it for 12 months. This reduces the noise and gives you clean, comparable benchmarks.

If your platform does not export raw journey data, ask your vendor for a historical reattribution report under your new model. Most enterprise tiers offer this. Smaller teams may need to switch to a platform with better export options.

Reality Check: When CAC Still Looks Wrong
-----------------------------------------

After reattribution, if CAC still seems out of line with historical trends, check three things. First, confirm that spend figures include all costs (ad spend, platform fees, labor if you allocate it). Second, verify that the conversion definition is the same across periods (a "lead" in Q1 might be defined differently in Q3). Third, check for seasonal shifts or campaign changes that actually did move CAC, independent of attribution.

What to Do Next
---------------

Export your raw conversion data this week and reattribute one channel as a test. Compare the result to your platform's native report. If the numbers differ by more than 5–10%, you have found model noise worth tracking. Once you confirm the process works, apply it to all channels and lock your baseline for the year.

FAQs
----

**Q: Can I use my platform's built-in multi-touch report to reattribute?** A: Not directly. Most platforms apply their model automatically and do not let you export the raw journey and reapply a different model. You need the raw data to reattribute yourself.

**Q: How far back should I reattribute?** A: At minimum, one full year before the model change. This gives you four quarters of baseline data to compare against. If you changed models last quarter, reattribute Q1–Q4 of the prior year.

**Q: Is reattribution accurate if I have offline touchpoints?** A: Only for digital channels. Offline touchpoints (phone calls, store visits, events) are rarely captured in your analytics journey. Reattribute digital conversions only and track offline CAC in a separate system.

**Q: What if my CAC looks the same before and after reattribution?** A: That means the model change had minimal impact on channel credit. Your channels may have similar journey patterns, or the new model distributes credit evenly. Document this finding and use it to decide whether future model changes are worth the effort.

People Also Ask
---------------

**Q: How do I choose between last-click and multi-touch attribution?** A: Last-click is simpler to audit and reattribute. Multi-touch (time-decay, position-based) is more accurate for longer customer journeys. If your average journey is 2–3 touches, last-click is fine. If it is 5+ touches, multi-touch is worth the complexity.

**Q: Should I reattribute every quarter or stick with one model?** A: Stick with one model for 12 months. Reattribution is labor-intensive. If you must change, do it once per year, not quarterly. The stability is worth more than the incremental accuracy.

**Q: What is a good CAC benchmark across channels?** A: This depends on your LTV (customer lifetime value). As a rule of thumb, CAC should be no more than 30–50% of LTV. If your LTV is $1,000, target CAC under $300–500. This varies by industry; SaaS companies often run tighter CAC/LTV ratios than e-commerce.

**Q: Can I use Google Analytics 4's default attribution model for CAC?** A: GA4's default is data-driven attribution (a machine-learning model). It is accurate but not easy to reattribute manually. For CAC calculations, use last-click or time-decay instead, which you can replicate in a spreadsheet.

**Q: How do I explain CAC changes to stakeholders?** A: Show them the split: X% of the change is due to the attribution model shift, Y% is due to actual performance change. This prevents false alarms and keeps focus on real optimization.

**Q: What if I do not have access to raw journey data?** A: Request it from your analytics admin or platform vendor. If unavailable, use your platform's multi-touch report as a proxy and note that reattribution is approximate. Plan to switch platforms if raw data access is critical to your CAC tracking.

**Q: Should I reattribute conversions from paid and organic channels the same way?** A: Yes. Use the same model for all channels. The model logic (last-click, time-decay, etc.) is channel-agnostic. Applying different models to different channels creates inconsistency and makes CAC comparisons meaningless.

**Q: How do I handle direct traffic in reattribution?** A: Direct traffic is often a placeholder for missing data (no referrer). In reattribution, credit the previous known touchpoint, not direct. This reduces false attribution to direct and improves accuracy for the real channel.

**Q: What is the difference between CAC and cost-per-lead?** A: CAC is the cost to acquire a customer (a paying user or closed deal). Cost-per-lead is the cost to acquire a lead (a prospect who expressed interest). CAC is always higher because not all leads convert to customers. Use the same reattribution logic for both, but track them separately.
