Home / Blog / CRM
CRMAttribution Modelling: How to Measure the Impact of Marketing Campaigns
A practical guide to marketing attribution models — first-touch, last-touch, linear, time-decay, and data-driven — and how to pick one that actually measures campaign impact in B2B.
TL;DR
Attribution modelling assigns credit for a conversion across the touchpoints that led to it, so you can tell which campaigns actually drive revenue. No single model is 'correct' — first- and last-touch are simple but biased, multi-touch and data-driven models are fairer but demand clean CRM data. Pick the model that fits your sales cycle, and instrument your CRM so the data is trustworthy in the first place.
What is attribution modelling and why does it matter?
Attribution modelling is the method for assigning credit for a conversion across the marketing touchpoints that influenced it — so you can measure which campaigns actually generate revenue instead of guessing. In B2B, where a single deal involves six to ten touches across months and multiple stakeholders, this is the difference between funding what works and funding what merely happened to be last in line.
The stakes are budget-sized. When 44% of marketers still credit only the last touch, they systematically overfund bottom-funnel channels and starve the awareness and nurture activity that built the pipeline in the first place. Attribution modelling exists to correct that distortion.
The core attribution models
Each model splits credit differently. None is objectively “correct” — each encodes an assumption about what matters.
| Model | How it assigns credit | Best for |
|---|---|---|
| First-touch | 100% to the first interaction | Measuring top-of-funnel / demand gen |
| Last-touch | 100% to the final interaction | Simple, short sales cycles |
| Linear | Equal credit to every touch | Long cycles where all touches count |
| Time-decay | More credit to recent touches | Deals where late-stage nudges close |
| U-shaped | 40% first, 40% last, 20% middle | Valuing both discovery and closing |
| Data-driven | Algorithm weights each touch by real influence | High-volume, data-rich teams |
Single-touch models are easy but biased — first-touch ignores everything that closed the deal, last-touch ignores everything that created it. Multi-touch models are fairer but demand clean data. Data-driven attribution is the most accurate but needs enough conversion volume for the algorithm to learn from.
Why single-touch fails in B2B
The B2B reality is a committee reading a blog post, attending a webinar, comparing vendors, downloading a case study, and finally requesting a demo — over weeks or months. Credit the demo alone and you’ll conclude the case study and webinar don’t work, cut them, and watch pipeline dry up two quarters later. Single-touch attribution doesn’t just mismeasure; it actively misleads budget decisions. Teams that move to multi-touch report roughly three times better budget allocation.
How to choose your model
Match the model to your sales motion:
- Short cycle, few touches — last-touch or U-shaped is fine and cheap to run.
- Long, multi-stakeholder cycle — linear or time-decay reflects reality far better.
- High conversion volume, mature data — data-driven attribution earns its complexity.
Don’t over-engineer. A linear model you trust and act on beats a data-driven model nobody believes. Start simpler than you think you need, then add sophistication as your data matures.
Data quality is the real bottleneck
Here is the uncomfortable truth: 76% of teams say their attribution problem is data quality, not model choice. Attribution is only as good as the touchpoint data feeding it. If your CRM doesn’t capture every meaningful interaction against each contact and deal, no model can produce a trustworthy answer.
Getting this right means:
- Consistent tracking — UTMs and event capture on every campaign and channel.
- A CRM that logs the full journey — every touch tied to the contact and the deal, not scattered across disconnected tools.
- Clean, deduplicated records — merged contacts and standardized fields so touches aren’t split across ghost records.
- A closed loop to revenue — deals in the CRM mapped back to the touches that influenced them.
This is why attribution and CRM implementation are inseparable. The model is a few lines of logic; the data infrastructure that makes it trustworthy is the actual work.
Putting it into practice
Start here:
- Instrument tracking across every campaign before worrying about the model.
- Pick one multi-touch model that fits your cycle — linear is a fine default.
- Report influenced pipeline, not vanity conversions, so leadership sees revenue impact.
- Revisit quarterly — refine the model as data volume and quality grow.
Attribution isn’t about finding a mathematically perfect answer; it’s about making budget decisions from evidence instead of instinct. If your touchpoint data is too messy to trust today, that’s the place to start — a free audit will show you exactly where your CRM is dropping the journey and what to fix first.
Want this done for you?
Get a free audit →FAQ
Which attribution model is best for B2B?
For most B2B teams with multi-month, multi-stakeholder cycles, a multi-touch model — linear or time-decay — beats single-touch because it reflects the reality that six to ten touches drive a deal. Data-driven attribution is better still if you have the CRM data volume and quality to support it. Avoid last-touch alone; it systematically undervalues everything that builds pipeline.
Do I need special software for attribution?
For single-touch you can get by with your CRM and analytics. For reliable multi-touch attribution you need a CRM that captures every touchpoint against each contact and deal — which is why attribution and CRM implementation go together. The software matters less than whether your data captures the full journey.
