Rethinking attribution: from models to marketing judgment
"Data doesn’t make decisions. Marketers do."
Attribution is one of the most widely discussed, and often misunderstood, concepts in marketing. It’s tempting to treat it as a definitive way to answer the question: what worked? But in reality, attribution isn’t just about assigning credit. It’s about helping marketers understand outcomes in context, compare alternatives, and make better decisions over time.
Too often, attribution is treated as the final word. A number shows up on a dashboard, and action follows. But effective attribution is not about handing off decisions to a formula. It’s about using structured input to support experienced judgment.
Here are four principles to help marketers and analysts approach attribution as a strategic tool, without overpromising precision or losing sight of what makes marketing work in the first place.
Joram’s four ways to make attribution actually useful
1. Attribution starts with the right data foundation
In most marketing teams, data comes from multiple sources, each with its own view of reality. Meta reports conversions based on its own attribution settings. Google does the same. CRM platforms track customer behavior independently. None of these systems talk to each other, yet all claim to tell you what works.
This fragmented measurement environment leads to confusion and inconsistency. Decisions get made based on partial signals or conflicting stories, and true performance becomes hard to see.
To get value from attribution, your data needs to be brought together. Not just technically integrated, but aligned in scope and logic. Attribution only becomes meaningful when it's based on a shared view of the customer journey; across platforms, channels, and time.
From there, more advanced modeling can add structure. Tools like Marketing Mix Modeling (MMM) and Unified Marketing Measurement (UMM) help connect the dots between click-based activity and broader marketing effects. Platforms like Billy Grace offer this integration by design, but the principle applies universally: attribution is only as strong as the foundation it rests on.
2. Attribution is not just about clicks
Much of attribution today relies on digital traces: clicks, views, conversions. MTA models distribute credit across touchpoints based on this data, offering a more nuanced view than last-click or first-click models.
But not all influence is clickable. A podcast, a billboard, a mention on social media, a conversation with a colleague, these shape decisions too. If your attribution framework doesn’t account for these, you're not seeing the full picture.
This is where approaches like MMM and UMM are valuable. MMM estimates the contribution of all marketing inputs, even those that don’t leave a digital footprint. UMM brings this together with MTA to offer a more complete and flexible lens.
Try this: If you’re working with both MTA and UMM, compare how each explains performance for the same campaign. What’s consistent? What’s different? Those differences highlight the role of indirect influence, and remind you why a single number rarely tells the whole story.
3. Your knowledge completes the picture
Attribution frameworks are built on observed patterns. But marketers work with plans, shifts, and assumptions, many of which happen faster than attribution systems can track.
You know when a campaign changes direction mid-flight. You know when you’re testing a new channel, running a limited-time offer, or pausing creative due to external events. Attribution can’t always surface these nuances, especially in the moment.
That’s why attribution should be seen as input, not output. It supports thinking, it doesn’t replace it. Especially when campaigns are new or changing, your expertise provides essential context to interpret what attribution is showing (or missing).
Try this: Ask yourself: What does attribution reflect well in this case, and what does it probably overlook? Treat attribution results as one source of insight, best used in combination with your own strategic awareness.
4. Build for both short-term tactics and long-term strategy
Different attribution methods serve different time horizons. MTA excels at short-term decision-making: testing creatives, reallocating spend, and optimizing conversion paths. UMM supports long-term learning; understanding how branding, media mix, and macro factors contribute to growth over time.
In practice, most marketing teams need both. Precision in the short term is necessary for efficiency. Perspective in the long term is necessary for resilience.
The goal isn’t to choose between models, but to combine them. MTA helps you act today. UMM helps you invest for tomorrow.
Try this: Use MTA to guide platform-level decisions; such as pausing underperforming ads or testing messaging variations. Use UMM to evaluate your full channel mix and validate strategic shifts in brand investment. Over time, this two-speed approach helps align daily execution with overall business growth.
Attribution is a starting point, not a destination
Data doesn't dictate decisions, people do. Attribution is a tool that helps you see more clearly, but it doesn’t replace your experience, your goals, or your context.
Especially in a landscape where marketing channels overlap, interact, and evolve, the most valuable insights often come not from choosing the “best” model, but from combining models with interpretation.
If you’re a marketer, your role isn’t just to read attribution reports. It’s to challenge them, contextualize them, and use them as inputs in a larger strategic process.
Three moves to start today
- Seek integration. Bring your data together across platforms to support more unified analysis.
- Compare perspectives. Analyze one campaign using both MTA and UMM. Note what each view captures, and what it doesn’t.
- Debrief with context. After major campaigns, ask: What story does attribution tell, and what do I know that adds to it?
Attribution doesn’t need to be perfect to be useful. But it needs to be interpreted, and that’s where marketing judgment makes the difference.
Remember: “use data as a tool, not the rule”.
Written by Joram - on using data not just to measure marketing, but to guide it.
