You know that moment in the boardroom when your CEO leans back and asks, “So which of these ads actually made us money?”
Yeah. That moment.
For the past decade, we’ve been sold a story: multi-touch attribution will finally answer that question. We’ve graduated from the Stone Age of last-click attribution to sophisticated algorithmic models that promise to decode the customer journey. We’ve invested in platforms, hired analysts, built dashboards that would make NASA jealous.
And yet, here’s what nobody wants to admit at the marketing conference after-parties: we’re still mostly guessing.
Worse than that, our increasingly complex attribution models aren’t just failing to give us answers-they’re actively leading us toward terrible decisions. And the brands winning right now? They’ve figured out something the rest of us are missing.
The Fundamental Flaw Nobody Talks About
Every multi-touch attribution model-whether it’s linear, time-decay, U-shaped, or some proprietary algorithmic black box-rests on the same shaky foundation. They all assume the customer journey is a series of discrete, trackable touchpoints moving in a mostly linear direction toward purchase.
That assumption made sense in 2014. In 2024? It’s fantasy.
Here’s the distinction that changes everything: attribution models distribute credit, but what we actually need is to understand influence. Those aren’t the same thing, and confusing them is where the trouble starts.
Attribution tells you what people touched. Influence tells you what actually changed their mind. One is correlation. The other is causation. And we’ve built entire marketing departments that can’t tell the difference.
Three Reasons Your Attribution Model Is Lying to You
The Dark Social Black Hole
Here’s a stat that should terrify anyone relying heavily on attribution data: 84% of sharing now happens through what researchers call “dark social”-private messages, texts, Slack channels, closed Discord servers, email forwards.
Your attribution model sees exactly none of this.
Think about how decisions actually get made in your own life. Your business partner texts you a screenshot of an ad. A Reddit thread convinces you a product is worth trying. Someone in your company Slack shares an Instagram post that shifts how you think about a vendor. These are often the most influential moments in the buyer journey, precisely because they come from trusted sources instead of the brand itself.
But to your attribution model? They don’t exist. It’s not that these touchpoints are undercounted-the model is architecturally incapable of seeing them. You’re making budget decisions based on data that systematically ignores the most persuasive moments in the modern customer journey.
We’re Measuring the Wrong Things
I’ve noticed a pattern across dozens of client accounts: the things that are easiest to measure are usually the least influential. And the things that actually move the needle? Nearly impossible to track properly.
Consider these two scenarios:
- A display ad impression that someone scrolled past in 0.3 seconds: Perfectly measurable, basically zero impact
- A 40-minute podcast episode where the host authentically raves about your product: Massive impact, attribution nightmare
We’ve optimized entire marketing organizations around what’s trackable instead of what’s meaningful. Multi-touch attribution makes this worse by creating a veneer of precision-your dashboard confidently states that “email contributed 23.4% to this conversion”-while completely missing that the prospect’s colleague recommended you after seeing your TikTok ad at a conference.
The data looks scientific. The decisions based on it? Often completely divorced from reality.
The Cross-Device Illusion
Ad platforms will swear they’ve “solved” cross-device tracking. What they mean is they’ve gotten better at guessing when different devices belong to the same person through probabilistic matching and device graphs.
But here’s how real customer journeys actually work now:
- Initial research on mobile during the morning commute
- Deeper dive on desktop at work (probably behind a VPN)
- Discussion on a partner’s tablet over lunch
- Decision-making on a home desktop three days later
- Actual purchase on mobile while running errands on Saturday
Every one of those transitions is a potential attribution break. Every context switch might create a “new” user in your system. Your multi-touch model is trying to connect dots that may not even be visible to it, often creating connections between the wrong dots while missing critical moments entirely.
And we’re basing million-dollar budget decisions on this.
The Dangerous Feedback Loop
Bad attribution doesn’t just give you wrong numbers-it warps your entire strategy over time.
Here’s the insidious part: when your attribution model systematically undervalues brand awareness campaigns (they’re hard to track) and overvalues retargeting (it’s easy to measure), you don’t just get inaccurate reports. You get a completely skewed understanding of what drives growth.
I’ve watched this movie play out more times than I can count. A company starts relying heavily on attribution data for budget allocation. Gradually, imperceptibly, budgets shift toward the measurable-more retargeting, more bottom-funnel tactics, more direct response. Meanwhile, the brand-building work that actually creates demand gets starved for resources.
The attribution model creates a self-reinforcing loop:
- Measurable channels get credit
- They receive more budget
- They generate more measurable conversions
- They get even more credit
- The cycle continues
Meanwhile, the YouTube pre-roll campaign that made people prefer your brand? The TikTok ads that made you culturally relevant? They get systematically defunded because their influence is harder to measure.
The cruel irony: the more religiously you optimize based on attribution data, the worse your actual performance becomes over time.
The $680,000 Wake-Up Call
Let me tell you about an e-commerce brand spending roughly $2 million annually across Facebook and Instagram. Their multi-touch attribution model showed Instagram Stories “contributing” to 34% of conversions. Based on that data, Stories was one of their star performers.
Then we ran a holdout test-we stopped serving Instagram Stories ads to a controlled segment of users.
What happened? Almost nothing. The conversions simply shifted to other touchpoints with virtually no drop in total conversion volume. The Instagram Stories ads weren’t driving incremental conversions; they were just visible in the path. The attribution model saw correlation and confidently assigned causation.
The cost of that misunderstanding? Nearly $680,000 in wasted annual spend.
This is the difference between attribution and incrementality. Attribution asks: “What did people interact with before buying?” Incrementality asks: “What caused purchases that wouldn’t have happened otherwise?”
Most attribution models can’t tell the difference. So you end up maintaining-and expanding-campaigns that add cost but not value.
What the Smart Money Is Doing
Trading Precision for Truth
The most sophisticated marketers I know have stopped chasing perfect attribution. Instead, they’re building what I call “triangulated understanding”-combining multiple imperfect data sources to develop directional confidence:
- Post-purchase surveys: Simple questions like “What made you decide to buy today?” provide qualitative insights that numbers never capture
- Holdout testing: Systematically pausing channels for controlled segments to measure true incrementality
- Market-level analysis: Using geo-testing to understand aggregate impact instead of trying to track individual journeys
- Branded search monitoring: Tracking branded search volume as a proxy for upstream marketing effectiveness
None of these is perfect. Together, they paint a more accurate picture than any attribution model alone.
Measuring Influence, Not Just Interaction
The cutting-edge teams are moving beyond “what did they touch” toward “what changed their likelihood to convert.”
They’re measuring shifts in conversion probability after exposure. They’re using synthetic control methods to estimate what would have happened without a given campaign. They recognize that some touchpoints create lasting preference changes while others just capture intent that already existed.
When we work with clients at Sagum across platforms like TikTok, YouTube, and Pinterest, we’ve shifted our reporting. Instead of saying “This ad contributed X% to conversions,” we focus on insights: “This creative test revealed your audience responds to authenticity over polish” or “This campaign shifted brand consideration scores in a way that historically predicts sustained revenue growth.”
It’s less precise. It’s more useful.
Making Peace with Uncertainty
The best marketers I know have accepted something radical: some value is inherently unmeasurable, and that’s okay.
They allocate budgets based on frameworks that acknowledge reality:
- Long-term brand building deserves investment even without direct attribution
- Marketing mix modeling and econometric approaches complement (not replace) digital attribution
- Qualitative signals matter as much as quantitative metrics
- Strategic conviction sometimes outweighs data certainty
This isn’t anti-analytics. It’s pro-reality.
A Portfolio Approach to Budget Allocation
Here’s the framework that’s working for growth-focused brands right now:
60% of budget-The Foundation: Channels with clear, measurable direct response performance. Search ads capturing existing demand. Retargeting converting warm traffic. Email nurturing known prospects. This is your baseline revenue engine.
30% of budget-The Growth Engine: Channels requiring multi-touch analysis and incremental testing. Social ads. Display campaigns. YouTube pre-roll. This builds the pipeline that your direct response captures. Attribution helps here, but incrementality testing validates it.
10% of budget-The Future: Unmeasurable or experimental tactics that build long-term equity. Brand campaigns. Cultural moments. New platform testing. Creative that expands your addressable market. This is where you accept that measurement will be imperfect.
Don’t let an attribution model dictate what belongs in each bucket. Let business results, incrementality testing, market dynamics, and strategic judgment guide those decisions. Use attribution as supporting context, not as the decision-maker.
The Real Competitive Threat
Here’s what actually keeps me up at night: while we’re obsessing over attribution models and arguing about credit distribution, there are brands out there building powerful marketing engines despite attribution’s limitations.
They’re creating scroll-stopping TikTok content without perfect attribution to revenue. They’re investing in brand building that won’t show clear ROI for 18 months. They’re testing Pinterest and new platforms before the measurement infrastructure exists to “prove” it works.
They’re winning not because they measure better than you. They’re winning because they move faster and think broader while their attribution-obsessed competitors are paralyzed by the need for measurement certainty before action.
Your real competition isn’t the brand with the better attribution model. It’s the brand willing to make strategic bets based on directional confidence rather than false precision.
Five Questions to Ask Right Now
If you’re running significant ad spend across multiple platforms, it’s time for an attribution audit:
1. Are we rejecting potentially profitable strategies because we can’t measure them perfectly? This is the silent killer. Opportunities lost not because they won’t work, but because they won’t generate the data that satisfies your measurement requirements.
2. Has our channel mix shifted dramatically toward easily-attributable tactics over the past two years? If you’re spending substantially more on retargeting and bottom-funnel tactics than you were 24 months ago, attribution might be creating strategic drift.
3. Are we confusing correlation with causation? Just because something appears in the conversion path doesn’t mean it influenced the decision. This is attribution’s original sin.
4. Do we have budget deliberately allocated to unmeasurable brand building? If not, you’re optimizing entirely for the short term. That works until it doesn’t.
5. When did we last run a true incrementality test? Holdout tests are uncomfortable. They require patience. They’re also the only way to validate what your attribution model suggests.
The Truth About Multi-Touch Attribution
Here’s the paradox we need to embrace:
Multi-touch attribution is dramatically better than last-click. It’s also significantly worse than we pretend. It’s absolutely necessary for modern marketing. And it’s critically insufficient for strategic decision-making.
All of these things are true simultaneously.
The marketers who win over the next decade won’t be those with the most sophisticated attribution models. They’ll be those who use attribution judiciously-as one input among many-while building the judgment, intuition, and testing muscle to make decisions when data falls short or points in uncertain directions.
Because here’s the reality nobody wants to say out loud at the conference keynote: if you’re waiting for perfect attribution before making strategic moves, you’re already losing to competitors who are comfortable with ambiguity.
The Question That Actually Matters
Stop asking “How do we attribute this conversion?”
Start asking “How do we build a marketing system that reliably drives business growth, whether we can measure every piece or not?”
The goal isn’t to measure everything. The goal is to build profitable, sustainable growth engines. Sometimes those require measurement. Sometimes they require conviction. Often they require both.
Multi-touch attribution is a tool-useful in the right context, dangerous when over-relied upon, and never a substitute for strategic thinking.
Treat it accordingly, and you might just stop leaving millions on the table while your competitors pull ahead.
