Treat Attribution Like a Toolkit, Not a Theology with Drew Smith

Why Picking “The Right” Attribution Model Is the Wrong Question With Drew Smith

Introduction

Most organizations get stuck on attribution for the same reason: they’re trying to pick one model to represent the whole business. Drew Smith, founder and CEO of Attributa and host of the Attribution Nation podcast, has spent three years telling clients that’s the wrong question entirely.

On this episode of Ops in Motion, Drew joined Kawal to talk through why attribution models should be chosen per report rather than per company, why adoption, not integration, is the real bottleneck in most martech stacks, and where marketing operations is actually headed as AI gets folded into the role.

Who should read this: marketing operations leaders stuck in attribution model debates, or managing a tech stack where tools get bought but never really used.

Key takeaways

• There’s no single right attribution model for an organization. Different reports need different models, the same way different bolts need different wrenches.
• Adoption, not integration or consolidation, is usually the biggest reason a martech investment fails to deliver value.
• Tool consolidation stalls because it’s political, not technical. People don’t want to give up the tool they championed, even when a usage audit says they should.
• AI isn’t going to replace marketing operations, because it can’t make the judgment call on which process or model actually fits a given situation. The bigger risk is not integrating AI into the role at all.

Stop trying to pick the right attribution model. Pick the right one per report

Drew’s clients spend a lot of energy debating first touch, last touch, and multi-touch models as if the organization has to commit to one. His answer is that there isn’t a right model for a business, only the right model for a specific report.

He compares it to a wrench set: you don’t buy six wrenches and throw away five because you found a favorite. Each one fits a different bolt. Attribution models work the same way. Once a team stops trying to select one model to represent the whole organization and instead picks the right model for each individual report, the internal arguments mostly disappear, and teams can actually start building reports instead of debating which one is correct.

Adoption, not integration, is why most martech investments fail

Asked what’s hardest about managing a modern martech stack, adoption, integration, or consolidation, Kawal, co-founder of Digital DI Consultants and host of the episode, named adoption as the consistent front-runner. Buying and implementing a tool is the easy part. Getting a team to actually embed it into daily process, understand why they’re using it, and see measurable value over time is where most investments quietly fail. A tool that overlaps with something the team is already using, or doesn’t fit their existing workflow, ends up sitting unused no matter how good it is on paper.

Integration is a close second. With enough tools connected, a single misfiring webhook or duplicate field mapping can break reporting or lead routing in seconds. Kawal’s team handles this with a minimum viable integration approach: connect only what’s genuinely needed first, test those workflows, then expand, rather than plugging everything in at once and hoping nothing breaks.

Consolidation is the most political of the three. Everyone wants fewer tools in the stack, but no one wants to give up the one they’ve become the internal champion for. Kawal’s team treats consolidation like a product sunset: run a usage audit, score each tool’s actual value, build a migration plan, then handle the communication and training needed to move a team off a tool they’re attached to.

Drew sees the same adoption gap from the attribution side specifically. Teams onboard an attribution platform, get it set up correctly, and then stall at the exact same question: now that we have this data, what do we actually do with it?

“Adoption isn’t just about buying and implementing the tool. You have to embed it into the process, get the team aligned on how and why to use it, and measure the value over time.”

— Kawal

AI isn’t replacing marketing operations. Not integrating it is the actual risk

Drew doesn’t share the concern that AI is coming for marketing operations roles. His reasoning is specific: AI can’t yet make the judgment call on which strategic process or attribution model actually fits a given business situation. That kind of decision still requires a person who understands the context.

What he does expect is a widening gap between marketing ops professionals who fold AI into their day-to-day work and those who don’t. He pointed to the emerging go-to-market engineer concept, a role built around integrating AI and automation directly into go-to-market execution, as an early signal of where the discipline is heading. His prediction is that marketing ops professionals who get comfortable with AI early are the ones most likely to move into broader marketing leadership roles over the next several years, precisely because they’ll already understand how to apply it. If you’re trying to understand where that role is headed, the responsibilities of a GTM engineer are a useful starting point.

Kawal’s addition is worth sitting with: AI can execute a repeatable process once it’s told what to do, but it isn’t going to decide on its own which attribution model fits a given report or how a workflow should actually be built. That judgment still sits with a person.

Action items

FAQs

Talk to Digital DI Consultants

If your team is stuck picking one attribution model, or your martech stack has tools nobody actually uses, talk to our team about your setup.