For decades, marketers have equated “data quality” with cleanliness – merge/purge routines, NCOA updates, and formatting fixes. But in a world that’s defined by real-time targeting, cross-platform engagement, and customer identities that refuse to sit still, that definition no longer holds up.
Michael D. Fisher, CEO of Allant Group, thinks it’s time for a new standard – one grounded not in hygiene, but in composability. In this model, identity isn’t pulled from a single curated file. It’s orchestrated across sources, validated in context, and constantly refined. The goal? Decision-ready data that’s built for action – not just analysis.
In the conversation that follows, Fisher outlines where traditional models fall short, what a modern definition of data quality should look like, and how a composable foundation enables marketers – and enterprises – to act faster, smarter, and with more confidence.
Legacy hygiene still dominates how most marketers think about “data quality.” What are the core flaws in that model—and why has it remained so persistent?
Fisher:
The process hasn’t changed in 35 years. Most data providers still rely on a single compiled file to define identity. They treat that file as a source of truth, even though each one uses different logic to define individuals, households, and addresses. But we know from experience: no one source is sufficient. If a single file were enough, then adding new sources wouldn’t improve performance. But every time we add another source, we see incremental lift in accuracy and reach. That tells you everything. The legacy merge/purge mentality assumes a source is complete because it’s been curated. But that thinking is outdated. Today, we need to harmonize across sources, not pick one and pretend it’s perfect.
If so many data providers are using outdated models, why hasn’t the industry evolved beyond the single-source paradigm?
Fisher:
The reason marketers and even data providers still follow legacy models is simple: there hasn’t been an easy way to operationalize a multi-source strategy. Data providers don’t collaborate with each other. They protect their files.
Allant takes a different approach. We don’t own the data. We orchestrate it. That means we’re not beholden to one source. We use all available sources to build the most complete, accurate, and analytically sound profile of each individual. We map each person to their household using rules that differ by source, and then we build our own logic to ensure accuracy. That’s what enables the composability our clients rely on. We can assemble and enrich each ID with data from across the ecosystem.
You’ve said that relying on a flawed identity spine creates a cascade of downstream problems. What’s really at stake when brands don’t modernize?
Fisher:
If brands stick with single-source identity graphs and outdated measures of quality, they’re going to target the wrong people—or miss entire segments. And worse, every decision they make downstream is compromised. If your identity spine is flawed, your enrichment data will be flawed. Your first-party data modeling will be skewed. Garbage in, garbage out. It snowballs quickly. Starting with a rich, multi-source foundation isn’t a nice-to-have. It’s essential.
In practical terms, how does Allant redefine “data quality” in a composable architecture?
Fisher:
Data quality today means being open. You have to be comfortable working across formats, across sources. There’s no longer a single source of truth.
What we’ve built at Allant is a 360-degree view that’s actually real – not a limited slice of the market. AMP+ brings everything together. We assemble and link the data, validate it, standardize it, and then provide advanced analytics that allow marketers to begin their work at floor 25, not floor one. Our goal is decision-readiness. The data is enriched, correlated, prioritized and ready for action. Marketers don’t have to wade through billions of rows. We surface the most valuable signals so they can go straight to modeling and activation.
Can you walk us through how a persistent, multi-source identity is actually built—and why it matters more now than ever before?
Fisher:
We don’t validate identity using one file. We validate it across five. We constantly compare and contrast attributes to find the best representation of a person. Let’s say you’re targeting people who buy high-performance cars and do their own maintenance. One source might have decent data on that audience. But when we check another source, we fill in missing details. Maybe source three disagrees; so we go deeper. If three out of five agree, we know what’s likely true. That’s the orchestration we bring. And the ID we build from it? It’s persistent. It supports marketing, customer service, risk management, underwriting—any function that needs a trusted view of the individual.
You’re making the case that this approach isn’t just a marketing advantage. Where else does composable data show up as a competitive edge?
Fisher:
This isn’t just a marketing issue. Customer service, media buying, fraud detection – they all benefit from high-fidelity data. We key the data once and make it available across the enterprise. The same identity spine runs through every application. That creates consistency. It eliminates fragmentation. And it turns identity from a tactical asset into a strategic one.
What’s a concrete example where your composability model uncovered something other providers would have missed?
Fisher:
We recently worked with a major credit card issuer and uncovered a high-net-worth, under-25-year-old audience segment that didn’t exist in any of the major compilers. These consumers didn’t fill out forms. They didn’t use traditional credit. They operated in digital-only environments. Because we compose across compilers and digital signal data, we found them, and they’re incredibly valuable. Single-source strategies would’ve missed them entirely.
From where you sit, what does the next phase of data strategy look like—and how should brands prepare?
Fisher:
The idea that one file can provide all the answers? That’s going away. Brands are realizing they need composability. They need orchestration. And they need data that’s ready to activate across every channel, not just the one it came from. And the shift doesn’t mean ripping out your martech stack. Keep the platforms you love. Just make sure the data feeding those platforms is enriched, real-time, and composed for the task at hand.
If you could erase one myth that still haunts the data space, what would it be—and why?
Fisher:
If I could erase one myth, it’s that data must be purged to be useful. That myth assumes an engineer knows the analytical value of a data point the moment it lands. They don’t. We keep everything. Then we analyze. We experiment. We model. That’s how you unlock value and innovate. We’re building what I call a data innovation factory. You bring in new ingredients, try new recipes, iterate continuously. There’s no penalty for asking more of your data.
Closing thought.
In a time when customer identities span channels, devices, and databases, the old question – “Is your data clean?” – no longer feels relevant. The better question might be:
Is your data ready to move your business forward?
Fisher and his team at Allant are betting the future depends on answering that with a confident yes … every single time.
