Trustworthiness trumps structure in the new era of data quality. Our webinar with Yali Sassoon from Snowplow Analytics revealed that organisations obsess over clean schemas and proper formatting while neglecting the more crucial factor: whether decision-makers actually believe and use the data. AI applications can often work with less structured data, but they absolutely require data that humans and systems can trust implicitly. The true measure of data quality isn't technical perfection—it's whether your dashboards drive real business decisions.

Tasman co-founder Thomas in't Veld & Snowplow CTO Yali Sassoon discuss data quality and trustworthiness.

Data isn't collected; it's manufactured. This fundamental shift in thinking transforms how successful organisations approach quality assurance. Rather than treating data as something that naturally exists to be gathered, smart teams recognise it as a complex product with a multi-stage production process. Each stage requires clear ownership and appropriate quality controls. When quality issues arise, the question becomes not "Why is this dashboard wrong?" but "Which stage in our production line is causing the defect?" This manufacturing mindset enables targeted improvements rather than endless firefighting.

Technical brilliance without business context creates sophisticated systems nobody uses. The most effective data teams either hire directly for domain expertise—bringing marketers, product specialists, or finance experts into the data function—or invest heavily in building cross-organisational relationships. This expertise bridges the vital gap between technical capabilities and business needs, ensuring data products deliver genuine insights rather than technically accurate but meaningless metrics. A performance marketer within your data team might be more valuable than another engineer.

Remember This:

  • Trust trumps structure: Prioritise data stakeholders will actually use for decision-making over perfectly formatted tables
  • Measure adoption: Dashboard usage reveals data quality more clearly than technical metrics
  • Prioritise ruthlessly: Apply different governance standards based on business impact
  • Think like manufacturers: Implement quality controls at every stage of your data production process
  • Value domain expertise: Technical skills mean little without the business context to apply them
  • Start small: Begin with one high-impact use case and build momentum through tangible success
  • Define metrics collaboratively: Business terms like "lead" or "conversion" need shared definitions

From chaos to clarity in twelve weeks—this isn't just theoretical advice. When we implemented these principles with Paired, they transformed their marketing performance by building a centralised data model with proper governance. This gave them unprecedented visibility into campaign effectiveness, allowing them to quickly reallocate spend from underperforming channels. The result wasn't just cleaner data but substantially improved return on ad spend and revenue growth. This real-world impact demonstrates why quality data isn't a technical nicety but a genuine business advantage.