[{"@context":"https:\/\/schema.org\/","@type":"BlogPosting","@id":"https:\/\/mytechanalyst.net\/blog\/data-quality-problems-start-guessing\/#BlogPosting","mainEntityOfPage":"https:\/\/mytechanalyst.net\/blog\/data-quality-problems-start-guessing\/","headline":"Data Quality Problems? Start Guessing!","name":"Data Quality Problems? Start Guessing!","description":"Last updated: April 2026 Are You Guessing When You Don’t Realize It? There is an old cartoon that made the rounds in business schools for years. A senior executive stands at a whiteboard in front of his leadership team. The caption reads: “We can’t figure out which half of our marketing budget is wasted, so […]","datePublished":"2026-04-25","dateModified":"2026-04-25","author":{"@type":"Person","@id":"https:\/\/mytechanalyst.net\/blog\/author\/sselip\/#Person","name":"sselip","url":"https:\/\/mytechanalyst.net\/blog\/author\/sselip\/","identifier":1,"image":{"@type":"ImageObject","@id":"https:\/\/mytechanalyst.net\/wp-content\/wphb-cache\/gravatar\/3f8\/3f89a45b84faa5ac60f91a01db85e3f4x96.jpg","url":"https:\/\/mytechanalyst.net\/wp-content\/wphb-cache\/gravatar\/3f8\/3f89a45b84faa5ac60f91a01db85e3f4x96.jpg","height":96,"width":96}},"publisher":{"@type":"Organization","name":"Principal Consulting, LLC","logo":{"@type":"ImageObject","@id":"https:\/\/mytechanalyst.net\/wp-content\/uploads\/2019\/06\/150x40-NewPCLLCLogo.png","url":"https:\/\/mytechanalyst.net\/wp-content\/uploads\/2019\/06\/150x40-NewPCLLCLogo.png","width":600,"height":60}},"image":{"@type":"ImageObject","@id":"https:\/\/mytechanalyst.net\/wp-content\/uploads\/2019\/07\/1641x625-bw-SchemaDepiction.jpg","url":"https:\/\/mytechanalyst.net\/wp-content\/uploads\/2019\/07\/1641x625-bw-SchemaDepiction.jpg","width":100,"height":100},"url":"https:\/\/mytechanalyst.net\/blog\/data-quality-problems-start-guessing\/","about":["Assessing","data quality","Listening","Planning"],"wordCount":940,"articleBody":"Last updated: April 2026Are You Guessing When You Don’t Realize It?There is an old cartoon that made the rounds in business schools for years. A senior executive stands at a whiteboard in front of his leadership team. The caption reads: “We can’t figure out which half of our marketing budget is wasted, so we’re just going to cut it all and guess.”It is funny because it is true. And it is true far more often than most organizations are comfortable admitting.When your data is unreliable, every decision that depends on it becomes a guess. The problem is that most organizations do not frame it that way. They frame it as analysis. They build dashboards, run reports, schedule review meetings, and present findings with confidence. But if the underlying data is wrong, incomplete, or inconsistent, all of that structure is guesswork dressed up in professional clothing.Daniel Kahneman’s framework for thinking about judgment and decision-making is useful here. Kahneman distinguishes between two modes of thinking: fast, intuitive System 1 thinking, and slower, deliberate System 2 thinking. Good decisions under uncertainty require System 2. But when data is unreliable, System 1 fills the gap \u2014 often without anyone noticing. You think you are analyzing. You are actually rationalizing.The antidote is not better intuition. It is better data. And the first step toward better data is understanding where the bias is coming from.Personal BiasThe most immediate source of bias is the individual decision-maker. Everyone carries assumptions, prior experiences, and mental models that shape how they interpret data \u2014 especially ambiguous data.A marketing manager who built her career on email campaigns will tend to interpret mixed performance data in ways that favor email. A consultant who has seen a particular pattern across multiple clients will tend to see that same pattern in new situations, even when the evidence is thin. This is not dishonesty. It is how human cognition works.Personal bias is most dangerous when data quality is low, because low-quality data is inherently ambiguous. Ambiguous data does not constrain interpretation \u2014 it enables it. When the numbers could mean several things, they tend to mean whatever the person looking at them already believed.The practical defense against personal bias is structured interpretation. Before drawing a conclusion from a data set, document your hypothesis and the specific evidence that would confirm or disconfirm it. Then look for disconfirming evidence deliberately. This sounds simple. It is surprisingly difficult to do consistently.Team BiasIndividual bias is compounded when teams discuss data together. The organizational behavior research on this is consistent and sobering: group deliberation tends to amplify the prior beliefs of the most influential members of the group, not converge on the most accurate interpretation of the evidence.This happens through a well-documented mechanism called information cascades. Early speakers in a meeting signal their interpretation of the data. Later speakers, consciously or not, adjust their own stated views toward the emerging consensus. Dissenting interpretations get filtered out not because they are wrong but because voicing disagreement carries social cost.The result is that a team reviewing the same low-quality data set will frequently converge on a confident but incorrect conclusion faster than an individual working alone, because the social dynamics of the meeting suppress the uncertainty that should be present.The practical defense is structured dissent. Assign someone explicitly to argue against the leading interpretation before the group converges. Collect individual assessments in writing before group discussion begins. Make it organizationally safe to say “I am not sure the data supports that conclusion.”Organizational BiasThe deepest and most persistent source of bias is organizational. Every organization develops, over time, a set of shared assumptions about how its market works, what its customers value, and what kinds of decisions have worked in the past. These assumptions become embedded in processes, metrics, and incentive structures.When data quality is poor, organizational bias fills the interpretation gap with those embedded assumptions. The company that has always competed on price sees price sensitivity in ambiguous customer data. The organization that measures success by revenue growth interprets mixed signals as evidence that more growth investment is needed. The assumptions are not examined because they are not visible as assumptions \u2014 they are just the way things work.Organizational bias is the hardest to address because it requires stepping outside the interpretive framework that the organization uses to make sense of everything. It typically takes either a significant external disruption or deliberate leadership commitment to surface and examine embedded assumptions.The practical starting point is a regular audit of your key metrics. What are you measuring? Why those things and not others? What assumptions about your business are baked into your measurement framework? When did you last examine those assumptions against external evidence?The Real Cost of GuessingThe cost of data-driven guessing is not just bad individual decisions. It is the cumulative erosion of organizational confidence in data itself. When decisions made on the basis of data analysis consistently produce poor outcomes, people stop trusting the analysis. They start making decisions on gut feel and then using the data to justify what they already decided.At that point, you are not just guessing. You are investing resources in a data infrastructure that provides the appearance of rigor without the substance.The way out is not a new dashboard. It is a serious commitment to data quality at the source, combined with structured processes for interpretation that acknowledge and counteract the three sources of bias.If you are not sure where your data quality problems are hiding,\u00a0that is exactly the kind of diagnostic we do."},{"@context":"https:\/\/schema.org\/","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Blog","item":"https:\/\/mytechanalyst.net\/blog\/#breadcrumbitem"},{"@type":"ListItem","position":2,"name":"Data Quality Problems? Start Guessing!","item":"https:\/\/mytechanalyst.net\/blog\/data-quality-problems-start-guessing\/#breadcrumbitem"}]}]