What dimensional inconsistency costs.

The single most important property of dimensional data costs is who pays them. The procurement analyst who typed “aws” instead of “Amazon Web Services” experiences no consequence. The cost is paid by the finance analyst three weeks later who tries to produce a spend-by-vendor report and finds that the same supplier appears under five names. The producer of the variance and the consumer of the consequence are different people on different teams on different time horizons.

This distribution is why dimensional data problems persist. Producers have no incentive to enter values consistently because they bear no cost when they do not. Consumers have no power to fix the producers because the producers belong to different teams with different management chains. The cost externalises, and externalised costs are reliably undermanaged.

Some costs are visible enough that they get measured. Three of them.

Wasted time. Analysts and engineers in organisations with significant dimensional inconsistency spend a substantial fraction of their time on data preparation rather than analysis. The commonly-reported range is thirty to fifty percent. A competitive analysis that should take two days takes five because three are spent reconciling vendor names and segment labels. An engineer maintains CASE WHEN statements as ongoing overhead. A data scientist hired for predictive modelling spends a quarter of their time cleaning categorical features.

Hidden spend. Vendors negotiated under variant names appear in spending reports as multiple suppliers, each below the threshold for strategic review. When the variants are consolidated, total spend turns out to be larger than any single line suggested, and the consolidated total qualifies for volume discounts and contract renegotiation that the fragmented view did not surface. Procurement teams that have done this consolidation consistently report uncovering between five and fifteen percent of total addressable spend that was previously invisible.

Compliance and audit risk. Regulatory reporting requires defensible counts and categorisations. A regulator asking how many active contracts of a specific type exist cannot get a defensible answer until the four label variants of that type are reconciled. An audit finding triggers remediation under deadline pressure, at premium cost, on work that has been deferred for years.

These three costs are visible because they can be measured: time, money, regulatory findings. They are not the largest costs.

The larger costs are invisible because the harm is several steps removed from the dimensional data that caused it, and the chain is rarely traced.

Wrong reports and the decisions made on them. A report that aggregates by an inconsistent dimension produces wrong numbers. The wrong numbers feed decisions. The decisions produce outcomes. The outcomes get attributed to execution failures, market conditions, or strategy errors. The dimensional data that caused the original wrongness is never identified as the root cause.

Blocked analyses. A cross-functional analysis that would require reconciling segments across four systems takes three weeks of preparation. The analysis does not get started, because the prep time outweighs the perceived value. The insight that would have been produced is never produced. The decision that would have been informed by the insight is made on intuition.

Degraded models. Models trained on inconsistent dimensions produce weaker predictions. Forty-seven variants of ten real industries dilute the signal. Teams experiment with algorithms and hyperparameters while the actual lever is the data. The model underperforms, the underperformance is attributed to the algorithm, and the dimensional cleanup that would have produced a step-change in performance never happens.

Erosion of trust. When data products are unreliable, the data function loses credibility. Stakeholders default to intuition. Personal spreadsheets proliferate. The data team is invited to fewer strategic conversations. Trust recovery, once dimensional remediation is complete, takes longer than the remediation itself, often by a factor of two or three.

The categories above are not parallel; they interact. A wrong report produces a wrong decision, which compounds the cost of the wasted preparation time, which contributes to the eroded trust that makes the next investment harder to justify. Most organisations underestimate the total because they measure only the visible categories and miss the larger compounding ones.

The detailed taxonomy of downstream impact is in Layer 4 of the problem catalogue, which enumerates nine specific categories of business harm with the mechanisms that produce each. The buyer-facing summary is in The Cost of Unmanaged Dimensions.