Your IT data isn’t dirty, it’s multi-truth

Nanda Vijadev

We break down why enterprise IT numbers rarely match reality and how contradictory identity, application, and spend data erodes trust. We dig into how inflated user counts, noisy app discovery, and fragmented purchasing emerge from systems that are each locally correct, and why true accuracy requires curated, decision‑grade data enriched with context, normalization, linkage, and consistent scoping.

In this first part of our series "Closing the IT Data Trust Gap: From Raw Records to Decision‑Grade Intelligence" we talk about why enterprise IT numbers don’t match reality and why asset management tools can’t fix it.

You just connected your identity provider, CASB, and procurement system to a new IT asset management platform. The dashboard loads. And immediately, something feels wrong.

Your directory says you have 312 users. You have 140 employees. Your CASB reports 1,847 applications. Your team uses maybe 80. Shadow IT flags Cloudflare, Akamai, and accounts.google.com. Those aren’t shadow IT; they are infrastructure services.

The natural reaction isn’t curiosity. It’s skepticism. And once skepticism sets in, teams stop acting on the data and fall back to spreadsheets and one-off reconciliations. The platform becomes expensive shelfware.

But here’s what most people get wrong about this problem: the data isn’t dirty. It’s contradictory.

Every system is locally correct

Enterprise IT data is best understood as “multi-truth.” Every system was designed with a point of view that is locally correct. Entra ID tracks every identity object — service accounts, guest accounts, test accounts, disabled accounts because that’s its job. A CASB detects every service touching the network (CDNs, auth endpoints, browser extensions) because that’s its job. Your ERP captures negotiated contracts, while credit card transactions scatter across departmental budgets.

Each system is doing exactly what it was designed to do. The problem appears when you combine these points of view without an enterprise entity model that explains how they relate.

Four trust killers

Through our work with enterprise customers, we see the same four patterns break trust every time:

Identity inflation. Directories track accounts, not people. Service accounts, guests, test accounts, and disabled accounts inflate raw user counts to 2–3x actual headcount.

Application inflation. CASB and network discovery detect services touching traffic, not tools employees consciously use. CDNs, auth endpoints, and infrastructure services produce app lists 10x larger than reality.

Noisy Shadow IT. When every discovered app without a direct entitlement gets flagged including Akamai, Cloudflare, and apps already licensed through bundles, the signal-to-noise ratio collapses. Teams stop looking.

Fragmented spend data. Purchase data lives in disconnected silos: ERP contracts, departmental credit cards, direct vendor subscriptions, making it impossible to connect what was bought to what’s actually being used.

In every case, the underlying source data is technically accurate. The gap is in interpretation: raw data hasn’t been enriched with the business context needed to make it meaningful.

From “clean data” to “decision-grade data”

The industry talks about “clean data” as if the problem is data hygiene. It’s not. The problem is data curation. What IT leaders actually need is decision-grade data, data that has been:

Normalized. Vendors, products, identities, and payments resolve to consistent canonical records

Classified. Entities reflect how humans reason: people vs. accounts, user-facing apps vs. infrastructure services

Linked. Usage, entitlements, and spend connect so rollups are explainable

Scoped. Consistent filtering applied once and reused everywhere, so different screens never contradict each other

Scored. Data quality is explicit: missing fields and weak matches are surfaced as known gaps, not hidden errors

Until data is curated to this standard, every downstream workflow from optimization to renewals and risk assessment produces the wrong answer.

This is why asset management tools that stop at inventory and reporting will always disappoint. You can’t optimize what you haven’t curated. And you can’t act on data you don’t trust.

In the next post in this series, we’ll look at the most common trust killer in detail: why your user count is 3x your headcount, and what it takes to fix it.

The difference between raw data and trusted intelligence is enrichment.

And enrichment is only as good as the knowledge that powers it.