Case Study: Turning Azure Tagging into Showback and Cost Ownership

Problem

A large enterprise had an Azure estate with more than 6,000 resources across 740 resource groups and 32 subscriptions, but tagging was not reliable enough for meaningful showback.

Only around 2% of resources had tagging that was usable for cost ownership. The rest of the estate was difficult to group by owner, service, application, or business area.

Context

The organisation needed to understand where cloud spend belonged, but the platform had grown faster than the governance model around it.

Without consistent tags, reporting could show the total Azure bill, but it could not clearly explain who owned the spend or which services were driving it. That made cost conversations slow and sometimes subjective.

Environments were scattered across subscriptions, resource ownership was unclear, and tagging had grown organically with no agreed standard. Some resources had tags, some did not, and many tags used different names or values for the same purpose. So the organisation could not reliably explain who owned what, what workloads were running, or how cloud costs should be attributed.

Management had been asking for clear workload ownership and showback for years. Internal teams had attempted tagging clean-up several times, but the work repeatedly started and stopped. The scale was too large, the process was too manual, and there was no reliable automation or shared view to drive progress. After five years, little meaningful progress had been achieved.

What was accomplished

Our team reviewed the estate, automated tag exports into one shared view, rationalised the tagging model, and helped turn tagging from a technical clean-up activity into a practical showback capability.

The team began with only around 10 resource groups per engineer. This gave them time to investigate each resource group properly, identify owners, understand the purpose of the workloads, and agree the correct tag values.

Daily workshops were established to review progress, unblock questions, validate ownership, and refresh the shared tagging view. This gave the team a firm grip on the work and avoided the common problem of tagging projects losing momentum.

Key decisions and trade-offs

The tag model had to be useful, not perfect on paper. Too many mandatory tags would have created friction and poor adoption. Too few tags would not support showback.

The practical decision was to focus on the tags that helped with ownership, reporting, and cost action first.

Result

The tagging implementation took approximately two months. The following two months were used to act on what the new visibility showed, including legacy resources, proof-of-concept workloads, non-production resources in production subscriptions, unclear ownership, and services that no longer had a business owner.

The work enabled showback across the Azure estate and improved usable tagging from around 2% to a level where cost conversations could be tied to owners and services.

The wider cost programme achieved around 35% cloud cost reduction. Tagging was not the only lever, but it was a key control because it made ownership and accountability visible.

Conclusion

Tagging only matters when it helps people make better decisions.

Many organisations talk about Azure tagging, but very few complete the work properly. Tagging is often treated as a policy or governance discussion, when the real challenge is execution.

If your Azure spend is hard to explain because ownership is unclear, platform governance and an Azure cost optimisation review should start with the data needed for showback.