To benchmark a client's Azure environment against other MSPs' clients, count high-impact open findings per 100 resources, so big and small environments compare fairly. 91 or fewer puts a client in the best-kept quartile of The Hidden Azure Work Index (Spotto, 2026), the ANZ baseline. Azure Advisor score and Defender for Cloud secure score rate each environment against Microsoft's own checks, and neither compares it with similar environments at other MSPs.
Download the report (PDF) · Read the key findings
Free to download, no form.
Why benchmark against peers at all?
A peer benchmark shows a client what good looks like, so you can agree where they want to be and prove the change. If you sell outcomes, the client has to be able to see them, and seeing them starts with a baseline.
Microsoft's scores give you two reference points: its own checks, and the client's history. A peer baseline adds a third: where the client sits among similar environments at other MSPs.
What does a well-managed Azure environment look like?
A well-managed Azure environment still has work open, just far less of it. The Index sorts client subscriptions into four equal groups, called quartiles, from fewest high-impact findings per 100 resources to most. The typical (median) subscription in the best-kept quartile has 67, against 227 in the quartile with the most open work: about a third as much.
Even the best-kept subscriptions have open work, so a long list doesn't by itself mean a client is behind. The best-kept estates are much further through the basics. Share of subscriptions with each item still open:
| Still open | Best-kept quartile | Most open work |
|---|---|---|
| Storage without zone redundancy | 64% | 95% |
| App Service on a single instance | 24% | 83% |
| Disks without zone-redundant storage | 33% | 52% |
| No Azure Site Recovery | 39% | 59% |
| No backup configured | 31% | 59% |
Most of the gap sits in recovery and resilience. For App Service, Microsoft advises deploying multiple instances, and its disaster recovery guidance says to test restores regularly to validate backups.
Why don't Advisor score and Defender for Cloud secure score answer this?
Both measure an environment against Microsoft's checks and track it over time. They're good tools, so keep using them.
Advisor score rates assessed resources against Advisor's recommendations, as a percentage for each category and overall. It doesn't yet tell production workloads from dev-test ones.
Defender for Cloud secure score combines Microsoft's built-in security recommendations into one score. Its classic model and the newer risk-based Cloud Secure Score give different values, so name which one you report.
Microsoft's peer comparison sits in Microsoft Secure Score in the Defender XDR portal, which shows an anonymised average for organisations with a similar seat count. It covers identity, devices and apps rather than Azure resources, so it doesn't benchmark Azure infrastructure.
How do I benchmark a client step by step?
Count high-impact open findings per 100 resources and place the result against the Index's quartile cut-offs. The ranges come from Spotto's checks, so compare a count made the same way.
- Count high-impact open findings, the same way. Leave out policy checks, which carry one blanket rating. Count each finding on every resource it applies to, so a backup gap on five virtual machines counts as five. Without Spotto, you can count Advisor's high-impact recommendations per affected resource. That sits on a different scale, so treat the quartile as a rough guide and track the client against their own number. For a count that matches the Index, run a Spotto estate analysis or the 14-day trial on one client.
- Scale it to 100 resources. Divide the count by the number of resources in the subscription, as Spotto counts them, then multiply by 100. For example, 40 findings across 80 resources gives 50. The benchmark covers subscriptions with at least 10 resources.
- Place the client in a quartile. The quartiles are cut at 91 and 171. A client at 91 or fewer is in the best-kept quartile, 171 or more is in the quartile with the most open work, and anything between sits in the middle half. So a client at 140 sits in the middle half. The best-kept quartile's median is 67.
- Pick the gaps that matter. Start with the table above.
- Agree a target and track it. Agree where the client would like to be. Your service delivery lead owns the number. One line in the monthly review does it: You're at 140 per 100 resources, in the middle half. We're aiming for 91 by [month]. This month we closed: …
The Index covers ANZ, and you know your client, so you decide what good looks like.
What should I check before recommending changes?
Check whether each subscription runs production, and agree the client's requirements before you scope and price anything.
- Site Recovery and backup add to the bill, and dev/test estates may leave them off on purpose.
- Zone redundancy needs a region with availability zones. Australia Southeast has none.
- A second App Service instance doubles that plan's compute cost.
- Record the reasons as client context in Spotto.
What is a realistic target?
Over 90 days, the typical environment barely moved: it cut its open high-impact findings by 2%. The quarter of environments making the most progress cut theirs by about a fifth (19%). A client who cuts a fifth in 90 days is keeping pace with them. A drop can also mean resources were removed, so check what changed before you report the progress.
How do I try it on three clients?
Pick three clients: one large, one in the middle, and one that's due a fresh technical review. That last one often holds the most surprises.
- Count and compare. Work out each client's number per 100 resources and compare it with the best-kept cut-off of 91.
- Group and plan. Sort the findings into the report's five bundles (identity and access, network exposure, resilience and recovery, database posture, and cost and cleanup), and use them to shape the next 12 months of work with each client.
What is a benchmark worth to my team and my client?
The client sees where they sit, what the best-kept have done, and the change month to month. Your team gets a reason for the next job the client can already picture. Most of these jobs are medium effort: in the report's six worked examples, a job takes about 2 to 16 hours, by Spotto's estimates, so many can be scoped and quoted on their own.
Before a first meeting, run an estate analysis on a prospect and set its count against the Index, so the conversation opens on where they sit. With Spotto, the AI-Assisted Cloud Operations platform for Cloud MSPs, your team sees findings ranked across six categories (cost, security, compliance, reliability, performance and operational excellence) by impact and effort, and turns agreed work into a draft statement of work to review and price, then PSA tickets. At the monthly review, a service delivery report shows what changed, and your clients log into a client portal to see their own cloud.
About the data
Source: Spotto platform data in The Hidden Azure Work Index (2026). The benchmark uses 353 client subscriptions with at least 10 resources. The 90-day change uses 201 subscriptions tracked over the same 90 days, leaving out days affected by changes to Spotto's checks and estates whose resource count fell by 20% or more. The category split uses the full pull of 417 client Azure subscriptions with open findings, read on 15 September 2026: security and reliability make up 85.6% of open findings, and cost is 3.5%. Findings are counted as resource instances, and each finding is assigned to one type of work. The figures cover ANZ only.
Try it on three of your own clients. Where would they sit? Download the report (PDF), or talk to us about a discovery call, then a free 14-day trial on one client.