EfficientEther Research
Packaging budgets, Cloud PC business cases and Copilot rollouts are usually approved on a vendor's promise. EfficientEther Research publishes the evidence instead: open studies, price benchmarks and field guides, each stating its sample, its method and what it does not prove.
Each percentage uses all 3,261 cases. The stages overlap, so they must not be added together.
Open-access papers published on Zenodo under CC BY 4.0, each with its own DOI to cite.
ResearchEfficientEther Ltd · 2026 · CC BY 4.0DOI 10.5281/zenodo.21985871
ResearchEfficientEther Ltd · 2026 · CC BY 4.0DOI 10.5281/zenodo.20556143
ResearchEfficientEther Ltd · 2026 · CC BY 4.0DOI 10.5281/zenodo.20497114
Studies, benchmarks, data notes and field guides, newest first. Each card states the finding before you open it.
AI can write a remediation script in seconds. This sets who approves it, what the reviewer checks and what evidence is kept.
Measured MSIX results, Microsoft platform changes, published packaging prices and a cost model with every assumption stated.
Microsoft keeps up to seven days of tenant configuration history, sized for incident recovery rather than a year of audit evidence.
Platform logs expire long before an annual certification cycle ends. How to prove a setting held all year.
Microsoft repriced Microsoft 365 and moved what sits inside E3 and E5. Check whether you still pay for an add-on you now own.
The difference is the shape of the cost, not just its size. Compare per user group, then migrate in piloted batches.
Assess licensing prerequisites, oversharing, data governance and adoption before you deploy, not after.
Steady-state launch times were similar. App Attach is justified by image externalisation and dynamic assignment, not by speed.
PSF is not inherently unsafe, but the decisive risk sits before signing, not after deployment. Treat it as a full-trust layer.
Four rules every study follows, and the claims we will not make even when the numbers would let us.
Cases that stopped early stay in the denominator, so a headline rate cannot be inflated by quietly dropping failures.
Applications come from the public Windows Package Manager catalogue. Each case is recorded at build, install, launch and clean-up.
Case-level results and the Python scripts that recalculate every table ship with the paper, under a DOI you can quote.
We build EtherApps Forge, the tool used in the packaging study, and say so in the paper. Preprints are not yet peer reviewed.
The limits of the MSIX at Scale study, as published with it.
"It worked" always needs qualifying. Rungs 1 to 4 can be automated for every package; rungs 5 and 6 remain your decision gates.
Learn to capture, build, sign and test MSIX packages using the same staged method as the MSIX at Scale study, so your team can tell a package that exists from an application that runs.
The research gives the method. These calculators apply it to your estate in a few minutes, with no tenant connection.
Calculator
Compare your current Azure Virtual Desktop run cost against Windows 365 Enterprise or Flex for your deployment type. Directional, no signup.
Run the numbersCalculator
A directional comparison of your manual packaging cost against an EtherApps Forge-assisted baseline, before you start a trial.
Estimate the savingJoin the EfficientEther newsletter to hear when a new study, benchmark or field guide is published. No sales sequence attached.
Unsubscribe from any email in one click.
