Statistics, speculation and stationery
What can several decades of Microsoft naming history tell us? Mostly that a product name is less a permanent identity and more a medium-term tenancy agreement.
Analysis uses the registry’s reproducible “as of” date.
Days since last rename
—days
The nameplates have remained still
Measured against the registry’s reproducible “as of” date.
◐ Analysing the paperwork…
○ We couldn’t analyse the registry.
Rename vital signs
The paperwork at a glance
The family league table
Where names have the shortest lease
Families are compared without silently merging categories. “Limited history” means fewer than two completed renames, where a confident conclusion would require more optimism than data.
Rename waves
When the nameplates moved
Completed name periods grouped by the year the next identity began.
Repeat offenders
Most documented identities
A leaderboard nobody explicitly asked to win.
The Rebrand Risk Index
What might be renamed next?
The index weighs current-name age (45%), prior identities (35%), and family rename frequency (20%). It measures historical rebrand pressure, not probability. Corporate naming committees remain stubbornly non-deterministic.
How this was calculated
Enough method to keep the joke honest
Measured
All figures are calculated in the browser from the canonical registry. Durations use completed name periods; current-name age ends at the dataset’s “as of” date. Medians limit distortion from unusually durable names.
Limited
The registry includes surviving products only, so discontinued products are absent and the analysis has survivorship bias. Family sizes differ, partial dates retain their documented precision, and correlation is not a secret product roadmap.
Predicted, technically
Scores are deterministic rankings of three historical signals. Status bands are deliberately broad: ● Elevated, ◐ Watchlist and ○ Not imminent. A high score means “good setup for a joke,” not “confirmed for Build.”