Skip to content

Case studies

What Beamer's products find on real codebases.

Every study says whether it comes from a customer or from public data, in the same place on every page.

All case studies

  • Where a quarter's extra engineering hours went: the extra hours as bricks, by domainPassword vault clients carried $810K of the $1.2M of sync tax.Ideal-structurecost8,500 hSync tax12,000 hSpill-overwork86%Coordinationoverhead10%Repairs andworkarounds5%Password vaultclients$810KShared cryptoSDK$140KServer coreAPI$130KMobilepasswordmanager$63KDirectory sync$8KPassword vaultclients$810KShared cryptoSDK$120Kto removeServer coreAPI$130KMobilepasswordmanager$63KDirectory sync$8KPassword vaultclients$720Kto removeShared cryptoSDK$120Kto removeServer coreAPI$130KMobilepasswordmanager$63KDirectory sync$8K
    DefusorPublic data study2026Q3

    Where a quarter's extra engineering hours went

    2.3×what the work cost on 1,286 finished changes, against an ideal structure

    On 1,286 of Bitwarden's 2026Q3 changes, work cost 2.3× what it could have taken with an ideal structure. Most extra hours came from spill-over changes.

How a study works

  1. Step 1

    Read the history

    Defusor reads the git history as one product: every repository and every finished change in the period.

  2. Step 2

    Price the work

    It estimates what each change could have taken with an ideal structure, and what the current structure added on top.

  3. Step 3

    Publish it with its clarifications

    Every study says where its data came from, which figures are assumed, and what could not be measured.

Get a study of your own code

Defusor reads your git history and returns this report for your product.

Request a free test