GiveWell last laid out a business plan in November 2008.

Judged as of 2010-01-07. True: GiveWell published or laid out a business plan in November 2008 and laid out none between then and January 2010. Mostly true: the plan dates to late 2008, within about a month of November, and no later plan exists. Mixed or mostly false: a business plan was laid out in November 2008, but a later one also appeared before January 2010. False: there was no business plan around November 2008, or the most recent plan before January 2010 is clearly at a different date.

Set before the research began.

1. E1 Self-evaluation: GiveWell as a donor resource - The GiveWell Blog GiveWell Blog (givewell.org) · 9 February 2010 Own record

   “Our progress over the last year (specifically, since our last business plan in 11/2008);”

2. E2 Self-evaluation: GiveWell as a project - The GiveWell Blog GiveWell Blog (givewell.org) · 11 February 2010 Own record

   “As of 11/08, we had only made one extended attempt to hire a third employee”

3. E3 Microsoft Word - Vision Document 2008 11 17.doc GiveWell (files.givewell.org) · 17 November 2008 Own record

   “The path we are focused on now, outlined in a separate document, focuses for the short term on the depth, breadth, and presentation of our research”

4. E4 Revisiting GiveWell Createquity · 22 January 2009

   “GiveWell has published a new “steady-state” vision document describing where the founders see the organization in 3-6 years.”

E3 is a GiveWell document filed under 'Business Plans and Reviews' in a folder dated 2008-11-17, which points to planning documents from November 2008. E1 and E2, GiveWell's own self-evaluations from February 2010, describe 11/2008 as the date of its last business plan, so no later plan is indicated. The independent source E4 (January 2009) mentions a newly published vision document, which fits a late-2008 date but does not confirm the business plan's exact date. Because the evidence comes mostly from GiveWell itself and the November date rests on file and folder names, the claim resolves mostly true rather than fully true.