The artifact

The ranking that goes out on Monday

Not a Slack alert. The leaderboard itself, corrected, with the raw rank each unit came from and a plain statement of who there is not yet any evidence about.

Rep close rate · trailing four quarters5 major rank changes40 reps · team rate 30.0% · prior 39 opps
#RepOppsRawCorrected95% intervalWasEvidence
1Hannah B.64046.7%45.8%42.0%49.5%#3Above team, q = <0.0001
2Rafael C.52040.4%39.7%35.6%43.7%#6Above team, q = <0.0001
3Paulo R.21041.0%39.2%33.2%45.3%#5Above team, q = 0.0036
4Kwame A.joined May 20261656.3%37.6%25.0%50.3%#2Not resolved
5Ruth C.10540.0%37.3%29.4%45.1%#7Not resolved
6Ivy M.47037.0%36.5%32.3%40.7%#9Above team, q = 0.0047
7Greta L.38036.3%35.7%31.1%40.3%#10Above team, q = 0.0268
8Samir P.61034.8%34.5%30.8%38.1%#12Above team, q = 0.0356
9Yuki S.19534.9%34.1%28.0%40.1%#11Not resolved
10Claire V.41034.1%33.8%29.4%38.2%#13Not resolved
11Priya N.joined Jul 2026560.0%33.4%19.7%47.2%#1Not resolved
12Tomas V.joined Jun 20261145.5%33.4%20.5%46.3%#4Not resolved
Do not act on these

8 reps have fewer than 20 opportunities: Kwame A., Priya N., Tomas V., Ines B., Sofia L., Marcus O., Ben H., Dani R.. Their raw rates span 11.1% to 60.0% and none of that range is evidence about anybody. They occupy the top and the bottom of the raw board: Priya N. is raw #1 on 5 opportunities and Dani R. is raw #40 on 9.

changes worth explaining in the QBR

  Sofia L.    #37 -> #23  up   14     6 opps
  Marcus O.   #38 -> #26  up   12     7 opps
  Priya N.    # 1 -> #11  down 10     5 opps
  Ben H.      #39 -> #29  up   10     8 opps
  Dani R.     #40 -> #30  up   10     9 opps
  Tomas V.    # 4 -> #12  down  8    11 opps
  Kai N.      #26 -> #33  down  7   900 opps
  Ines B.     # 8 -> #14  down  6    13 opps

screening: 12 of 40 reps differ from the team rate
           after Benjamini-Hochberg at FDR 5%
           Priya N. (raw #1, 5 opps) does not clear it
What this does not decide

It does not decide compensation, promotion, or termination. It produces a better estimate of a rate with an interval attached, and an honest list of the people about whom there is not yet any evidence. It also assumes the reps are comparable draws from one population. If enterprise and SMB territories are on the same board, split them and fit a prior per segment, or the correction will pull a genuinely strong enterprise rep toward a mean that was never theirs.