Question

Topic: Research/Metrics

Validity Of 'win Rate' On Small Sample Size

Posted by Anonymous on7/21/2009 at 2:17 PM ET 250 Points
I am working with 2 sales groups, and there is a debate over who has a higher 'win rate' for prospects over a certain size.

Over the last three years, results are as follows:
Group 1 - 52 opportunities, 31 'wins' (60% win rate)
Group 2 - 52 opportunities, 20 'wins' (38% win rate)

Question 1 - is this difference meaningful?

Question 2 - Group 2 insists that it has a better close rate for the most important prospects within the opportunity set above, namely the largest prospects. For the 17 largest prospects out of the 52 opportunities, they won 9, or 53%. Is this 'win rate' meaningful?
To continue reading this question and the solution, sign up ... it's free!
SUBSCRIBE
或,SIGN IN

RESPONSES

  • Posted bymelissa.paulikon7/21/2009 at 3:00 PM Member
    Group 2 may have a valid point if these larger customers mean higher revenue, or did Group 2 have to discount heavily to win the business? Initial revenue isn't always the deciding factor if there's significant opportunity for additional revenue from larger customers.

    In addition, if these customers were marquis names that the company badly needed to establish their presence in a market, that could be significant as well.

    It really depends on what's important to the company.

    All the best!

    Melissa
  • Posted byCarolBlahaon7/21/2009 at 3:29 PM Member
    why is it meaningful? You don't understand the salesperson psyche. You are either winners or losers. You either make the number or you are a loser. No in-betweens. You are a profit center. You are as good as your last 30 days. That is what drives winning sales people.

    Its about dollars, not # of closings. It takes more time to close a $200K sale vs a $20 sale. Its much more complex sale. What brings impact to the bottom line--# of closings or dollars-- revenue. Try paying the mortgage with close ratio alone. Its all about the dollars in the bank.

    It is however good to track both #'s. And to track missed opportunities. Maybe the higher close ratio group sells something better-- a good add on sale than the other. There is much to learn from the entire group intelligence.

    When I conduct sales meetings I ask for SOFT reports-- Success, Opportunites, Failures and Threats. When someone in the group hits a stumbling block-- the group as a whole gets involved. Almost always they cone up with suggested counters to the issue. Its very powerful.

    Sell Well and Prosper tm
  • Posted on7/21/2009 at 4:05 PM Author
    Thanks for the responses. This was actually intended to be a statistics question. Sorry if not clear. By 'meaningful' I meant statistically significant.
  • Posted bykoen.h.pauwelson7/22/2009 at 5:35 AM Accepted
    Statistically, you can set up this question either as
    1) do the confidence intervals around the two hit rates overlap or not?
    2) can we reject the null hypothesis that the win rate is the same across both groups?
    Going with the former, the next question is whether the 52 opportunities in both groups can be assumed to be drawn from the same population. In other words, is their lead quality the same? Assuming this is not in doubt, we next ask whether we can assume the hit rate can be approximated by the Normal distribution. That is the case, as n is large enough according to the central limit theorem. Therefore, we calculate the confidence intervals around the two hit rates of p = 0.60 and p= 0.38, using the 95% confidence interval (Z = 1.96)

    p +/- 1.96* Sqr[(p(1-p)/n]

    For the first group = 0.60 +/- 1.96*sqr[(0.60(1-0.60)/52]
    = [0.467 0.733]

    For the second group = 0.38 +/- 1.96*sqr[(0.38(1-0.38)/52] = [0.248 0.512]

    Because these confidence intervals overlap, we can not say with 95% confidence that the hit rates of the two groups statistically differ.
  • Posted bykoen.h.pauwelson7/22/2009 at 5:40 AM Accepted
    My above response raises the question at which confidence level the two responses would differ. If you use 90% confidence instead (Z = 1.645), we instead get:

    [0.488 0.712] and [0.269 0.490]

    仍然有些重叠。因此,您将需要to use 85% confidence levels or less to obtain the result that the hit rates differ statistically. Is that reasonable? We are now going out of statistics (which prefers to use 90% or 95%) and into management. It depends on your managerial tradeoff of making an ‘error’ either way. Ask yourself: do you really need to be more than 85% sure that the actual hit rates/capabilities of these two groups differ before making a decision such as giving the first group a higher bonus? I think not. But still, your call.
  • Posted bykoen.h.pauwelson7/22/2009 at 5:42 AM Member
    Just noticed your second question; and I agree with Melissa and Carol: it all depends how much profits certain prospects bring in. Instead of the hit rate, you can calculate the total profits brought in by each group and compare these two totals
BACK TO TOP

Post a Comment

Related Discussions

Community Info

Top 25 Experts
(Research/Metrics)