[Mcls-print-storage] Allocation numbers
Randy Dykhuis
DykhuisR at mcls.org
Wed Feb 22 15:41:18 EST 2012
Holy cow, this is great! Yes, let's pull the trigger.
=========================
Randy Dykhuis
Executive Director
Midwest Collaborative for Library Services (MCLS)
Lansing, MI & Indianapolis, IN
Phone: (800) 530-9019 x119
Fax: (517) 492-3879
Cell: (517) 927-5121
E-mail: dykhuisr at mcls.org
Skype: randy dykhuis
"Rick Lugg" <rick at r2consulting.org> wrote on 02/22/2012 01:04:50 PM:
> From:
>
> "Rick Lugg" <rick at r2consulting.org>
>
> To:
>
> <rick at r2consulting.org>, "'Randy Dykhuis'" <DykhuisR at mcls.org>,
> <mcls-print-storage at mcls.org>
>
> Date:
>
> 02/22/2012 01:05 PM
>
> Subject:
>
> RE: [Mcls-print-storage] Allocation numbers
>
> Hello again, everyone:
>
> Iteration 15 of the algorithm is now complete, and it looks really
> good. The following are the actual allocations for each library—
> lined up next to Randy’s final statement of the agreed proportions:
>
> Randy SCS
>
> CMU 37,000 37,438
> EMU 67,145 67,221
> GVSU 48,228 49,564
> MTU 49,177 48,655
> SVSU 52,673 53,724
> WMU 110,000 111,607
> WSU 169,844 165,858
>
> I don’t think we can get much closer than this, but thanks to Eric
> it’s quite a bit better than we originally thought. Randy, will you
> let us know when we can release the hounds? I think the only one we
> haven’t heard from is Wayne, and this gets pretty close to their
> maximum allocable number so I’m thinking it’s OK.
>
> Obviously, in order to test these results, we had to actually do the
> allocations. Assuming they’re acceptable, we’ll still need some time
> to get them into formatted withdrawal candidate lists, design
> retention lists, and a few other tasks. But we’re in pretty good
> shape overall. We may be back with some questions about the
> retention lists, but that’s for another day.
>
> Thanks,
> Rick
>
>
>
>
>
>
> From: mcls-print-storage-bounces at mail.mlcnet.org [mailto:mcls-print-
> storage-bounces at mail.mlcnet.org] On Behalf Of Rick Lugg
> Sent: Monday, February 20, 2012 5:06 PM
> To: 'Randy Dykhuis'; mcls-print-storage at mcls.org
> Subject: Re: [Mcls-print-storage] Allocation numbers
>
> Hello, list denizens –
>
> We at SCS spent some time today discussing our approach to the
> blended model, so we can be ready to proceed if the group decides to
> move in that direction, using the ranges suggested in Randy’s
> message. We’ll be experimenting with it over the next day or two as
> your deliberations continue. Our approach will be along these lines:
>
> 1. Identify all allocable candidates with zero charges, and
> assign them according to the proportional goals outlined in Randy’s
message.
> 2. Next identify all allocable candidates with 1 charge, and
> allocate those according to proportional goals.
> 3. If this does not achieve what we need, make further
> adjustments, which may include some instances where a 1 charge
> candidate at a particular institution is favored over a 0-charge
> candidate at another.
>
> We understand that it may take more time to decide about allocation.
> In the meantime, it would help us to have a decision on *what* we
> will be allocating. Those who have spoken to date seem more
> comfortable with a circulation threshold of <4 (Scenario 3a) rather
> than <5 or <6 (Scenarios 3b and 3c).
>
> Scenario 3a yields 534,067 allocable candidates. It includes any
> title-holding that has circulated 3 or fewer times. It is the basis
> of the most recent set of numbers sent by Randy (below). Does anyone
> object to this – i.e., prefer a higher circulating threshold?
>
> Thanks,
> Rick
>
> Rick Lugg
> Sustainable Collection Services LLC
>
> rick at sustainablecollections.com
>
>
> From: mcls-print-storage-bounces at mail.mlcnet.org [mailto:mcls-print-
> storage-bounces at mail.mlcnet.org] On Behalf Of Randy Dykhuis
> Sent: Monday, February 20, 2012 9:07 AM
> To: mcls-print-storage at mcls.org
> Subject: Re: [Mcls-print-storage] Allocation numbers
>
> I thought I'd been hearing that folks were in favor of set targets
> but Barbara & Doug are advocating a blended approach. No one has
> advocated for Rick's second option - basing the lists only on
> circulation so I think we can throw that one out.
>
> As I understand Rick's message, the blended approach could mean as
> much as a 20 percent swing, either way, from the target amount. He
> did also say it could be 10 percent. Assuming the "worst" case. If
> we reduce GVSU's allocation by 12,000 titles, as Doug offered, and
> split those among EMU, MTU, and SVSU*, the range of allocations in
> the blended approach could look like this:
>
> CMU: 29,600 - 44,400 titles
> EMU: 56,916 - 85,374 titles
> GVSU: 28, 982 - 43,474 titles
> MTU: 42,542 - 63,812 titles
> SVSU: 45,338 - 68,008 titles
> WMU: 88,000 - 132,000 titles
> WSU: 135,875 - 169,844 titles (WSU's allocation was at the maximum
> number they can withdraw so in the blended approach they can go no
higher.)
>
> In this approch, we won't know the exact allocation until the lists
> are run. Some libraries are likely to get more titles on its list,
> which means that others get fewer because this is essentially a
> zero-sum game - there is a set maximum number of titles that can be
> withdrawn.
>
> The basic argument for strict allocation numbers without regard to
> circulation is predictability - each participant knows precisely how
> many titles will be on their deselection lists. While some libraries
> will keep items that have had zero circs locally while discarding
> some that have had as many as 3 circs locally, these titles will
> remain accessible and easly obtainable through MeLCat. In addition,
> none of the titles are high demand - any item that has more than 4
> circs since 1999 is automatically excluded from consideration for
deselection.
>
> The basic argument for the blended approach is that it is more
> likely that each library keeps only the titles that have circulated
> locally while discarding those that have zero circs locally but may
> have circulated in other participating libraries. This comes closer
> to matching deselection lists with local activity, and it would
> preclude some ILL/resource sharing activity to ask for a title for
> which there is known demand. It has the added virtue of perhaps
> arousing less opposition throughout campus.
>
> I think that fairly states the case for each approach. I'd like to
> see for a decision one way or the other in the next couple of days
> so that we can get SCS working on this.
>
> *Choosing these institutions is fairly arbitrary & could be done
> differently, including or excluding other participants.
> =========================
> Randy Dykhuis
> Executive Director
> Midwest Collaborative for Library Services (MCLS)
> Lansing, MI & Indianapolis, IN
> Phone: (800) 530-9019 x119
> Fax: (517) 492-3879
> Cell: (517) 927-5121
> E-mail: dykhuisr at mcls.org
> Skype: randy dykhuis
>
>
>
> Doug Way <wayd at gvsu.edu> wrote on 02/18/2012 02:41:28 PM:
>
> > From:
> >
> > Doug Way <wayd at gvsu.edu>
> >
> > To:
> >
> > Barbara J Cockrell <barbara.cockrell at wmich.edu>, Randy Dykhuis
> > <DykhuisR at mcls.org>
> >
> > Cc:
> >
> > "mcls-print-storage at mcls.org" <mcls-print-storage at mcls.org>
> >
> > Date:
> >
> > 02/18/2012 02:41 PM
> >
> > Subject:
> >
> > RE: [Mcls-print-storage] Allocation numbers
> >
> > I apologize for not chiming in earlier, but I am in agreement with
> > Barbara. I would strongly prefer the blended approach. I think it
> > sets a model for a best practice, could save money and staff time in
> > ILL/document delivery, it is an easier "sell" to our campuses
> > ('we're keeping books here that were used a lot'), and I think it
> > may maximize withdrawals and deduplication (if I have a local "high
> > use" title on my withdrawal list, I am very unlikely to discard it,
> > so why have another library that had little or not use keep that
book?).
> >
> > If some additional flexibility is need to help libraries reach their
> > targets GVSU would be willing to reduce its withdrawal number by 10,
> > 000 to 15,000 titles. Hopefully that will still allow us to
> > accomplish our goals, while hopefully allowing the group to
> > implement what is in my mind the best practice that could be used as
> > a model in the future.
> >
> > My 2 cents...
> > Doug
> >
> > From: mcls-print-storage-bounces at mail.mlcnet.org [mcls-print-
> > storage-bounces at mail.mlcnet.org] on behalf of Barbara J Cockrell
> > [barbara.cockrell at wmich.edu]
> > Sent: Friday, February 17, 2012 5:44 PM
> > To: Randy Dykhuis
> > Cc: mcls-print-storage at mcls.org
> > Subject: Re: [Mcls-print-storage] Allocation numbers
> >
> > I maybe in a minority but my preference is to go with the blended
> > approach (Rick's #3 from his Feb 10 email). If SCS allocate any
> > circulation ties so that they are distributed in favor of helping
> > libraries with low targets and particularly pressing needs to reach
> > their targets then that should get institutions close to their
targets.
> >
> > Barbara
> >
> > From: "Randy Dykhuis" <DykhuisR at mcls.org>
> > To: mcls-print-storage at mcls.org
> > Sent: Friday, February 17, 2012 4:50:51 PM
> > Subject: Re: [Mcls-print-storage] Allocation numbers
> >
> > I often take silence as assent & occasionally get in trouble for it
> > but so far there have been no dissenters to the proposed allocation
> > scheme I floated. If there are concerns, it would be good to hear
> > them by Monday so that SCS can get busy cranking out the deselection
> > lists. I know that some of you are anxious to get cracking on this.
> >
> > To restate:
> >
> > Deselection titles would be allocated as shown below. Because most
> > participants want a predictable & precise number of titles that they
> > can deselect, individual library lists will not use circulation data
> > as a way to allocate titles. Depending on what we learn this time
> > around, this could be modified in future iterations of this process.
> >
> > CMU: 37,000 titles
> > EMU: 67,145 titles
> > GVSU: 48,228 titles
> > MTU: 49,177 titles
> > SVSU: 52,673 titles
> > WMU: 110,000 titles
> > WSU: 169,844 titles
> >
> > =========================
> > Randy Dykhuis
> > Executive Director
> > Midwest Collaborative for Library Services (MCLS)
> > Lansing, MI & Indianapolis, IN
> > Phone: (800) 530-9019 x119
> > Fax: (517) 492-3879
> > Cell: (517) 927-5121
> > E-mail: dykhuisr at mcls.org
> > Skype: randy dykhuis
> >
> > "Rick Lugg" <rick at r2consulting.org> wrote on 02/17/2012 06:38:31 AM:
> >
> > > From:
> > >
> > > "Rick Lugg" <rick at r2consulting.org>
> > >
> > > To:
> > >
> > > "'Randy Dykhuis'" <DykhuisR at mcls.org>
> > >
> > > Date:
> > >
> > > 02/17/2012 06:38 AM
> > >
> > > Subject:
> > >
> > > RE: [Mcls-print-storage] Allocation numbers
> > >
> > > Randy’s summary of the situation seems right on target and very
> > > helpful. I think he’s framed clearly the decision(s) needed to move
> > > forward. It does seem that people have most interest in having a
> > > predictable number of candidates. The only real downside to favoring
> > > that factor is that in some cases a copy with 3 circs in a library
> > > might be withdrawn while a copy with 0 circs is retained in
> > another library.
> > >
> > > While this may not be optimal, we’d suggest it’s tolerable: the
> > > title remains secure and available; historical demand is not that
> > > high (3 circs in 11 years); and this approach enables numerical
> > > targets to be hit. That’s our take, but of course we’ll manage this
> > > any way the group decides, to whatever degree the data supports.
> > >
> > > If the group can make a decision today (Feb 17th), SCS can produce
> > > the corresponding allocation lists by early March. We’ll deliver
> > > them as quickly as we can once we know what to do, but it will take
> > > some time. We are also fine with a later decision, as long as
> > > everyone understands that delivery of the lists will be
> > > correspondingly later.
> > >
> > > From: mcls-print-storage-bounces at mail.mlcnet.org [mailto:mcls-print-
> > > storage-bounces at mail.mlcnet.org] On Behalf Of Randy Dykhuis
> > > Sent: Thursday, February 16, 2012 4:00 PM
> > > To: mcls-print-storage at mcls.org
> > > Subject: [Mcls-print-storage] Allocation numbers
> > >
> > > At the end of last week, we received data from SCS about several
> > > potential allocation schemes. Based on conversation during our last
> > > conference call, the preferences seemed to lean either toward an
> > > allocation distributed by collection size or by the proportions we
> > > agreed to when we thought we had 735,000 candidates. Neither of
> > > these were perfect for everyone. This week I spoke individually to
> > > several participants and I think with a little rearranging, we can
> > > make this acceptable to everyone.
> > >
> > > Barbara posted to the list that WMU would be happy with 110,000
> > > titles on their discard list and when I talked to Pamela, she
> > > indicated that CMU would be willing to reduce their title list as
> > > well. Given that, here's where we are:
> > >
> > > CMU: 37,000 titles
> > > EMU: 67,145 titles
> > > GVSU: 48,228 titles
> > > MTU: 49,177 titles
> > > SVSU: 52,673 titles
> > > WMU: 110,000 titles
> > > WSU: 169,844 titles
> > >
> > > That should add up to 534,066 titles. In addition, you all have your
> > > lists of unique titles that will increase your possible
> > withdrawals somewhat.
> > >
> > > That leaves open Rick's question about how to use circulation data
> > > when creating the title lists. Should the lists be created without
> > > regard to where an item has circulated most? Or do we want to make
> > > sure that libraries that have the most circulations retain the
> > > titles (or conversely titles can be discarded in libraries that have
> > > zero circs).
> > >
> > > The ideal is that we have both predictable target numbers, such as
> > > shown above, and the titles on the discard lists are titles that are
> > > low use within each library. Since those are "competing objectives"
> > > as Rick says, my reading of the group is that you would rather have
> > > a predictable number of titles to discard and if that means that
> > > some titles on your shelves have zero circs you'll live with it.
> > >
> > > Is that an accurate assessment?
> > >
> > >
> > > =========================
> > > Randy Dykhuis
> > > Executive Director
> > > Midwest Collaborative for Library Services (MCLS)
> > > Lansing, MI & Indianapolis, IN
> > > Phone: (800) 530-9019 x119
> > > Fax: (517) 492-3879
> > > Cell: (517) 927-5121
> > > E-mail: dykhuisr at mcls.org
> > > Skype: randy dykhuis
> > _______________________________________________
> > Mcls-print-storage mailing list
> > Mcls-print-storage at lists.mlcnet.org
> > http://lists.mlcnet.org/mailman/listinfo/mcls-print-storage
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.mlcnet.org/mailman/private/mcls-print-storage/attachments/20120222/eef7e67d/attachment-0001.html
More information about the Mcls-print-storage
mailing list