Copyright
Peter Hammann.

A re-issue policy model online

. (page 1 of 2)
Online LibraryPeter HammannA re-issue policy model → online text (page 1 of 2)
Font size
QR-code for this ebook


LIBRARY

OF THE

MASSACHUSETTS INSTITUTE
OF TECHNOLOGY






ALFRED P. SLOAN SCHOOL OF MANAGEMENT



A RE-ISSUE POLICY MODEL



634-72



By

Peter Hammann



December 1972



MASSACHUSI



. ITUTE OF TECHNOLOGY
50 MEMORIAL DRIVE



IRIDGE. MASSACHUSETTS 021



(MASS. INST. TECH.
FEB 10 1973'
DEWEY LlBRAriY



A RE-ISSUE POLICY MODEL
By
Peter Hammann
634-72 December 1972



Thanks are due to Jean Rutherford for kindly and meticulously typing the manuscript






RECEIVED
MAR 8 1973
M. I. T. LIBRARIES |



Abstract

Publishing firms of books and records follow a well-known policy of
reissuing their products in a different format some time after the initial
introduction. Paperbacks and budget-label discs are two typical formats.
As they usually are accompanied by some price cut, two questions arise:

- When should the product be reissued?

- At what price should the product be marketed again?

The following paper considers these questions from the viewpoint of
quantitative analysis. A simple model is presented, deriving from a case
study in the recording industry. Possibilities of implementation are dis-
cussed.



G37040



CONTENTS



1. Introduction: The Problem

2. The Model

3. Obtaining the Data

4. A Heuristic Evaluation Routine

5. Conclusion



1. INTRODUCTION: THE PROBLEM

In the book and recording industries it has been a common policy ever
since to reissue books and records in a different format, some time after
their first introduction to and subsequent withdrawal from the market. As
a particular book or record aims at some specific market segment there will
normally be some saturation or even a decline in sales. Reissuing the same
product, especially when the price is cut considerably, will overcome this
saturation or open another market segment to which the product's different
characteristics (including price) might appeal. Two typical formats with
which we are dealing here are

- paperback books and

- budget label records

as opposed to the original hard-cover editions or full price discs.

This reissue policy poses two difficult questions (see also (5) for a
general discussion):

1. Should the product be reissued and, if so, at what time?

2. At what price (or in which price category) should the product be
reissued?

In managerial practice, both problems are often dealt with in a lack-
adaisical manner, though this does not mean that decisions were "wrong".
Rather, support for these appears to be very weak. We shall try, therefore,
to supply a simple model, gathering together relevant details and the in-
formation available to the manager. In this, we are guided to some extent
by Little's principles of a decision calculus (4), in as much as the model
will be simple, robust, easy to control, complete on the important issues



and easy to communicate with. The model, however, shows adaptive properties
only to a limited degree as the decision will be a unique one. Though the
problem comes up every half year or year, the alternatives have changed
completely. Information on the performance of reissued products at stage
t-1 cannot be brought to bear directly on the list of alternatives for
decision at stage t.

In the following we shall restrict ourselves to the case of record
reissue policies which have to cover both of the questions revised above
(see (2)). Book reissues, quite often, do not touch the first of these,
as paperbacks follow closely on the heels of hard-cover editions. Fig. 1
shows this graphically:



^



SCO



Sj^Cfc) naiAj _




FiR- 1



Replacement of a hard-cover edition by a paperback one theoretically
may take place when - for the former - average sales equals marginal sales,



1.1



siti = dSitl
t dt



where average sales is maximum. At t*, the time the original edition is
withdrawn, another life cycle for the new edition starts, leading hopefully
to increased sales. With books, it may even happen that both editions con-
tinue to exist side by side for some time, supposedly to attract different
buyer classes in a particular market segment.

In the classical record business, on the other hand, which we shall
have to consider primarily (see (2)), there will be nearly always a con-
siderable time lag between the withdrawal of a disc from the catalogs and
its reinstalment on a budget label. A typical example is the appearance of
so called "historical" recordings from the Thirties or Forties that had been
buried in the companies' archives for some decades. Fig. 2 illustrates this
clearly:

at )




>/V\/WvV^




Fig. 2



The market segment of any record is typically composed of the following
sub-sets of the population:

- devotees of the work or piece of music (e.g., a particular symphony
by Beethoven)

- devotees of the artist(s) involved

- devotees of the record make (for reasons of ourstanding recording
technique, finish or presentation)

lastly, if the record offers a coupling of works or artists

- devotees of a particular coupling

The market audience for budget records, thus, is made up of the following
categories of potential customers

- those who missed the record the first time round for some reason

- those who had to miss it the first time because they could not
afford it

- those who might like to replace a worn-out copy

- those who were interested to have the disc but did not think it

worth the full price asked

o

- those that came new to the scene during the interval t*t„

As will be clear from these specifications, there is nothing more overly
bothersome in record marketing planning than estimation of relevant market
potentials. Things can even get harder, if older titles are coupled across.
In the following, we shall assume that older material will be reissued "straight'
i.e., without changing the coupling, as was the general rule with the company
for which this case was investigated. However, even with "straight" reissuing
of deletions, the firm will incur costs, not only of pressing, handling and
marketing the disc. In many cases, there will be additional costs such as

- cost of technical refurbishing

- cost of new sleeves (cover art, design and sleeve notes)



- cost of new matrices

With every disc, therefore, we can associate some fixed Investment.
Cost of recording, on the other hand, will be regarded as sunk, also con-
sidering the long time lags between deletion and reissue.

As regards production, companies usually set aside some capacity for
pressing and packaging of reissues. This is done with respect to seasonal
fluctuations in production when some job smoothing is called for. Typically,
record companies will subject the decision on what, when and how to reissue
older material to a budget constraint as is also the case with new projects.
Another characteristic seems to be that reissues are produced in limited
lots only at the beginning. In case of commercial success, companies are
willing to follow up rapidly with additional output on a standing basis.
This requires a priori surplus production of a base stock of sleeves, cost
for which has to be incorporated in the fixed investment mentioned above.
The base stock, incidentally, is an arbitrary figure set by management
according to experience. We have not subjected this to sensitivity analyses
in the process of the study. Therefore, analysis of the system's flexibility
was not undertaken in detail.

As will be seen later, the model depends in a critical way on a variety
of subjectively estimated data or subjectively estimated corrections of
historical data that lend themselves to adaptation for the present case.
However, this should not hamper any successful application of the model as
long as data changes can be dealt with effectively and quickly.

2. The Model

As the problem has been structured in detail, we can now proceed to
develop the model. Assume that we have a list of n discrete reissue alter-
natives (i = 1, , n) . Typically, there will be a number of discrete price



categories in which the reissued product could be marketed. Let pik be the

k - th price category for product I. (k = 1, , e). We shall develop the

model for a planning period of T time units (t = 1, , T) . During this

time span the product may be either released immediately or postponed to a
later date. Hence, we define a binary variable:

;ime t and
"0, otherwise
(i = 1, - -, n; k = 1, - -, e, t = 1, — -, T)



jl, if i is selected for reissue at ti
I price k



Assume further that the market has been divided into a set of market

segments (s = 1, , m) . Of course, product i will not necessarily appeal

to potential customers in segments. Therefore, the market potential in any
such segment will be zero. Le the market potential of segments be denoted
M (in units). As the company, regularly, will have to face competition in
its markets, it will eventually succeed in capturing only a fraction of the
market potential in any segment. This fraction of the market potential in
segment s will be denoted by Ot , where

2.2 o < 0( < 1 (s = 1, -— , M)

As we have to regard market segments as open to a variety of products,
this market share applies to the entire line of products appealing to a par-
ticular segment. For any product out of this line there will be again only
a fraction of the firm's total share in segments. Let this product share be
denoted by P iks . it will not only depend on i, but on the price category k
in which the product i has been placed. On the other hand, there will be a
dependency of / iks on a variety of other - more or less indirect - factors
such as

- competition through other albums of the firm's existing catalog (C.)

- competition through albums of competitors' existing catalogs (["*,)

- forthcoming new releases by tl e company (D, )



- forthcoming new releases by the company s competitors (A )

where we could further differentiate according to the price categories appli-
cable to the various products. We may assume, on the basis of empirical ev-
idence, that price categories in this highly oligopolistic market are the same
for all firms concerned. Thus we should rewrite Piks implicitly as follows:

2-3 /^iks = ^iks 2, K^'h, 'l^^'^h, \^^'^h, K^^^'h

and i'(l,2,3)„ the six non-feasible subsets.

Again, we save the items from the non-feasible subsets for inclusion

in the list for t+1, i.e., i .

This ranking and checking procedure is repeated T times, i.e., according
to the length of the planning period and for, possibly, different price cate-
gories. In the end, we arrive at a reissue program, telling us what (and
whether at all) should be reissued as well as the time and price of each
item under review.

The routine is best summarized in the following flow chart:



-16-



Initialize




1




,|






List 1^ ' 1, -
> (■t=l, -— ,T'>


""■


"t


^


I


li


.1




I Calculate





irtition i

' 1"



No Reissues



D~<






balcu



— ,jP_

late B'



f^o I Partition Sets

vB;t>o^ »


1

Online LibraryPeter HammannA re-issue policy model → online text (page 1 of 2)