in
o
LJ
>-
ay .tt*J**>f tt g
The
Service Bulletin
of the
Bureau of Personnel Research
H. G. KENAGY. Editor
NOVEMBER, 1922
No, 3
The Bureau
exists
to serve
business
organizations
in the
SELECTION
TRAINING
and
SUPERVISION
of
PERSONNEL
In this Issue:
TURNOVER IN
SALES ORGANIZATIONS
BY
WILLIAM E. LANGE
The Division of Cooperative Research
Carnegie Institute of Technology, Pittsburgh, Pa.
THE SERVICE BULLETIN
The Service Bulletin is the medium through which the Bureau of
Personnel Research publishes the results of its scientific research in the
various fields of personnel. Each issue of the bulletin discusses some
problem of general interest, outlines the plan of research which is being
carried through, or gives the results which have been secured.
Copies are furnished free to all companies which are associated
with the Bureau in its work.
The following titles of bulletins indicate the general lines of re-
search and service with which the Bulletin concerns itself:
PROBLEMS OF SALES ORGANIZATION.
MANAGEMENT OF A BRANCH SALES AGENCY.
BASIC EXPERIMENTS IN SCIENTIFIC SELECTION.
ANALYZING THE SALESMAN'S JOB.
PREDICTING THE CAREERS OF CLERICAL EMPLOYES.
TURNOVER IN SALES ORGANIZATIONS.
CHARACTERISTICS OF SUCCESSFUL SALESMEN.
TENDENCIES IN PERSONNEL PRACTICE.
The April issue of the Bulletin will contain the results of time-
studies of specialty salesmen in four different selling lines.
TURNOTCR HI THE SAIES ORGANIZATION
A Study of Actual Turnover Oonditions in
a Large ITumber of Matlonal Sales
Organizations V/ith a Discussion
of Causes and Remedies
By
William B. Lange
Research Assistant
The Bureau of personnel Research
Carnegie Institute of Technology
November, 1922.
All Rights Reserved
• «.. . . •
TURNOVER IN SALES 0BG/>3IZATI0HS
TABLE OF CONTEOTS
VAGS
IS TURNOVER A SEllOUS IM.TTER , 1
Instances of High Turnover 2
Costs of High Turnover 3
TEE FACTS OF TURNOVER
How Data Were Gathered 4
Results of Questionnaire 5
Reliability of Data 6
AMLYSIS OF (SKER.iL TDRSO^R CONDITIONS 7
COMPARISON OF TCRFOVER AMOT^G 0EGA3:iZATI0NS HAIDLING
DIFFERENT PRODJCTS 10
COIviP .PRISON OF TORI«TOSR IN OFFICE SPSCIALTF OEGAMZATIONS
WITH TURNOVEE IN ALL OTHER BUSINESSES 12
TBB NEED FOR BETTER TURNOVER DATA ^ y //{ /4 â–
Insufficiencies of Present Data 14
LIMITATIONS ON VALUE OF fURl^OVER DATA 16
WHAT TURNOVER RECORDS SHOULD BE EEPT 18
Record Form - Txirnover of Seles Force
Analysis of Exits - Termination Record
VALUE OF PROPER BECOIDS 18
r>24'H3
i*jewM(9».-j!-\*jr
.'{finqikii-:
TURI?OVER IN SALES ORGANIZATIONS
Is Turnovor A Sorlous Matter.
Is largo tiirnovor ^vithin sslos organizations a serious ;
matter? Evidently it has not beon so rogardod by nany companios,
for only a small per cont of nstionul selos orgfjnizations havo
given the matter iriuch consideration. Convincing proof of this fact
comes readily to any organization or concern which seoks to investi-
gate conditions. For one reason or another, most companies have not
thought it if.'orth xvhile to keep turnover data . Records of factory
labor turnover hare been considered essential for years; salos labor
turnover records are still regarded as fads.
Unintentionally, manf/ sales executives hare adopted an
indifferent attitude tov/ards the matter. Salesmsn turnovor has
beon uniformly high during the past few years and, as a result, the
condition has come to be accepted as a natter of course. A turnover
percentage which, considered alone, might havo caused serious concern,
lost much of its alarming quality t/hen considered in connection with
general business conditions and the situation in other companies.
The abnormal conditions since 1915 hav'B led other sales
executives to adopt a hopeless attitude - to believe that high
turnover is unavoidable. The fact that causes other than abnormal
conditions may. be operating within ji'.a*les orgaEization to cause high
turnover has not beon considered. Believing that abnormal conditions
alone were to blame and that those conditions could not bo changed by
their efforts, these sales managers have simply let turnover run its
-a-
coursa. <-
Some executives have not evi^n been aivare that turnover with-
in their sales organizations has been high. In the prose of other
matters, they have noi^locted to secure the facts. Altho\jgh a"'are
that many men â– vrore leaving their organizations or being dropped,
they havo not appreciated havr serious the condition wes, since they
have not hsd figures showing the turnover in percentages.
Instances of High Turnover.
A turnover of 100 per cent in any year indicates that the
nvimber of changes in personnel equals the average number employed,
and yet percentages in excess of this figure are not uncommon,
especially during the Inst year or t-wo. A real estate sales organi-
zation, employing a sales force v/hich at times has numbered one
hundred men, had an average annual sales turnover for the period from
1915 to 1922 of over 400 per cent. The pcrcontago of turnover for
each year is indicated below.
Year
Average Rumber On
Total Number
Total
Per Cent
il
les Force
Hired
Exits
Turnover
1915
60
250
200
333
1916
80
400
320
400
1917
40
150
110
275
1918
35
125
90
257
1919
100
525
425
425
1920
60
360
300
500
1921
20
125
105
525
1922
(to
Aug
.)
25
100
75
515
Conditions .Ith this organization are exceptional and yet
turnover figures in excess of 100 per cent in other sales organi-
zations can be cited. A notable example of high turnover is furnished
by an organization handling an office specialty. Since 1919 its
». f
100
42
110
126
76
101
281
181
130
308
268
394
-3-
percentage of turnover has not fellen helavi 100 per cent. In fact,
for the seven months of 1922 > the turnover was 394 per cent. The
range of txirnover percentages is shown below %
Year Average Kxjinber On Total' Number Total For eent
Sales Force Hired Exits Turnover
1919 38
1920 75
1921 139
1922 (to Aug.) 117
This organization increased its sales force from an average
of 38 in 1919 to 117 in 1922. With this increasing sales force it
has had increasing turnover. During the seven months of 1922 alone,
268 salesmen left the company and 308 wore added. Conditions such
as this are surely worthy of serious analysis.
Costs of High Turnover
The seriousness of a high sales turnoTrer lies in its costli-
ness. The sales manager of a company employing over 600 salesmen
states that each salesman represents an expenditure on the part of
the company of one thousand dollars for his training and supervision-
One hundred and eighty salesmen left this company during the first
seven months of 1922, v;hich means t\ cost to the company of ^^180,000.
This is a considerable s\an of monfeyl
Most companies probably do not have sc expensive a plan of
training as the company mentioned, but whatever method of training
is used involves some expense. The new salesajian may simply travel
with an experienced salesman for a certain period of time, or ho may
bo trained in the company's course. In either event, he represents
an investment v;hich is lost to the ccmpany when he leaves its ert^loy.
Furthermore the new salesman requires more supervision than
-4-
-fche more experienced, aad thus increases sales cosbs. In an organi-
zation 'vith a rapidly changing personnel this item of increased
supervision cost becomes of considerable importance.
The increased supervision cost caused by a rapidly chang-
ing sales personnel is dwarfed, however, b^'• the cost of lost business
due to high salesman turnover. A eustomer does not feel confidence
in a company which has a rapidly changing sales personnel. In the
customer's eyes, the salesman is the company. If the company changes
representatives often, the customer's confidence in the houso and his
willingness to buy c??.nnot be conserved or developed. As a reiBult,
business is lost to the company.
Hot all companies lose a thousand dollars every tine a
salesman leaves. For most companies the loss may be conservatively
placed at tv'o hundred dollars. A company "which has a sales fcrco of
one hundred and a turnover of thirty per cent, would lose six thousand
dollars per year on this basis. Such a sum is worth saving, and a
great part of it can be saved by reducing turnover.
Turnover can be reduced if its causes are known- In order
to discover these causes, however, sales turnover conditions must be
studied and accurate turnover records must bo maintained for regular
analysis.
THE FACTS OF TURKOl^R
How Data Were Gathered.
One hundred and fifty national sales organizations were ad-
dressed by the Biireau and asbsd to cooperate in an investigation of
salesman turnover by furnishing their turnover figvires. These organi-
zations ^nere selected so that they included concerns handling many
different types of product.
-5-
For the convenience of the companies and to Insure
uniformity in the calculation of turnover porcontages, each company
â– was fxirnished with a Sales Turnover B'la:ik, a copy of which is in-
cluded in this report. Annual data were as]<-ed for covering the
period from January 1915 to August 1922 and full directions for all
calculations were included on the blank.
Results of Questionnaire
Responses wore received from sixty-nine of the one hxindred
and fifty companies addressed by the Bureau. Of these sixty-nine
companies only thirty-five v^^ere actually able to furnish turnover
figures. Several other companies, thovigh vmable to furnish figures,
estimated general turnover conditions within their org-inizations and
manifested interest in the investigation.
An analysis of the responses indicates the wide variety of
concerns addressed by the Bureau and the status of turnover records
among sales organizations handling various types of products.
TABLE I
ANALYSIS OF THE RESPONSES TO THb' BUREAU QIESTIONKAIRE
Number of
Number
Type of Company
Responses
16
FvT n i s hi ng F i gur e s
Office Specialties
10
Automobiles
7
3
Rubber Products
5
1
Life Insurance^
4
3
Real Estate, Mortages,
Stocks &"Bonds
3
1
Steel
3
Clay Products
2
2
Heating Equipment
2
2
Commercial Feeds
1
1
Drugs & Drug Store Suppl:
ies 2
2
Paint
2
Electrical Supplies
1
1
Vacuum Cleaners
1
1
Electric Lights
1
D
SALES TURNOVER RECORD
(Confidontial laformEtion)
COMPAKY DATE
ADDRESS OF COMPAIIY
Important: Ploase rood directions carefully before recording data. If you
cannot supply the data according to the headings given, then
fiirnish turnover figures as you have them, explaining how the
percentage figvire is sooured.
YEAR :
ATORAGE mUEER OH :
SALES FORCE :
TOTAL limiBER
HIRED
TOTAL EXITS • â–
PER CEIIT
. TURKOVKtt
1915 :
•1916
• .
1917 :
1918 '
â–
1919 •
1920 :
1921 :
1922
to
AuOTSC
« . _^
1
DIRECTIOIIS : (1) The "average number on 6%les force" for any year may be
obtained by adding the number on the payroll at the begin-
ning cf each nonth rnd dividing by 12 .
(2) "''Total number hired" means the entire number of additions
to tha sales force during the year — in other words, the
men hirod.
(3) "Total exits" moans tho number of salesmen "who loft tho
company during tho yoar for any reason, including discharges.
(4) "Por cent turnovor" is tho figure obtained by dividing
"total exits" by "average number on sales force".
(5) When tho information called for has been filled in, mail
tho blank at once to:
â– THE BUREAU OF PSRSCIITIIEL RESEARCH
CARIEGIE INSTITUTE OF TECHl^OLOOT
PITTSBURGH, BElfiJSYLVANIA.
#
Table I ((Ccartinaed)
at
jtyp^
ot osrqssy
3es-?ns-r s
Corlring Utaasils
Fil)re BoKxd
Food Br odu B ts
looriac R-oteets
loveltiss
M^iBS Honsa Bro ducf
SoKp
B*B Cl0thil«
OU
Sopplies
Plate GftBM
Taps ft DIbs
Brass Fro^scte
1
1
1
1
1
1
1
o
o
o
o
o
o
In aAslfeiOB to the flgodres fsvislBd diroetlf ^ tiiir^^
tlTs coafianies, data were reeeired frca tbe t^i^TBMw^p^
IBfftitvte and the Joaepk ^idiarda Coa^ry '»iio were £lao
iarasti^tica of Salsa ?«m0rer. Tbeir d^ta ecrscrsd i3drty>-#iTe
ecapacioe aiKxis ^^ich vare several -::^idi hsd ftlrcadr ra po t tod tc tlis
â– â– reav*
Be liability of Pats.
Statistics ot Ma r n pr e r have been, in ^na>«l« aaBstis-
faetcry. tba aaall andier of coapanies f^miiAin^ data is aiaian is
tlaa freeeding table. Bvea in those coapaaies '•inure emylctB raeords
are kept, the preamce of a lan^ naiber of -variables avy eawae tlie
sales zsnsger to dexn^ that a hi^ turasrer rate has ca^ eieaificaisce.
Ihe abnsmelity of the tues aid the aggresslre tactics sf ssles
Mana j^ers in "cleaaiag taoaae* faarve caosed a hS^ tanc««r* OBnr
variables affection tvraonrer src ecapeasatisa. the liiMilai cf the
liae, tl»e aetfaods of selection and training, and e-vesntte sine of tlw
-7-
company. These variables, however, do not so seriously affect the
individual turnover ratios as to pre\-ent the indication of the
general tend of turnover or the relation of the individual turnover
ratio to other ratios.
AKAiySIS OF CEHBRAL TORI^OIER COI33ITIOHS
probably the best index of general sales turnover condi-
tions is the average yearly percentage of txirnover for all the
companies reporting. In Table II, the year, the average percentage
of turnover for that year, and the number of companies reporting for
that year are listed.
tli/ff^ TABIE II
(SIIjEBAL TURIiOVER COIffiilTIONS AS SHOWII BY
A^^ ^GE FERCEIJTAGE OF TURHOVE^
Mumber of Companies Average Percentage
Yeay ; Reporting of Tnr never
1915 12 41
1916 16 49
1917 20 45
1918 24 51
19X9 26 60
1920 32 60
1921 S3 77
1922 35 85
It can be seen that during the period from 1915 to August
1922, there has been a gradual increase in the percentage of pales-
man turnover, though this may be the result of the fact previously
mentioned, - that the nxanbcr of companies keeping turnover records
increased with each year. In 1915, only twelve of the thirty-fire
companies maintained accurate records. From then on there is a
gradual increase 'intil in 1922 thirty-fivo companies "/ere able to
supply data.
-8-
The general increase in turnover in these thirty-five
companies can be shown to better advantage by the use of a chart*
The average yearly percentage of turnover for all concerns report-
ing, compared to business conditions for the same period, is pictured
in Chart I. Business conditions are represented by Babson's curve of
business activity. With the exception of the year 1917, the increase
in turnover is gradual from year to year. In this year there is a
slight decrease in turnover as compared to the preceding year, but
still an increase over the year 1915.
On first thought, it seems logical that the curve
representing sales turnover conditions should be the reverse of that
representing business activity. As business improves, it would seem
that the turnover among sa lesmen should grow less and when business
is retrogressing turnover should increase. In timos of business
activity a seller's market prevails, sales are easily made, and men
of lesser ability can hold their jobs. In times of depression, with
sales difficult to make, the'se poorer men are dropped and swell the
turnover percentages.
Chart I, indicates however, that for the period vinder con-
sideration, turnover conditions in sales organizations were not the
reverse of business conditions. Despite the business activity of
the years 1915 to 1919 inclusive, turnover continued to increase.
This whole period was abnormal, perhaps, due to war conditions.
One company reports that 37 per cent of the salesmen vrho
left their employ in 1917, entered military service. Another company
reported that 19 per cent of the salesmen employed from 1914 to
November 1918, left to go to war.
â– â– r/tf;n
-Oaou JIO
i. i'
»€>*»*
^a>-3j-»j
'OJ»aA<
-9-
During 1919, although the war v/r,s over and husiiiess vms
booming, salesman turnover continued to increase. The demand for
salesmen was very great and the competition bctveen companies for
their services caused then to jvanp from one organization to another
as larger compensation was offered. For this reason, largely, turn-
over continued to increase despite business activity.
Bineteen-tventy and 1921 rrero yoars of business depression.
During this period salesman tvirnover increased rapidly. Sales were
difficult to make and those salesmen who had only the period of the
war as z.' Ijcckground of experience covin not meot tho conditions
that existed. Many of the older salesmen who had become spoiled by
the "easy sales" of tho preceding years could not accustor. them-
selves to the now state of affairs and were dropped also.
The largest cut in sales organizations came in the early
part of 1922. At the beginning of the year most sales organizations
adopted an aggressive policy and djefinitely "cleaned house". As ona
sales manager expressed it "We will continue to have a^ large turnover
vmtil we have built up oti r organization to one thousand salesmen who
will be hardboiled and -^ reterans" . This aggressive policy ie reflected
In the 1922 figures.
The turnover figures for the first part of 1922 are undoubt-
edly higher than the figures for the entire yaar will be. Fron
present indications it seems that the turnover ratio for 1922 will
not graatly exceed thit of 1921 v'hile the* turnov-r for 1?£5 -".ill pro-
bably be much loss. In other -vords, general salesman turnover con-
ditions are improving.
-10-
COMFiiRISOK OF TURrO^ /ES AI'.IOFG ORG;^ N IZATIONS
HiV.iDLING DIFFSREPT PRODUCTS
The Rverage t\a:nover of a large raunber of sales orsani-
zations is indicative of general tendencies in salosjuan turnover,
but of little value in considering the turr.cfror In ary particular
line of businoss. Organizations handling sir.il?>.r products are
affected in the same manner by general conditions and, other things
bsing equal, should tond to have the sano percentage of sales tvirn-
over .
T.ie av3r"^ge salceraan tjrnover in -. Ir.rgo nunbcr of sales
organizations, affords a means by v;hich sales executives may chock
up their orm turnover. If the turnover for any one company groat ly
exceeds the average for a largo number of organizations handling a
similar product, bhe causes for such high turnover must be sought
within the companv itself.
The turnover in the various companies, listed eccording
to product sold, is given in Table III. Such ferf companies handling
any one t^rpe of product reported, 'ath the exception of office
spocialties, that the percentages, althotigh interesting, do not
indicate general tendencies in these types of aganizations.
-11-
TABLE III
PERCEKTAGE OF TIffiKCfVER FOR V.^RIOUS TYIES CF COMPAHIBS
«,^^_ — >— — «.„-^ __^.-__
'â– "T*
''trp^ of Product
No. of
: Handled
Companies
1915
1916
•1917
:1918
.1919
:1920:"
1921:'
1922:
Bo'>.l Estate
1
333
400
275
; 685
425
: 500:
525:
515:
Oil Corapany
1
232:
108:
Rub'jer Products
1
: 60
. 80
50:
170:
189:
Cindy
1
83:
68:
89:
Roofing Products
1
. 59:
93:
52:
Automobiles
3
45.
• 81
â– 57
62
. 56:
69:
111:
Office Specialties
10
15
SC-
30
: 52
60
58:
70:
88:
Soap
1
38"
SI
. 51
: 47.
69
: 80:
58:
51:
31oetrical Supplies
1
• 38
: 44
60
: 62-
37
57:
57:
47:
Fncking House Products
1
: 48
50:
42:
60:
CoKimercial Feed
1
22:
39:
53:
Life Insurarca
3
', 33
67
61
: 34
42
. 38:
35:
40:
Heatiiig tquipnent
2 :
0-
12
25
: 57-
15
17:
27:
41:
Women's Clothing
1 :
271
16
22-
18
11:
31:
50:
tJovslties
1
is!
21
20
. 18
38
9:
17:
33:
Drugs
2
5
10
9
12:
20
27:
25:
23:
Clay Products !
2
0.
0.
17
• 31:
15
18:
22:
35:
Food Products !
1
44:
43:
40:
Vacuum Cleaners
1
166:
230:
Several companies estimated their turno-rer figures. These
figures are shorm in Table IV. Estimated figures are, as a rule, too
lovr since such companies tend to minimise the impor-csnce of turnorer.
Such estimates are presented for twenty-one companies representing
fourteen different lines of business. A comparison of the figures
vrith those for similar lines in Table III shw? that the estiiuates are
very lo'.^s on the averag:c. Al•^ogotbor the figures do not form rv basis
for reliable conclusions.
-12-
TABLE IV
SSTDiATED TU!?WC!7ER W FOURTEEN LINES
Per Cent
Type of Product Turnover
Poriod Covered
Storaso Batteries 50
Past two years
M.:;chin^ry SO
1921
Aluminum Goods 48
Past year and r. hf.lf
Ins\irance 44
1921
Weighing devices, scales, etc. 25
Annual figure
Sashes and doore 25
Annual figure
Electrical appliances 20
Annual figure
Corsets 15
Annual figure
Hard-vare 12'i-
Annual figure
Slectrical vehicles, hoists, etc. 10
Annxial figxire
Metal bods 8
Annual figure
Buttons 5
Annu?il f igtire
Corrugated Boxod 3 to 5
Annual figure
Flour 3 to 5
Annual figure
COfcPARISQg OF TUHNOVEg IN OFHCE SPECIALTY 0RG.M:IZATI0NS
-^ wifa tufjover in all other busii-esses
Ten office specialty organizations furnished their turn-
over data. From these data accurrto concluBions can bo drawia as to
turnover conditions in this field. Table V compares the annual
average percentage of turnover for office epocialtics vith the turn-
over in all other lines represented in Table III,
TABLE V.
COMPARISON OF TUI^NOVER M O FfflCE SESCIILTY 3AL5S ORGANIZATIONS
TO TURNOVER IN OTHER SALES Ot^GAivIZATIOWS
Type
•1915
1916
1917-
1918
1919
.1920
1921
1922;
Office Specialty
15
30
30 .
52
60 •
58 :
70
88 :
Other businesses
47
54
51
t 51
. 55
61
• 80
. 83 :
-13-
This snme information is pictured in Ohart II. Durin^; 1915,
1916, and 1917, turnover â– 'ms considerably higher in all other businssess
than in office specialty organiznticns. In 1918, however, turnover in
the office applicance group increased considerably and fro-.Ti then on
remained as high or higher than tho turnover in other businesses.
During the period of prosperity it seems evident that many
companies purchased all sorts of office applionces, vhereas in less
prosperous tines they ivould not hr.ve done so. The demand for office
specialties caused expanded sp.les forces as vroll as expanded mantxfactur-
ing facilities. The fact that e;:pansion in office specialty organi-
zations was much more rapid than in organizations handling other tj'pes
of products, is illustrated in Chart III.
The chart shows the expansion of office specialty sales
organizations and all other organizations in terms of their size in
1920. The year 1920 is used as a base since sane of the companies did
not report figures previous to this time.
Rapid expansion by office specialty companies naturally
resulted in hi^ sslesman ti'mcver. Salesmen were scarce and many
men were hired who could not become salesrtien. This T/as true of
organizations selling other types of product but more trxie of office
specialty organizations.
Figures for individual office appliance ccmpanies fur.iish
some interesting variations from the general averrges f;iven above.
Such figures are given in TablG VI.
w
c
-f-
N
(^
c
c
u
o
O
-5
CO
c
a
u
cr>
(O
^
O
D-
+-
w
(y
CO
W(^
t cu:
+-
*•
o.-
!P
5§
s-
j:
V
X
h
-b-
>
^
^^