Multichoice Subscriber Management (Pty) Ltd and Tiscali (Pty) Ltd (72/LM/Sep04) [2005] ZACT 23; [2005] 1 CPLR 224 (CT) (20 April 2005)

60 Reportability
Competition Law

Brief Summary

Competition Law — Merger Approval — Conditional approval of merger between Multichoice Subscriber Management (Pty) Ltd and Tiscali (Pty) Ltd — The Competition Tribunal issued a Merger Clearance Certificate on 12 January 2005, allowing Multichoice to acquire Tiscali's internet access business, excluding its cellular telecommunications business — The merger was deemed to enhance competition in a stagnating market, with no adverse effects identified — The decision was based on testimonies from industry experts and the rationale behind Tiscali's divestment due to regulatory changes in Europe.

COMPETITION TRIBUNAL 
REPUBLIC OF SOUTH AFRICA
     Case No: 72/LM/Sep04
In the large merger between: 
Multichoice Subscriber Management   (Pty) Ltd  
and
Tiscali (Pty) Ltd
Non­Confidential Reasons for Decision
______________________________________________________________
APPROVAL
On 12 January 2005 the Competition Tribunal issued a Merger Clearance  
Certificate   conditionally   approving   the   merger   between   Multichoice  
Subscriber Management   (Pty) Ltd (Pty) Ltd and Tiscali (Pty) Ltd (Tiscali)  
in terms of section 16(2)(b). The reasons for the approval of the merger  
appear below.
Transaction description
i.The   acquiring   firm,   Multichoice   Subscriber     
Management   Services   (‘M­Web”),   is   a   subsidiary   of  
M­Web   Holdings,   whose   ultimate   owner   is   Naspers  
Limited.    1       
2. M­Web   is   acquiring   the   entire   issued   share   capital   of   Tiscali   from  
Tiscali International BV.   The internet access business of Tiscali will,  
post­merger, be incorporated into the internet access business of M­
Web.   The   Tiscali   Cellular   Mobile   Telecommunications   business   is  
excluded from this transaction. 2
1  Although M­Web is the name of Multichoice’s holding company, it is the brand and business  
unit  that is relevant for the purpose of this transaction, so we will for convenience refer to the  
acquiring firm as M­Web
2  In a separate transaction we have approved of the sale of Tiscali’s cell phone business to  
Vodacom. See case number 87/LM/Oct04. 
1

Hearing 
3. The hearing was held on 12 January 2005. We heard testimony from  
Anthony Brooks, General  Manager of the Internet Service Providers’  
Association   (“ISPA”),   who   was   called   as   an   expert   witness   by   the  
Commission, Chiwashira from G ­Soft who had noted objections to the  
merger,   Kim   Reid,   CEO   of   M­Web,   and   Diego   Massidda,   CEO   of  
Tiscali, who testified on behalf of the merging parties.
Rationale 
4. Tiscali’s Italian based parent decided in 2004 to sell its South African  
subsidiary. The reasons for this decision, we were told, had nothing to  
do   with   conditions   in   the   South   African   market   ­   the   business   was  
considered to be doing well ­ but were rather rooted in changes in the  
regulatory   environment   in  Europe.   As  a   result   of   these  changes   the  
group   had  decided  to  invest  more  heavily  in   its  primary   market   and  
needed to generate cash to fund this. It decided to sell its South African  
investment and called for offers.
5. After a bidding process it decided to sell the internet business to M­
Web and the cellular business to Vodacom.
6. M­Web saw an opportunity to increase its subscriber base in a market,  
where growth appears to be driven more by acquisition than organic  
growth. Both parties consider that the market is stagnating presently. 
Merging parties
M­Web
7. M­Web   was   launched   in   October   1996   by   the   acquisition   of   an   ISP  
from the CSIR. It focussed initially on selling technology as a consumer  
product   with   an   emphasis   on  handholding,   support   and  reliability.   In  
1998   it   bought   the   I­Africa   ISP   business   from   UUNet,   comprising  
approximately   60   000   subscribers.   In   1999   it   acquired   another   ISP,  
Netactive, with a consumer base of approximately 31   000 subscribers.  
By 2000 its subscriber base had reached approximately 250 000. Since

By 2000 its subscriber base had reached approximately 250 000. Since  
then,   M­Web   maintains   that   it   has   lost   subscribers   due   to   keen  
competition, especially from ABSA and Telkom.
8. Early   in   2004   it   launched   a   product   called   “Polka”,   a   lower   priced  
internet access brand, in order to increase its subscriber base.  
Tiscali
2

9. In 1998 Vodacom and World Online, a Dutch company, started an ISP  
in which each had a 50% stake. The joint venture grew by acquiring  
many subscriber bases. Historically Tiscali had always been number  
two in terms of size after M­Web. 
 
10. In   2000   Tiscali,   an   Italian   company,   bought   a   60%   interest   in   the  
company, buying out the Dutch based parent and part of Vodacom’s  
stake. Nevertheless it retained the rights to the brand and continued to  
trade as WorldOnline.  In 2001 Tiscali bought the remaining 40% from  
Vodacom   and   became   the   sole   shareholder.   In   2002   the   company  
acquired   a   relatively   small   ISP   called   Netactive,   focussed   on   the  
corporate market. That allowed the company to increase its presence  
in   the   corporate   market,   especially   in   leased   lines   dedicated  
connections. The re­branding of WorldOnline to Tiscali only started in  
2003, being completed in January 2004.  3
Identifying the markets
11. The parties   provide the following services:
M­Web Tiscali
Internet access including  
leased line
Internet   access  
including leased line
Local access Cellular   mobile  
telephony 
E­commerce
Subscriber management
Market analysis
12. Both merging firms as we have seen from the table provide a service  
that   enables   consumers  to  access  the  internet.   The  merging  parties  
would have it that this is where the boundaries of the relevant market  
are to be found, and that it would be wrong to segment it further on  
some more narrow definition. Indeed they go as far as suggesting that  
even   defining   a   market   as   one   for   internet   service   provision   is   a  
concession to narrowing rather than broadening the market.  Thus in  
their filing they state:
“   A   narrower   definition   of   the   market   (that   is   narrower   than  
telecommunications services) is the provision of access to the Internet  
3  Record page 1155
3

(or Internet connectivity) generally.” 4
13. They   go   on   to   assert   that   further   segmentation   on   the   basis   of   the  
licence   required   or   the   technology   used   to   provide   access   is  
unwarranted as these are “ technical details not integral to the service  
being provided.” 5
14. The Commission has distinguished between a corporate access and a  
dial­up access market. This segmentation is not specifically motivated,  
but   appears   to   arise   from   the   parties   own   business   practice   in  
regarding these as distinct services with distinct customers. What this  
amounts   to,   despite   the   labelling   offered   by   the   Commission,   is   a  
segmentation   premised   on   customer   needs   and   profiles.   Large  
customers, who the Commission locates in the leased line market, are  
corporations with high numbers of users willing to incur the costs of this  
kind   of   service.   Dial­up   customers   on   the   other   hand   are   largely  
consumers who operate from homes and comprise both families and  
small businesses.
15. We will now consider if the Commission’s delineation of the market is  
appropriate or whether we should segment it in some other fashion, if  
at all.
16. Internet   service   provision   is   about   getting   the   user   access   to   the  
internet via a range of technology and intermediaries to the networks  
that carry the data from one consumer of services to another.
17. A consumer wanting to access the internet has a range of technical  
options to choose from which are disparately priced. What influences  
the choice finally made is not only the consumer’s propensity to pay,  
but also the quality and quantity of access required.
18. The simplest form of access and the most pervasive at the moment is  
called   ‘dial­up’   access.   The   consumer   accesses   the   internet   via   a  
telephone through a modem. The consumer gets access by having a

telephone through a modem. The consumer gets access by having a  
contract,   pre­paid   or   subscription   with   an   internet   service   provider,  
such as one of the merging firms. Note that dial­up access is defined  
not by the purpose for which it is used i.e. for private or office use, but  
the means of access. 6 The disadvantage is that access is not constant  
and   the   consumer   must   dial­up   to   access   the   service   and   then  
terminate it each time it wants to connect to the Internet.
4­ Record page 43. 
5  Record page 43. 
6  See  for instance “Internet Access in South Africa, 2004, An annual study of the Internet  
access market in South Africa”,  by World Wide Worx (Pty) Ltd, ( Research led by Arthur  
Goldstuck.), page 37, where the authors say that for the purpose of their research that  “No 
distinction is made between dial up accounts at homes and offices.”  
4

19. The service is limited to one user at a time and is considered slow. Its  
advantages are price and technical simplicity.
20. The   next   stage   up   is   the   ISDN   line   which   is   also   used   mainly   by  
households and small businesses. ISDN stands for Integrated Services  
Digital Network. This is a digital service between a customer’s home  
and   the   dial­up   telephone   network. 7  Both   Tiscali   and   M­Web   again  
provide this service. Tiscali’s pricing at present is more competitive but  
the services are not directly comparable and vary in relation to the size  
of the mailboxes for e­mail and the number of e­mail addresses per  
subscription.  8
21. ISDN is more expensive but faster than analogue dial­up, according to  
BMI.9  Subscriber   figures   indicate   that   this   service   is   much   less  
pervasive than analogue. Tiscali’s figures indicate that it has about one  
tenth the number of ISDN subscribers than analogue dial­up.
22. The   next   mode   of   access   is   ADSL   (Asymmetric   Digital   Subscriber  
Line).   This   technology   allows   the   user   to   send   faxes,   talk   on   the  
telephone and surf the internet simultaneously.  The advantage of the  
service is that it is always on with no dial­up necessary. The service is  
charged at a fixed monthly fee. Fees for ADSL vary according to the  
size of the gigabytes offered. However pricing even for the lowest is at  
a premium to that of the most expensive analogue service.
23. At   the   next   level   is   the   leased   line.   Here   the   customer   is   given   a  
dedicated line to an internet service provider. This again is on all the  
time  but   allows  for   use   by   a  large  number  of   users   and  thus  is  the  
preferred   method   of   access   for   large   organisations   with   a   need   for  
many users. Not surprisingly this is the most expensive service and is  
not used by households or small businesses.

not used by households or small businesses.
24. Other more ambitious forms of internet access exist as well, albeit that  
they are very much in their infancy at the time of this decision. We are  
told that Sentech offers wireless broadband service which users will be  
able to access by carrying a modem in their pockets and connecting  
from   anywhere   that   has   reception. 10  But   at   the   same   time   it   is  
suggested that there have been significant delays in the roll out of this  
technology.11 Mr Brooks, the Commission’ s expert, made much of the  
fact that from 1 February 2005 there would be significant deregulation  
in the industry following an earlier announcement to this effect by the  
7  BMI report record page 801
8  See record page 1168.
9  BMI report page 801
10  See World Wide Worx op cit record page 1031
11  See World Wide Worx  op cit record page 1031
5

Minister. He predicted that this would lead to an expansion of business  
for existing private sector providers who would no longer be dependant  
on having to use Telkom’ s infrastructure, as they were obliged to do by  
law up till now. We were also informed that the mobile cell companies  
had targeted internet access as basis for business expansion, and that  
access could come from a cell­phone.
25. What all these methods have in common is that they provide a means  
of access to the internet, but this does not mean that they constitute  
substitutes for one another for the purpose of competition analysis. 
26. Now   it   must   be   noted   that   firms   that   provide   access   to   the   internet  
typically offer more than one of these modes and that the merging firms  
are no exception in this respect.
27. We   have   up   till   now   examined   the   demand   side   of   the   market.   But  
delineating the correct antitrust market also requires an examination of  
the supply side. Here again the task is no easier. In the first place the  
market consists of various tiers. 
28. At the bottom of the structure is Telkom who to date have enjoyed a  
monopoly over the supply of infrastructure a situation that may change  
with   deregulation.   Telkom   sells   bandwidth   to   a   range   of   firms   who  
operate   largely   as   wholesalers   of   this   bandwidth.   These   firms   are  
sometimes referred to as first tier firms. First tier firms then wholesale  
this bandwidth to other firms who in turn act as retailers in the sense  
that they sell internet access to the ultimate consumer. The merging  
parties fall into this latter category of retailer, also known as second tier  
firms.   There   is   also   a   category   of   firm   who   are   considered   by   the  
market to be third tier firms or virtual ISP’s who, in the sense that we  
understand this distinction at all, sell bandwidth to customers but are  
little more than marketers who depend entirely on the first tier supplier

little more than marketers who depend entirely on the first tier supplier  
for all the back­up including that of a call centre. For our purposes this  
distinction   is   academic   as   firms   in   the   so­called   third   tier   are   still  
competitors of the merging firms and other firms in the second tier. We  
have reflected this structure of the industry in the drawing below.
6

Diagram 1: Structure of industry 
29. But this model is far from rigid. Although Telkom is presently a sole  
supplier  to  the   wholesalers,   it  is  also   a  competitor   itself   in  the   retail  
market and thus of the merging parties. 
30. Although the wholesalers are not in the retail market for dial­up access,  
they do sell directly to customers who want leased lines.
31. The   industry   is   thus   characterised   by   vertical   relationships   where  
suppliers   either   compete   with   their   customers   or   at   least   have   the  
potential to do so.
32. What distinguishes the merging firms as retailers is their access to a  
mass   customer   base.   This   means   that   they   have   to   concentrate   for  
their competitive advantage on customer service and marketing. This  
at present distinguishes the merging firms from their suppliers in the  
first tier, as the latter, whilst geared up to service a niche corporate  
market, are not at present competitive in the mass market. 
 
Conclusion on defining the relevant market
33. Both M­Web and Tiscali offer analogue and digital dial­up, ISDN, ADSL  
and   leased   line.   M­Web   currently   provides   a   limited   internet   access  
service via leased lines. Tiscali too, although the parties state that a fair  
TELKOM
Sell 
bandwidth
FIRST TIER  
ISP’S 
  UUNET, Internet Solutions,  SAIX  
and MTN Network Solutions
SECOND TIER  
ISP’S
  M­WEB, Tiscali,  Storm, Xsinet,  
@lantic 
THIRD TIER  
ISP’S
ABSA
=
=
Sell 
bandwidth
7

proportion of its customers use leased lines. 12
34. The   merging   parties,   as   we   observed   earlier,   commenced   with   an  
opening   observation  that   the   market   was   one   for   telecommunication  
services.   That   this   observation   was   barely   credible   seems   to   have  
been   tacitly   acknowledged   by   them,   as   they   proceeded   almost  
immediately with a concession that the market could be characterised  
as one for internet access. 
35. What we have to evaluate is whether this is an adequate description or  
whether the market may be further delineated. As the discussion above  
has   indicated,   user   requirements   differ   vastly   depending   on   the  
sophistication of the service required and this is reflected not only in  
the   choice   of   different   technologies,   but   also   in   their   prices.   This   is  
further   reflected   in   the   way   both   merging   firms   have   marketed   their  
products at consumers.   We have had access to the business plans  
and marketing materials of both the merging firms. The home­based  
consumer,   whether   a   private   household   or   a   small   business,   is  
identified as a distinct segment of the broader access market. These  
consumers are overwhelmingly in dial­up access, to a lesser extent in  
ADSL and even less ISDN. This is not surprising given that the more  
sophisticated the service, the greater the price to the consumer. But it  
is not possible to delineate the relevant market on the basis of choice  
of technology alone. Rather it is the nature of the consumer and the  
type of service they require.  13
36. For  this  reason we suggest,  as a point  of departure,  that  there is a  
distinction between a home­based market and a corporate market.
37. Although the Commission has not followed this approach, instead, as  
we noted above, using the technology utilised to segment markets into  
a   leased   line   and   dial­up   access   market,   these   segments   roughly

a   leased   line   and   dial­up   access   market,   these   segments   roughly  
correspond   to   the   division   between   a   corporate   and   home­based  
consumer market that we have preferred.
38. Leased lines  are far  more  expensive to  operate than other forms of  
access   particularly   dial­up   access,   and   are   typically   only   used   by  
corporates or large institutions. On the supply side this market looks  
different as well. 
39. Support   for   this   distinction   also   comes   from   industry   commentators.  
12  Record page 46
13  Figures show that as at 2004 at least 88% of Tiscali customers utilise dial–up access. See  
Annexure E Subscribers Warranted. M­Web did not provide figures but stated too that a large  
proportion of their customers utilised dial­up access.
8

According to World Wide Worx;
“The   corporate   market   split   off   almost   entirely   form   the   dial   –up  
services   arena   with   the   two   major   ISP’s.,   Internet   Solutions   and  
UUNET,   focussing   entirely   on   the   former.   They   remain   the   market  
leaders.”14
40. This highlights an important distinction between the corporate and the  
home­based market. The leading firms in this former market are not the  
merging  parties,  but  their  suppliers.  Indeed,  according  to  the  parties  
they would post merger only have a [less than 10%] per cent share of  
this market. For this reason if we assume that the corporate market, or  
in the Commission’s estimation the leased line market, is a separate  
relevant market the merged entity is not post merger likely to be able to  
behave anti­competitively given its low market share. We need not then  
consider this market further.
41. It remains for us to consider the home­based segment of the industry.  
Let   us   assume   that   in   the   home­based   segment   at   present,   dial­up  
access is predominantly the technology of choice.   Does this exhaust  
all   the   possible   ways   to   delineate   this   segment   for   competition  
purposes?   There   is   some   suggestion   that   there   may   be   a   further  
relevant   market   distinction   in   the   home­based   segment   between   a  
premium and non­premium service. This is one of the issues that we  
explored   during   the   hearing,   because   if   there   is   a   distinct   premium  
market,   the   merged   firms   would   be   its   only   significant   players   and  
hence the merger would lead to a near monopoly.
42. The emergence in the market of a distinction between premium and  
non­premium offerings was a response to new entry into the market by  
Absa and Telkom as low cost providers.
43. Absa caught the market by surprise when it entered in 2001 by offering  
free internet access. The response of consumers was overwhelmingly

free internet access. The response of consumers was overwhelmingly  
favourable.   However   the   free   offer   did   not   last   long   and   soon   Absa  
began   charging   its   customers   for   access,   discriminating   in   price  
between those who were its banking customers and those who were  
not. Absa’s strategy was to use its internet offering not as a means to  
get   into   the   internet   business,   but   to   get   customers   for   its   banking  
business. The Absa offering was a no­frills one that was cheaper even  
for the non­Absa customer than the services offered by the traditional  
incumbents, M­Web and Tiscali. 
44. Telkom’s entry into the internet access home­based market should not  
have come as a surprise. Indeed the surprise is that it had taken so  
14  See World Wide Worx op cit record page 1040.
9

long   for   it   to   do   so.   Given   Telkom’s   dominance   of   every   tier   of   the  
access edifice it is baffling that it realised so late in the day of the value  
of being in the second tier. The Telkom offering, marketed through its  
access to fixed line consumers, was also priced significantly below that  
of the merging firms.
45. Thus faced with a twin­prong attack on their traditional market by two  
firms with deep pockets, and access, through their related businesses,  
to a large customer base, M­Web and Tiscali were forced to respond.  
Interestingly, both responded in the same way. They first introduced a  
low priced service to be competitive with Telkom and Absa.  M­Web’s  
offering is separately branded with no apparent link in the name and  
marketing to the M­Web business. Polka, as the product is known, has  
only   recently   been   introduced   and   at   this   stage   it   is   too   early   to  
evaluate its success. 
46. Tiscali’s also other, simpler products with less “bells and whistles” at a  
price of R79.00 ­ R 90.00 per month, but this was not offered to the  
public   at   large,   but   only   to   select   corporate   client’s   customers   who  
could only have access to the cheaper option, by virtue of being that  
corporate’s client. 15
47. Of course in developing a lower cost service the firms risked having  
their clients on the more expensive packages migrate downwards. To  
avoid   this   scenario   their   second   strategy   was   to   differentiate   their  
offerings  between  premium  and  non­premium.   The  premium  offering  
contains more features than the non­premium product and these range  
from access to free content such as educational material news sites  
and dating services, to additional e­mail addresses, larger data storage  
capacity   and   greater   safety   features.   Whether   this   distinctiveness   is  
sufficient to justify the price premium is not something we are called

sufficient to justify the price premium is not something we are called  
upon   to   decide.   What   is   of   interest   to   us   is   whether   the   product  
differentiation  has   been   successful   enough   to   justify  segmenting   the  
relevant market between premium and non­premium. 
48. According to World Wide Worx;
“The number of dial­up subscribers in South Africa reached 782 000 by  
the  end  of  December 2000,   and  the dial   up user  base  consolidated  
behind   M­Web   and   WorldOnline,   [now   Tiscali]   which   both   began   to  
raise their fees, partly in order to position themselves as high­quality  
ISP’s   with   extensive   value­add,   and   partly   because   it   had   become  
apparent that subscribers were not willing to pay separate additional  
amounts for additional services.” 16
15  Transcript page 83
16  Word Wide Worx op cit , record page 1039.
10

49. The table below compares the rates of these particular services.
Current Monthly Charge of the Various ISPs
ISP Rate Per Month
Telkom Internet R79
XSINET R90
M­Web’s Polka R75
Tiscali Lite R79
ABSA R39   (ABSA   clients),   R65   for   Non­
ABSA clients
@lantic R65
50. The cost of the premium services offered by Multichoice and Tiscali is  
R145 per month.
51. Note that the cost to the consumer comprises the monthly charge to its  
ISP and the per minute call rate charge to Telkom.
52. The   parties   argue   strenuously   in   their   competition   filings,   and   thus  
contrary   to   the   thrust   of   their   marketing   materials,   that   this   market  
segmentation   is   not   justified.   They   argue   that   there   is   still   a   single  
market for internet access and that if the consumer is paying too much  
for  a  premium  service  she  will  migrate to  a  cheaper  one.  With  both  
Absa   and   Telkom,   inter   alia,   waiting   in   the   wings   to   take   on   the  
dissatisfied   consumer   the   premium   customer   is   not   to   be   taken   for  
granted by them and has choices if the premium does not justify its  
value. 
53. But   the   evidence   would   suggest   that   the   act   of   changing   service  
providers is less ubiquitous  than the  merging  parties would have  us  
accept.   One   of   the   major   reasons   it   appears   that   consumers   are  
reluctant to lightly change their existing service is that like telephone  
numbers,   e­mail   addresses   are   not   portable.   The   price   conscious  
consumer wanting to change to a cheaper supplier must factor in the  
inconvenience of a changed email address. 
54. Whilst   Mr   Reid   of   M­Web   did   not   consider   this   an   insurmountable  
obstacle   to   changing   one’s   provider,   his   company’s   own   practices  
suggest otherwise. When M­Web has acquired other businesses it has  
allowed   consumers   to   retain   existing   e­mail   addresses.   In   the   sale

allowed   consumers   to   retain   existing   e­mail   addresses.   In   the   sale  
agreement with Tiscali it has provided for the right to use the Tiscali  
name for [confidential ] years before it will have to migrate customers  
11

on   to   another   name.   Prior   to   the   conclusion   of   the   sale   M­Web’s  
anxiety   to   secure   this   right   is   manifest   in   its   written   offer   to   Tiscali  
where Mr Reid states: 
“For this reason it is imperative that M­Web retains the Tiscali.co.za  
and   wol.co.za   domain   names   on   the   terms   set   out   in   the   Sale  
Agreement in order to minimise the risk of losing subscribers. It is our  
experience that e­mail domain related cut over comes with significant  
churn risk and that in the event that we are not able to manage such a  
risk   in   the   manner   proposed,   this   will   have   a   material   effect   on   the  
retention of subscribers and therefore the consideration offered.” 17
55. The M­Web strategy document indicates that Polka consumers will be  
allowed   to   migrate   upward   and   get   the   premium   service,   whilst   still  
retaining their Polka e­mail address, but that the opposite will not be  
allowed – if M­Web customers want to migrate to Polka they forfeit their  
e­mail address. This strategy appears to be an acknowledgment of the  
‘sticky’   quality   of   the   e­mail   address.   M­Web   lets   you   keep   your  
address if you migrate upwards towards the more expensive service,  
but it is less accommodating if you want to migrate downwards.
56. The   Commission’s   expert,   Mr   Brooks,   also   suggested   that   there   is  
inertia amongst consumers to change service providers, although he  
attributed   this   more   to   a   reluctance   to   change   stop­orders,   than   a  
reluctance to lose an address. 
57. Yet   this   churn   inertia,   whatever   its   provenance,   is   not   one   which   is  
exacerbated by the merger. Even if the merging parties were to raise  
the price of the premium service, it is likely that consumers, once they  
have come to the realisation that they may have to forfeit their address,  
would as easily opt for a non­premium service as a premium one. Thus

would as easily opt for a non­premium service as a premium one. Thus  
the existence of an independent Tiscali is not decisive in constraining  
an exercise of market power by the merged entity, as once a consumer  
has decided to churn it is as likely, given the questionable value of the  
premium   product,   that   they   would   churn   ‘downwards’   rather   than  
‘sideways’. For this reason, we believe that the merger itself would not  
contribute to the lack of portability already in the market, because each  
provider has some relational dominance in relation to its customer and  
that   dominance   is   a   function   of   the   value   of   the   inconvenience   in  
changing e­mail addresses,  not market power .  
58. This feature would suggest that at the time of this merger, a premium  
segment has not yet been carved out as a stand­alone relevant market  
for competition law purposes and that the market is one for   internet 
access by a home­based consumer.
17  Letter from Mr Reid to Evert den Hollander of Tiscali dated 21 June 2004.
12

59. That being said there is no doubt from all the materials that:
i.M­Web and Tiscali are the two largest firms in the market.
ii.Both   have   grown   more   by   acquisition   than   organically   and   that   this  
suggests that the scope for smaller rivals to grow organically is slight.
iii.Pricing   behaviour   suggests   that   they   behave   as   a   duopoly   and   a  
pricing move by one is followed by a response by the other. At present  
their   prices   are   identical   for   both   premium   and   non­premium   dial­up  
access. According to Mr Massidda from Tiscali when on one occasion  
they priced above M­Web they suffered the consequences and decided  
to keep pricing in line in future. It would appear that M­Web was the  
price leader in this relationship.
iv.Each regarded the other as its primary rival and that the merger leads  
to the elimination of M­Web’s most effective competitor.
v.If  the market is  a home­based market the  merged parties  combined  
share would be 34.4 % made up as follows:– 
Firm Market Share
M­Web 24.3 %
Tiscali 10.1 %
Telkom Internet 15.8 %
ABSA 14%
Atlantic Internet 3 %
XSInet 1.2 %
Internet Solutions 16 %
Other  15.4 %
Source: Parties’ figures given at hearing
60. Precise   data   for   this   share   is   nevertheless   not   available   and   the  
merging   parties   do   not   claim   that   their   knowledge   on   this   point   is  
definitive.   Nevertheless,   even   accepting   a   broader   home­based  
market, one not segmented into a premium and non­premium market,  
the   merging   parties   have   a   large   enough   share   in   a   concentrated  
market for it to raise some concerns.
Barriers to entry
61. It is trite that even if concentration levels are high in a market as long  
as entry barriers are low, a merger is unlikely to be anti­competitive as  
any   attempt   by   the   merged   firm   to   exercise   market   power   will   be  
countered   by   new   entry,   provided   that   entry   is   timely,   likely   and  
13

sufficient.  18
62. The   merging   parties   argue   that   entry   barriers   in   the   internet   access  
market are low for three reasons:
62.1 Firstly, new entrants may come from firms with access to a large  
client   base   of   their   own   who   see   the   provision   of   internet  
services   as   an   add­on   to   attract   clients   to   an   unrelated   core  
service.   In   this   regard   we   have   seen   already   the   successful  
entry   of   a   firm   like   Absa.   At   the   time   of   this   decision   we   are  
informed   that   Discovery,   a   medical   aid   scheme   and  
administrator, is offering internet services at a very low cost to  
its members.
62.2 Secondly, is the fact that this is an innovation market. The Act  
requires us in section 12A(2)(e) to recognise:
“   the   dynamic   characteristics   of   a   market   including   growth,  
innovation, and product differentiation.”
These are all features inherent in this market. We are advised  
that wireless internet access for the home consumer is around  
the corner, that the new  incumbent in the fixed line market is  
likely to target this market as an area for growth and that the  
three   mobile   phone   operators   have   plans   up   their   sleeves   to  
provide an internet access service as well. Whilst we must be  
cautious about the optimistic hype that surrounds the plethora of  
would­be but not yet proven entrants, the history of the market  
to   date   has   shown   that   entry   has   come   from   unexpected  
quarters   (e.g.   Absa)   and   that   new   technology   is   developing  
continually   and   with   it,   the   potential   for   new   entry   and   newly  
priced   offerings.   This   has   been   the   decisive   factor   in   our  
allowing the merger. Were this not a dynamic innovation market  
we might have found against the merging parties.
62.3 The third reason given to justify likely entry is its low cost. Both

62.3 The third reason given to justify likely entry is its low cost. Both  
the merging parties and Mr Brooks indicate that a firm can enter  
the   market   at   the   access   provider   level   without   great   capital  
outlay. Even if we accept that is the case, it does not mean that  
entry will be effective. G­Soft, the intervenor, indicated that this  
was precisely the problem and hence their interest in acquiring  
Tiscali.   The   history   of   both   parties   indicates   that   they   have  
grown   more   by   acquisition   than   organically,   and   that   post­
merger,   most   of   the   jewels   in   the   market   will   have   been  
18  See United States Department of Justice and Federal Trade Commission’s “Horizontal  
merger Guidelines, 1997.”
14

acquired.   Of   course   some   new   entrants   such   as   financial  
institutions have a built­in marketing advantage by their ability to  
mine their existing or potential customer basis. 
63. The brand new entrant without a client base may have to rely on new  
sign­ups  and  this  requires  access   to  the  retail   market   where  people  
purchase their new computers and software. 
64. For this reason, both Tiscali and M­Web have successfully tied up the  
major   computer   retail   outlets   with   exclusive   agreements.   Tiscali   has  
exclusive relationships with Game and Dions.  M­Web has an exclusive  
relationship   with   Internet   Connection,   which   clearly   has   a   large  
presence   in   the   retail   market,   operating   chains   in   the   major   centres  
around the country and described by the merging parties as one of the  
largest computer retailers in the country. 19   Kim Reid, the MD of M­
Web, admitted that Incredible Connection was their most important and  
had   been   their   most   effective   channel   to   market. 20The   Commission  
has   recently   received   a   complaint   from   a   small   computer   retailer  
complaining about the fact that the exclusivity impacts on his ability to  
compete in the sale of computers. As he cannot sell computers that  
come with free internet software of one of the merging parties as part  
of the deal, his offering, he alleges, is less attractive. 
65. The analogue of this complaint would be the difficulty of small internet  
companies   to   enter   the   market   for   sale   of   their   product   if   they   are  
precluded   from   the   major   retail   outlets.   For   this   reason   we   have  
proposed a condition that seeks to address this problem by requiring  
the   merged   firm   to   terminate   its   exclusive   distribution   arrangements  
with its retail customers. The merging parties were consulted on this  
condition and had no objection.

condition and had no objection.
66. The   other   potential   problem   the   merger   raises   concerns   the   buying  
power of the merged firm. UUNet, which along with Internet Solutions  
is   one   of   the   first   tier   firms   in   the   internet   market,   has   raised   this  
concern with us in a submission dated 4 January 2005. Its concern was  
that the merged entity could enjoy  “ extremely high buying power” . The  
submission   was   far   from   clear,   but   we   understand   UUNET   to   be  
concerned that the balance of power in the industry could shift from the  
first to the second tier. However UUNET ended its objection by stating  
that it has not raised an “ official objection to the merger going ahead  
but, that “ it is of the opinion that if it does go ahead the current balance  
of power in the market will be significantly disturbed.” 21
19  Transcript page 36­37
20  Transcript page 39
21  See Record page 70 –71.
15

67. It is difficult to pursue this concern further when the protagonist has  
raised   it   in   such   a   faint­hearted   manner.   Granted   that   M­Web   in   its  
internal   documents   does   put   forward   the   ability   to   negotiate   better  
terms with suppliers as one of the advantages of the merger. Since the  
input costs of its provider constitutes most of its input costs there is no  
doubt where the squeeze on suppliers will be exercised. However, it  
seems   unlikely   that,   given   the   structure   of   the   foundations   of   this  
industry that first tier firms such a UUNET will be effectively squeezed  
in a manner that may raise competition concerns. If the merged firm  
overreaches UUNET, so that it is no longer profitable for it to supply it  
at post merger prices, then it can enter the market itself at the second  
tier and compete with the merged firm. This threat is likely to restrain  
the merged firm’s exercise of market power as a purchaser of inputs.  
Recall that UUNET was once in this market with Iafrica.com before it  
sold that business to M­Web in return for the latter staying out of the  
first tier market. 
68. We   do   not   see   that   the   merger   will   lead   to   the   merged   firm   having  
monopsony power in respect of first tier firms .At present there are at  
least four firms in the first tier market. 22  
Vertical issues
69. As   outlined   above,   Naspers   is   the   ultimate   shareholder   of   M­Web.  
There   is   therefore   a   vertical   relationship   between   a   primary   internet  
service provider and a firm which is a significant content contributor to  
its   portal.   According   to   G­Soft,   content   provision   is   crucial   to   the  
success of any ISP. G­Soft raised the concern that the merger would  
give   M­Web  a  big  advantage  over  other  incumbents  in  the  industry,  
insofar as Naspers is a multinational media group with operations in,

insofar as Naspers is a multinational media group with operations in,  
inter   alia,   pay­TV,   print   media   and   publishing.   Naspers   for   instance,  
controls ­Net, MultiChoice, News 24, and prints and publishes Beeld,  
Volksblad,   City   Press   and   Daily   Sun.   It   thus   is   an   owner   of   a  
considerable amount of content, which, if considered a ‘must have’ by  
consumers, would give the merged entity a considerable competitive  
advantage. 
70. M­Web   has   explained   that   Naspers   offers   it   two   types   of   content  
product.   The   first,   of   which   news   on   Nasper’s   Media   24   site   is   one  
example, is what it terms ‘open content’. Although available on dial­up  
to its subscribers, the content can be accessed freely by anyone with  
internet access.
71. The second type it terms ‘closed content’. This content is only available  
22  See World Wide Worx op cit record page 1041.
16

to subscribers and is not open to view on the internet. An example of  
this   closed   content   is   Naspers’   A­to­Z   of   Diseases .   Although   this  
product is available to M­Web, Naspers does not sell it exclusively to  
M­Web and it gives examples of other firms, unrelated to Naspers, who  
have purchased this content for their subscribers. According to figures  
given by Naspers, M­Web has had to pay for this content. The figures  
also show that it is paying more than at least two other customers for  
the content.
72. Although Tiscali also offers content none of this is acquired exclusively  
by it and rivals are free to purchase similar content. 23 
73. The commission argued that any vertical concerns that arose out of the  
relationship between Naspers and M­Web, existed prior to the merger  
and are not merger­specific. However, there is a possibility that if there  
were foreclosure problems before, this merger could exacerbate these,  
since the merged entity would command a greater share of the market,  
therefore the foreclosure effects would be more pronounced.
74. Nevertheless   even   if   foreclosure   may   be   attractive   to   Naspers   post  
merger,   we   must   consider   if   it   is   likely.     The   Commission’s  
investigations revealed that of the 22 content products M­Web provides  
to its customers, only 6 are sourced from Naspers. Of those 6, not all  
are offered exclusively to M­Web.
75. Our conclusion is that the content:
i.is not of the ‘must have’ variety. Indeed with all due respect to the  
merging parties, it seems particularly lacking in that respect;
ii.is not exclusive to the merged firms presently ;
iii.is not unique.  There is a vast amount of similar content available to  
new entrants;
iv.does not play a decisive role in consumers preference for providers.
76. For all these reasons we concluded that the merger raises no serious  
vertical concerns. 
Better buyer issue
77. An objector to the merger, G­Soft, raised certain concerns about the

Better buyer issue
77. An objector to the merger, G­Soft, raised certain concerns about the  
merger. G­Soft is a new BEE entrant, which was one of the bidders for  
the   Tiscali   business,   and   indeed   at   one   stage   was   considered   the  
preferred bidder. It was ultimately rejected as the buyer in favour of M­
Web, it appears for funding reasons, although this remains a subject of  
dispute between G­Soft and Tiscali.
23  Record page 909
17

78. G­Soft’s   objections   were   to   the   effect   that   the   merger   with   M­Web  
raised   both   vertical   and   horizontal   concerns   and   that   if   it   were   the  
purchaser of Tiscali, as a new entrant, none of these concerns would  
arise.
79. We have already dealt with the horizontal and vertical issues and found  
that the merger would, if accompanied by the condition we require, not  
have an anticompetitive effect.
80. Although   G­Soft,   had   it   been   successful   as   a   purchaser,   may   have  
presaged a more competitive market than M­Web, this is not the test  
we  have  to  apply.  Our  task is not  to  indicate which  firm  might be a  
preferred   buyer,   but   only   if   the   merger   as   proposed   would   be   in  
violation of the Act. If the answer to the latter question is in the negative  
then the merger as proposed must be approved regardless of whether  
a better bride waits in the wings.
81. Although sympathetic to G­Soft’s obvious disappointment at not being  
able to conclude the deal, we cannot interfere with the merger on this  
basis.
Public interest –employment condition
82. We   have   imposed   a   condition   in   relation   to   the   retrenchment   of  
employees.     This   condition   arose   not   out   of   the   hearings   but   an  
agreement reached between the merged entity and their employees as  
part of their collective bargaining. For this reason we do not have to  
explain it any further.
83. In imposing this condition, we have followed the approach adopted in a  
previous   merger,   in   terms   of   respecting   the   agreement   reached  
between employees and employer. 24
Conclusion
We approve the merger subject to the conditions, which are set out in the  
annexure to this decision.
_____________ 20 April 2005
N. Manoim
24  Cherry Creek Trading 14 (Pty) Ltd and NorthWest Star (Pty) Ltd  ­  52/LM/Jul04
18

Date
Concurring: D. H. Lewis, M. Madlanga
For the merging parties:   J. Meijer, Cliffe Dekker Attorneys
For the Commission:  .   Mabusa,   K.   Moodliar,   M.   Langa,   Competition  
Commission
19