LMA8CCLu un1lL lL8 12, 2013 aL 8:30 am L1

MC8lLL SLA8CP AuvL81lSlnC A8Cunu 1PL CLC8L

Pow SmarLphones and 1ableLs Are Changlng ald Search













2013 Annual 8eporL

Moblle Search AdverLlslng around Lhe Clobe - 2012

age 2 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.




!"#$% '( )'*!%*!+


A1 A CLAnCL ............................................................................................................................................... 3
ln18CuuC1lCn ............................................................................................................................................ 4
A81 l - unl1Lu S1A1LS ............................................................................................................................... 6
MC8lLL SLA8CP 18LnuS ............................................................................................................................... 6
8LnCPMA8klnC SLA8CP L8lC8MAnCL AC8CSS uLvlCLS ................................................................................. 7
A81 ll - unl1Lu klnCuCM Anu 1PL Lu8CZCnL .................................................................................... 10
MC8lLL SLA8CP 18LnuS ............................................................................................................................. 10
8LnCPMA8klnC SLA8CP L8lC8MAnCL AC8CSS uLvlCLS ................................................................................ 11
A81 lll - MC8lLL SLA8CP AC8CSS 1PL CLC8L ................................................................................................. 13
C1lMlZlnC MC8lLL SLA8CP CAMAlCnS ..................................................................................................... 13
A8Cu1 MA8ln SCl1WA8L ........................................................................................................................... 17




Moblle Search AdverLlslng around Lhe Clobe - 2012

age 3 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.


"! " ,$"*)%

We llve ln a world LhaL's geLLlng more connecLed by Lhe mlnuLe. 1he rapld prollferaLlon
of smarLphones and LableLs ls changlng how consumers lnLeracL wlLh brands, consume
medla and make purchase declslons. AdverLlsers who undersLand Lhe broad lmpllcaLlons
of Lhls Lrend and adapL Lhelr markeLlng sLraLegles wlll ulLlmaLely wln Lhe baLLle for Lhe
consumers' hearLs, mlnds and walleLs.
DIGI1AL ADVLk1ISING A1 A CkCSSkCADS
1he lnLroducLlon of Coogle AdWords ln 2000 revoluLlonlzed Lhe search buslness by Lylng
adverLlslng spend Lo purchase lnLenL. 1he emergence of hlgh-powered moblle devlces ls
lncubaLlng anoLher revoluLlon: one LhaL wlll enable hyper-connecLed consumers Lo
search, buy and share anywhere, and aL any Llme.
IAS1Lk 1nAN A SLLDING ALGCkI1nM
Moblle adopLlon ls happenlng along an exponenLlal growLh curve. Whlle users rellsh Lhe
move Lo smarLer moblle devlces, Lhe rapld pace of change ls glvlng adverLlsers pause as
Lhey wresLle wlLh fundamenLal quesLlons. ls Lhls a fad? Pow much budgeL should be seL
aslde for Lhese emerglng channels? Should Lhls lnvesLmenL be neL new or dlsplaced?
ADA1ING 1C AN ALWAS-CN WCkLD
MarkeLlng funnels ln LexLbooks look very dlfferenL. ln Lhe real world, search, soclal and
locaLlon are lnLerLwlned from Lhe polnL of awareness Lo purchase. Moblle devlces, ofLen
aL Lhe hearL of Lhe modern shopper's buylng process, provlde lnsLanL access Lo producL
revlews, coupons and compeLlLors. 1o Lhrlve ln Lhls always-on world, markeLers need Lo
geL beLLer aL qulckly dlsLllllng volumes of daLa lnLo key lnslghLs and correcLlve acLlon.


Moblle Search AdverLlslng around Lhe Clobe - 2012

age 4 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
-*!.'/0)!-'*

ln Lhe few shorL years slnce Lhe arrlval of Lhe Apple lhone ln !une 2007,
smarLphones have become an lnLegral parL of how people relaLe Lo
brands, consume and share conLenL, and conducL commerce. So much so
LhaL Loday, many of us can'L lmaglne llfe wlLhouL our smarLphones. 8uL
Lhe rlse of smarL moblle devlces lsn'L [usL abouL Angry 8lrds or endless
sLorles abouL !"#". AL lLs core, moblle represenLs a paradlgm shlfL ln dlglLal
adverLlslng.

Accordlng Lo a research reporL from Cowen
1
, Coogle earned $2.3 8llllon ln moblle revenue ln 2011. And
by 2016, lL's esLlmaLed LhaL moblle could drlve $20 8llllon ln revenue for Lhe search glanL. 1he
exponenLlal growLh ln moblle adverLlslng largely Lracks consumer adopLlon Lrends. SmarLphones now
have a greaLer Lhan 30 peneLraLlon ln Lhe uS
2
moblle phone markeL. And moblle adverLlslng lsn'L [usL
llmlLed Lo phones. ln facL, sales of moblle devlces (phones and LableLs) runnlng Coogle Androld are
expecLed Lo have ecllpsed 'C' sales ln 2012
3
. And consumers worldwlde have already boughL more Lhan
a bllllon lCS and Androld devlces
4
.

1he emerglng world of Lhe moblle-enabled consumer ls already qulLe dlfferenL from Lhe pre-smarLphone
world. ln Lhls new reallLy, Lhe walls beLween onllne and offllne commerce are crumbllng. Case ln polnL:
Loday's consumers rouLlnely check producL revlews and compare prlces onllne whlle Lhey're ln a brlck
and morLar sLore.

WlLh smarL moblle devlces becomlng ublqulLous, adverLlsers need Lo adapL Lhelr onllne markeLlng
sLraLegles Lo more closely reflecL how people search, buy and share. Speclflcally, markeLers wlll need Lo
learn how Lo communlcaLe wlLh an always-on audlence, Lhlnk harder abouL allocaLlng ad budgeLs across
devlces, and develop beLLer aLLrlbuLlon models Lo accounL for Lhe fadlng boundarles beLween onllne and
offllne commerce.

WrlLLen for Lhe onllne markeLer, Lhls research brlef wlll provlde a prlmer on moblle search Lrends and

1
Cowen and Company: lnLerneL and new Medla - lndusLry CuLlook - !une 2011
2
hLLp://glgaom.com/moblle/7-slgns-LhaL-androld-ls-falLerlng-as-los-sLrengLhens/
3
hLLp://www.asymco.com/2012/01/17/Lhe-rlse-and-fall-of-personal-compuLlng/
4
hLLp://www.Lheverge.com/2013/1/23/3908602/apple-has-sold-over-300-mllllon-los-devlces-acLlvaLed
$%&'%%( *++,-% .(/
011-%2 3+(456%#4 7.8%
.-#%./9 :+5,7& 6+#%
&7.( . :"--"+( &.:-%&4 .(/
46.#&17+(%4
Source: Cowen and Company: Internet and New Med|a - Industry Cut|ook - Iune 2011
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 3 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
pro[ecLlons, explore cosL and performance across devlces, and help adverLlsers opLlmlze Lhelr moblle
search opporLunlLy.

Study Methodo|ogy:

1o underLake Lhls sLudy, we looked across leadlng brands and adverLlsers LhaL manage more Lhan $4
8llllon ln annuallzed pald-search spend. Cur daLa seL represenLed all ma[or lndusLry secLors alongslde
Lhe followlng counLrles and/or reglons:

AusLralla
8razll
Canada
Chlna
Lurozone
lndla
!apan
Mexlco
new Zealand
8ussla
Slngapore
unlLed klngdom
uSA

1he slze and dlverslLy of our daLa seL, coupled wlLh Lhe broad geographlc coverage, enables us Lo
provlde Lhe mosL comprehenslve reporL on how smarL moblle devlces are changlng pald-search. WlLh
LhaL sald, Marln's cllenLs are malnly large adverLlsers spendlng upwards of $100,000 per monLh on pald-
search. As such, Lhe lnformaLlon presenLed ln Lhls reporL ls blased Lowards larger adverLlsers, and may
noL reflecL moblle search Lrends for small or medlum slzed buslnesses.

Moblle Search AdverLlslng around Lhe Clobe - 2012

age 6 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
1".! - 2 0*-!%/ +!"!%+
3'#-$% +%".)4 !.%*/+

1he lasL Lwelve monLhs have seen conLlnued adopLlon of moblle search by consumers and adverLlsers.
8eLween !anuary and uecember 2012, our uS cllenLs saw Lhelr share of Coogle pald-search cllcks from
moblle devlces rlse from 14.2 Lo 23.4.

Spend, or ad budgeLs on moblle devlces, lncreased aL an even
fasLer pace. 8eLween !anuary and uecember 2012, adverLlsers
lncreased Lhelr share of search budgeL on moblle devlces from
10 Lo 18.4, an 83 lncrease. uesplLe efforLs Lo conLlnually
lncrease spend on moblle, ad budgeLs sLlll lagged Lhe cllck
volume arlslng from smarL moblle devlces.

As consumers rapldly expand Lhe use of smarLphones and LableLs, adverLlsers wlll need Lo follow by
lncreaslng Lhelr search budgeL on Lhese devlces. So how much budgeL should adverLlsers dedlcaLe Lo
speclflc moblle devlces? AllocaLlng budgeL across LableLs and smarLphones can depend on an
adverLlser's lndusLry, audlence and geography. ln Lhe absence of more preclse daLa, Lhe fracLlon of pald
search cllcks from LableLs relaLlve Lo LoLal moblle cllcks ls a good proxy for deLermlnlng how much
budgeL Lo allocaLe for LableLs.
;+:"-% /%8"3%4 '"-- .33+5(& <+# .
&7"#/ +< 1."/=4%.#37 3-"3>4 .(/
4%.#37 :5/,%&4 :9 ?%3 @ABC
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 7 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
#%*)43".5-*, +%".)4 1%.('.3"*)% ").'++ /%6-)%+

Moblle search behaves dlfferenLly from search on Lhe LradlLlonal compuLer. lor one, smarLphones and
LableLs have less avallable ad lnvenLory because of consLralned screen real-esLaLe. 1hls has Lhe
unlnLended effecL of reduclng Lhe number of ad lmpresslons LhaL can be served across Lhese newer
devlces. Powever, our daLa suggesLs LhaL users are acLually more engaged wlLh search resulLs on Lhelr
moblle devlces. 1he followlng charL looks aL Lhe average cllck-Lhrough raLe (C18) across devlces and
shows LhaL smarLphone and LableL users have C18s LhaL are respecLlvely 107 and 37 hlgher Lhan
users on deskLop compuLers.



AnoLher poslLlve for adverLlsers ls LhaL Lhe cosL-per-cllck (CC) on smarLphones and LableLs ls much
lower Lhan lL ls on deskLop compuLers. As Lhe followlng charL lllusLraLes, smarLphones cllcks are
approxlmaLely 36 cheaper Lhan deskLop cllcks. Along Lhe same llnes, LableL cllcks are abouL 17
cheaper Lhan Lhelr deskLop counLerparLs.






!%.#37 ./4 +( 46.#&17+(%4 D
&.:-%&4 7.8% . 6537 7",7%#
3-"3> &7#+5,7 #.&% &7.( 4"6"-.#
./4 +( &7% /%4>&+1
E.:-%& .(/ !6.#&17+(% FGF4
.#% 37%.1%# &7.( 3+615&%#4
:5& "(3#%.4"(, .& . <.4&%# #.&%
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 8 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
WlLh LhaL sald, Lhe sLory doesn'L [usL end wlLh LableLs and smarLphones belng more cosL-effecLlve.
ulgglng a llLLle deeper, we found LhaL CCs on LableLs and smarLphones lncreased 23 and 13
respecLlvely durlng 2012. ln comparlson, CCs on compuLers lncreased approxlmaLely 9. So whlle
smarLphones and LableLs have lower CCs, Lhey're lncreaslng aL a fasLer pace. Cur analysls suggesLs LhaL
LableLs CCs could equal deskLop CCs durlng 2013.





Whlle CCs are one slde of Lhe coln, converslon raLes are Lhe oLher. Looklng across our cllenLs, we
dlscovered LhaL Lhere ls a slgnlflcanL dlfference beLween converslon raLes across devlces. Cn average,
we found LhaL ad cllcks orlglnaLlng from smarLphones have Lhe lowesL converslon raLes, whereas ad
cllcks orlglnaLlng from LableLs are becomlng more comparable Lo compuLers. Pere lL's lmporLanL Lo noLe
LhaL converslon raLes for LableL devlces are lmprovlng relaLlve Lo deskLops and smarLphones. uurlng
2012, LableLs saw a 31 lncrease ln converslon raLes, compared Lo 9 for smarLphones and 7 for
compuLers.







H"&7 3+(8%#4"+( #.&%4
41">"(, :9 CBI /5#"(, @AB@2
&.:-%&4 .#% <.4& :%3+6"(,
.( %F+66%#3% 1+'%#7+54%
E.:-%& FGF4 3+5-/ %J5.-
?%4>&+1 FGF4 /5#"(, @ABC
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 9 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
1hough smarLphones underperformed on converslon raLes, more research ls warranLed here. 8elng
lnherenLly moblle, smarLphones are used dlfferenLly and many converslons may be happenlng ln a
physlcal sLore, maklng Lhem hard Lo Lrack. AddlLlonally, Lhe rlse of shopplng apps makes converslon
Lracklng more complex and somewhaL opaque. As such, Lhe value of smarLphone adverLlslng ls more
dependenL on Lhe Lype of adverLlser and producL, and markeLers should keep an open mlnd abouL Lhelr
efflcacy.





1".! -- 2 0*-!%/ 5-*,/'3 "*/ !4% %0.'7'*%
3'#-$% +%".)4 !.%*/+

uesplLe Lhe macro-economlc headwlnds, 2012 was a good year for moblle search ln Lurope. ln !anuary
2012, smarL moblle devlces accounLed for 14.8 of Coogle's uk pald cllcks. 8y uecember, smarLphones
and LableLs were generaLlng roughly 24.4 of all pald cllcks. ln a mere Lwelve monLhs, Lhe uk saw
Coogle's share of pald cllcks from moblle lncrease 63. 8y comparlson, Lhe Lurozone saw lLs share of
pald cllcks from smarL moblle devlces rlse from 3.9 Lo 14.3, a sLaggerlng lncrease of 146.







Spend, or ad budgeLs on moblle devlces, lncreased aL an even fasLer pace. uurlng 2012, uk adverLlsers
lncreased Lhelr share of search budgeL on smarL moblle devlces from 9.94 Lo 19.32, an lncrease of
94. Meanwhlle, Lurozone adverLlsers upped Lhelr lnvesLmenL ln moblle search from 4.8 Lo 11.8 of
pald-search budgeLs, reflecLlng an lncrease of 146. uesplLe efforLs Lo conLlnually lncrease ad spend,
pald-search budgeL share sLlll lagged Lhe cllck share of smarL moblle devlces.








$9 ?%3%6:%# @AB@2 46.#&
6+:"-% /%8"3%4 '%#%
/%-"8%#"(, . J5.#&%# +< &7%
1."/ 4%.#37 3-"3>4 "( &7% KL
?5#"(, @AB@2 &7% M5#+N+(%
"(3#%.4%/ "&4 6+:"-% .(/
&.:-%& 1."/=4%.#37 :5/,%&
47.#% :9 BOPI
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 11 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
#%*)43".5-*, +%".)4 1%.('.3"*)% ").'++ /%6-)%+

So how dld smarL moblle devlces perform ln Lurope? noL surprlslngly, lL wasn'L Loo dlsslmllar from Lhe
uS. 1he followlng charL looks aL Lhe average cllck Lhrough raLe (C18) across devlces, and shows LhaL ads
placed on smarLphones and LableLs have a hlgher C18 Lhan Lhose placed on compuLers. 1haL belng sald,
we also saw LhaL dlfferences ln C18 across devlces are more pronounced ln Lhe uk, and less sLrlklng ln
Lhe Lurozone.



Cn Lhe cosL slde, uk smarLphones cllcks are cheaper Lhan Lhe correspondlng deskLop and LableL cllcks. ln
Lhe Lurozone, LableL and compuLer cllcks cosL abouL Lhe same, whlle smarLphone cllcks are Lhe cheapesL
overall. Cur flndlngs also show LhaL LableL CCs ln Lhe uk and Lhe Lurozone are closlng ln on compuLer
CCs.



Across Lhe uk and Lhe Lurozone, we found slgnlflcanL CC lnflaLlon Lhrough 2012. Speclflcally, we saw
LhaL LableLs had Lhe greaLesL lncrease ln CC, averaglng a 36 lncrease ln Lhe uk and a 24 bump ln Lhe
Lurozone. Pere, lL's lmporLanL Lo noLe LhaL our analysls doesn'L Lake currency flucLuaLlons lnLo accounL.
FEQ4 <+# ./4 +(
46.#&17+(%4 3+(&"(5%
&+ %3-"14% &7% /%4>&+1
.(/ &7% &.:-%&
E7% KL .(/ &7% M5#+N+(%
.#% 4%%"(, 4",(4 +<
3+(8%#,%(3% "( /%4>&+1
.(/ &.:-%& FGF4
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 12 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.



Whlle CCs are lmporLanL, adverLlsers ulLlmaLely care abouL sales and revenue. Pere we found LhaL
deskLop and lapLop sLlll showed Lhe besL converslon raLes across Lhe uk and Lhe Lurozone. 8uL wlLhln
Lhe Lurozone ln parLlcular, LableL converslon raLes are sLarLlng Lo caLch up wlLh Lhose on compuLers.
SmarLphones had Lhe lowesL converslon raLes, buL here lL's lmporLanL Lo keep ln mlnd LhaL smarLphone
converslons may be happenlng vla a phone call or ln a physlcal sLore, leadlng Lo arLlflclally depressed
measuremenLs.




E.:-%&4 4.' &7% 7",7%4&
"(<-.&"+( "( FGF! .3#+44 &7%
KL .(/ &7% M5#+N+(%
F+(8%#4"+( #.&%4 <#+6
&.:-%& 3-"3>4 .#%
4&.#&"(, &+ #"8.- &7+4%
.#"4"(, <#+6 /%4>&+1
3-"3>4 "( &7% M5#+N+(%
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 13 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
1".! --- 2 3'#-$% +%".)4 ").'++ !4% ,$'#%

Moblle ls lnherenLly more global Lhan Lhe deskLop. WlLh handseLs becomlng smarLer and bandwldLh
more bounLlful, moblle devlces are now polsed Lo leapfrog compuLers ln brlnglng Lhe lnLerneL and
dlglLal commerce Lo bllllons of people around Lhe world.
1o sLudy lnLernaLlonal moblle search Lrends, we analyzed Coogle's daLa from across a number of ma[or
markeLs. 1he followlng charLs break ouL Lhe share of cllcks and ad-spend across devlces for each counLry
under conslderaLlon. AddlLlonally, Lhey compare devlce-speclflc cosL per cllck (normallzed Lo uSu) and
cllck Lhrough raLes for each reglon. lL's lmporLanL Lo re-emphaslze here LhaL Lhe charLs below only refer
Lo Coogle's performance, and do noL lnclude daLa from oLher search englnes such as 8lng, 8aldu, ?andex
or ?ahoo! !apan.
R+&%S T.-5%4 :%-+' .#% #+5(/%/ &+ +(% /%3"6.- 1-.3%2 '7"37 3.( -%./ &+ 3-"3> .(/ 41%(/ 47.#% .3#+44
/%8"3%4 %U3%%/"(, BAAI :9 . 46.-- 6.#,"(V

Moblle Search AdverLlslng around Lhe Clobe - 2012

age 14 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.




Moblle Search AdverLlslng around Lhe Clobe - 2012

age 13 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
'1!-3-7-*, 3'#-$% +%".)4 )"31"-,*+

8y uecember 2013, smarL moblle devlces could make up a Lhlrd of Coogle's
pald-search cllcks for large adverLlsers. Clven moblle's raLe of growLh and
geographlc reach, lL ls fasL becomlng an area of focus for markeLers.
Powever, moblle campalgns can be qulLe dlfferenL from Lhelr deskLop
counLerparLs ln Lerms of LargeLlng, blddlng and ad formaLs. AddlLlonally,
moblle users could be uslng search ln dlfferenL ways Lhan a deskLop user.
lor example, cllck-Lo-call ads are hlghly relevanL for a smarLphone buL noL
for a deskLop or a LableL. 1o LhaL end, Lhe followlng secLlon covers besL pracLlces for moblle pald-search
and wlll help search markeLers make Lhe LranslLlon from deskLop Lo moblle devlces.
1. 8reak out Mob||e Campa|gns:
Whlle lL may lncrease operaLlonal overhead, Lhere are good reasons Lo break ouL moblle campalgns.
llrsL, moblle CCs and C18s are slgnlflcanLly dlfferenL from Lhe deskLop. Second, desLlnaLlon u8Ls could
be dlfferenL for moblle ads. 1hlrd, locaLlon ls usually a more lmporLanL conslderaLlon ln moblle searches.
llnally, wlLh Lhe emergence of newer ad formaLs such as cllck-Lo-call ads, moblle ad campalgns are
geLLlng lncreaslngly speclallzed. 8y separaLlng moblle ouL, markeLers wlll have more conLrol over search
campalgns, be able Lo wrlLe moblle-speclflc ad copy, and drlve hlgher performance overall.






2. 1arget Users w|th Dev|ce-Spec|f|c Copy:
When browslng on moblle devlces, consumers ofLen look for assurance LhaL llnks wlll Lake Lhem Lo a
moblle opLlmlzed experlence. lncorporaLlng devlce-speclflc ad copy lnLo pald search ads glves
consumers LhaL assurance, lncreaslng cllck-Lhrough raLes ln Lhe process. ln Lhe example below, 8anana
8epubllc has Lallored moblle ad Lo LargeL lhone users. Whlle Lhls ls a relaLlvely nuanced change, moblle
copy LhaL lncorporaLes Lhe devlce name Lyplcally ouLperforms even Lhe besL deskLop copy.


033+#/"(, &+
*++,-%2 +(% "(
&7#%% 6+:"-%
4%.#37 J5%#"%4
7.8% -+3.- "(&%(&
Use mob||e spec|f|c ad types and ad copy
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 16 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.










3. r|or|t|ze Ad os|t|on:
1he effecLlve real esLaLe for pald search ads ls smaller on moblle devlces, lncreaslng Lhe lmporLance of
ad poslLlon. 8ecause smarLphone browsers supporL fewer ad unlLs Lhan deskLop browsers, lL's exLremely
lmporLanL Lo ensure LhaL your ads are ln poslLlon 1 or 2. As a rule of Lhumb, we recommend seLLlng blds
for moblle campalgns aL Lhe ouLseL Lo Lwlce LhaL of deskLop campalgns and Lhen ad[usLlng downward.
Whlle Coogle doesn'L allow for poslLlon-based blddlng, adverLlsers should LesL dlfferenL blds whlle
monlLorlng ad poslLlon Lo ensure an average poslLlon LhaL's ln beLween 1 and 2. 1ools such as Marln can
monlLor Lhls and auLomaLlcally alerL Lhe adverLlser when correcLlve acLlon needs Lo be Laken.

4. Cpt|m|ze the User Lxper|ence:
A key reason for lower converslon raLes on smarLphones ls Lhe lack of moblle-opLlmlzed webslLes. MosL
webslLes are sLlll dlfflculL Lo browse on smarLphone browsers, and usablllLy lssues can lnLroduce frlcLlon
lnLo Lhe purchase process. 1o make moblle a sLrong and susLalnable parL of Lhe converslon funnel,
adverLlsers should opLlmlze Lhelr webslLe for smarLphone browsers, and sub[ecL lL Lo Lhe same
sLandards of usablllLy and A|8 LesLlng as Lhey do Lhelr regular webslLe.
S. 1rack Mob||e Convers|ons:
Moblle searches ofLen resulL ln converslons LhaL happen vla a call or a physlcal sLore. unforLunaLely,
mosL markeLers lack Lhe ablllLy Lo glue Lhese cllcks LogeLher lnLo a unlfled converslon funnel. MarkeLers
should look Lo esLlmaLe Lhelr moblle-lnfluenced revenue Lhrough Lhe use of popular moblle ad formaLs
such as cllck-Lo-call and sLore-locaLor. 8y comblnlng Lhe Lyplcal converslon raLe for ln-sLore and phone-
based LransacLlons wlLh Lhe average revenue per LransacLlon, markeLers can esLlmaLe a revenue per
cllck for moblle devlces and ad[usL Lhelr moblle CCs and budgeL accordlngly.
Lffect|ve dev|ce target|ng across desktop and smartphones
Moblle Search AdverLlslng around Lhe Clobe - 2012

age 17 of 17

CopyrlghL 2013 Marln SofLware, lnc. All rlghLs reserved.
"#$%& '"()* +$,&-"(.

Marin provides a leading revenue acquisition management platform used by advertisers and
agencies to manage more than $4 billion in annualized ad investments. Offering an integrated
platform for search, social, display, and mobile marketing, Marin helps advertisers and agencies
improve financial performance, save time, and make better decisions. Headquartered in San
Francisco, with offices worldwide, Marin's technology powers marketing campaigns in more than
160 countries. For more information about Marin's products, please visit:
http://www.marinsoftware.com.


Un|ted States (G|oba| neadquarters)

123 Mlsslon SLreeL
23Lh lloor
San lranclsco, CA 94103
+1-413-399-2380

Lng|and (LMLA & AAC neadquarters)
1sL lloor, Crlon Pouse
3 upper SL. MarLln's Lane
London WC2P 9LA
+44-843-262-0404

New ork
213 ark Avenue SouLh
SulLe 1801
new ?ork, n? 10003
1el: (646) 490-2427

Ch|cago
203 n. LaSalle SL, SLe 2100
Chlcago, lL 60601
1el: (312) 267-2083

Austra||a
Level 31
88S 1ower [ Aurora lace
88 hllllp SLreeL
Sydney nSW 2000
1el: +61 (0)2 8211 0333



Irance
AcLualls Level 2
21 and 23 8oulevard Paussmann
arls 73009
1el: +33 1 36 03 66 44

Germany
Am kalserkal 1
20437 Pamburg
1el: +49 40 80 80 74-322

Iapan
3rd lloor, Sanno ark 1ower
2-11-1 nagaLa-cho
Chlyoda-ku
1okyo 100-6162
1el: +81 3 6203 3000

S|ngapore
Cne 8affles Cuay
norLh 1ower, Level 23
048383
1el: +63 6622 3888