Best Prac­tices In Mar­ket Mix Mod­el­ling

Point of Purchase - - CONTENTS - By Jim Nail with Chris Char­ron and Jen­nifer Joseph

Pre­sent­ing to you the sec­ond part of the “Mar­ket­ing Mea­sure­ment” se­ries.


In­creased de­mand by CEOs for greater ac­count­abil­ity of mar­ket­ing spend­ing comes just at a time when pro­lif­er­at­ing me­dia choices un­der­cut the sim­plic­ity of mea­sur­ing brand­ing or sales im­pact. Com­pa­nies can learn valu­able lessons from CPG mar­keters who have be­gun to un­tan­gle mul­ti­chan­nel mar­ket­ing’s ef­fects us­ing so­phis­ti­cated sta­tis­ti­cal mod­els. The keys to suc­cess lie in scop­ing the project cor­rectly, gain­ing the sup­port of stake­hold­ers, and work­ing with a mod­el­ling spe­cial­ist who can ap­ply the find­ings to busi­ness prob­lems


Re­sults of a sur­vey con­ducted by For­rester and the As­so­ci­a­tion of Na­tional Ad­ver­tis­ers on mar­ket­ing ac­count­abil­ity re­vealed sub­stan­tial in­ter­est in mar­ket mix mod­el­ing as a tool to bring greater ac­count­abil­ity to mar­ket­ing spend­ing.1


Left Brain Mar­ket­ing — an­a­lyt­i­cal mar­ket­ing strate­gies, skills, and pro­cesses — is not just for data­base mar­keters.2

Dur­ing the past 15 years, mar­keters have be­gun to un­tan­gle the sales im­pact of brand ad­ver­tis­ing and other el­e­ments of an in­te­grated mar­ket­ing mix, with the help of three tech­no­log­i­cal developments:

1) greater avail­abil­ity of store-level sales data;

2) in­creas­ing data pro­cess­ing power; and

3) the ap­pli­ca­tion of sta­tis­ti­cal tech­niques.3 CPG com­pa­nies, pulling from the vast data stores of In­for­ma­tion Re­sources, Inc. (IRI) and ACNielsen, have pi­o­neered the tech­nique of mod­el­ling mar­ket­ing tac­tics with busi­ness re­sults. In re­cent years, this suc­cess has be­gun to in­ter­est other in­dus­tries.

For those not fa­mil­iar with the con­cept, a mar­ket mix model:

Links mar­ket­ing spend­ing to sales re­sults

A mar­ket­ing model lines up a com­pany’s ad­ver­tis­ing, pro­mo­tion, and sales ac­tiv­i­ties with prod­uct sales over time (see Fig­ure 1). Us­ing sta­tis­ti­cal tech­niques such as lin­ear and mul­ti­vari­ate re­gres­sion anal­y­sis, statis­ti­cians look for cor­re­la­tions be­tween mar­ket­ing ac­tiv­ity and sales vol­ume and then tease out the sales im­pact of in­di­vid­ual me­dia or pro­mo­tional el­e­ments of a cam­paign. The re­sult is an equa­tion that math­e­mat­i­cally de­scribes con­sumers’ re­sponses to mar­ket­ing stim­uli more com­pre­hen­sively than a sim­ple di­rect mail re­sponse re­port and with greater so­phis­ti­ca­tion than a weekly sales re­port.

Quan­ti­fies the im­pact of mul­ti­ple de­mand driv­ers

Co­ef­fi­cients in the equa­tion de­scribe the rel­a­tive im­pact of each mar­ket­ing ac­tiv­ity. But mar­ket­ing isn’t the only thing that drives de­mand. For ex­am­ple, the price of gasoline has damp­ened con­sumer en­thu­si­asm for SUVs, while a scorch­ing heat wave may spur con­sump­tion of soft drinks. Modellers com­pile data on these ex­ter­nal fac­tors, an­a­lyze their im­pact on sales, and in­clude rel­e­vant fac­tors as ad­di­tional co­ef­fi­cients in the model equa­tion.

Re­veals mar­ket­ing’s in­cre­men­tal sales im­pact

Al­most two-thirds of mar­keters agree on “in­cre­men­tal sales gen­er­ated by mar­ket­ing” as the def­i­ni­tion of ROI.4 Mar­ket mix mod­els typ­i­cally show that

50% to 60% of sales would likely have oc­curred in the ab­sence of mar­ket­ing. The model then as­signs the re­main­ing sales to the mar­ket­ing ac­tiv­ity that stim­u­lated them and cal­cu­lates the sales (or profit) per dol­lar of spend­ing on each mar­ket­ing ac­tiv­ity.

An­a­lyzes ge­o­graphic, me­dia, or brand-level bud­get al­lo­ca­tions

De­pend­ing on the data in­volved and the type of anal­y­sis ap­plied, mar­ket mix mod­el­ling comes in three f lavours (see Fig­ure 2). A brand man­ager can build a mar­ket­ing mix model com­par­ing the im­pact of ad­ver­tis­ing, con­sumer coupons, and trade pro­mo­tion, while an ad­ver­tis­ing man­ager can cre­ate a me­dia mix model of the ef­fects of TV, ra­dio, In­ter­net, and print ads. For the CMO, who must de­cide how much of the to­tal mar­ket­ing bud­get to al­lo­cate to each of the com­pany’s brands, a port­fo­lio model can show how to max­i­mize the ROI and can also re­veal the halo ef­fect, where ad­ver­tis­ing for one brand may also drive sales for a re­lated brand.

Mod­els Re­quire Clar­ity And Com­mit­ment

Given the in­creas­ing clam­our for mar­ket­ing ac­count­abil­ity and the suc­cess of mar­ket mix mod­els in the CPG in­dus­try, it is tempt­ing to view this ap­proach as the an­swer to all mar­ket­ing in­vest­ment ques­tions. Al­though mod­els are pow­er­ful tools, mar­keters eval­u­at­ing this tech­nique should un­der­stand four lim­i­ta­tions. Mar­ket mix mod­els:

Re­flect his­tory

Statis­ti­cians need at least two, and prefer­ably three, years of weekly mar­ket­ing spend­ing and sales data to build a model. This re­liance on the past means that the model can’t project the im­pact of a medium that hasn’t yet been in the mix and can’t tell if spend­ing has been too low to reg­is­ter an im­pact. This is com­mon to­day in on­line ad­ver­tis­ing. Also, the model’s ac­cu­racy quickly de­clines when try­ing to com­pare the ef­fect of a bud­get shift of more than 15% with his­tor­i­cal trends.5

Start with a busi­ness goal

Mar­ket mix mod­el­ling is an um­brella term en­com­pass­ing a range of tech­niques that can an­swer a va­ri­ety of busi­ness ques­tions. Mar­keters must clearly ar­tic­u­late the mar­ket­ing de­ci­sions they are try­ing to sup­port and the hy­pothe­ses they want to test to en­able the modeller to bring the right data and an­a­lytic tech­niques to bear. But re­lat­ing the model to im­por­tant busi­ness is­sues helps to win ap­proval for the project.

Re­quire a sig­nif­i­cant com­mit­ment

Mar­keters can ex­pect to spend at least $50,000 to $100,000 to hire a mar­ket­ing mod­el­ling ex­pert, as well as ad­di­tional costs if they pur­chase ex­ter­nal data. As mod­els be­come a rou­tine part of mar­ket­ing, com­pa­nies will also need to put ef­fort into in­te­gra­tion projects to au­to­mat­i­cally pull data that is buried in the data ware­house or in­ter­nal fi­nance and CRM sys­tems. In­cor­po­rat­ing us­age of the mod­els throughout the mar­ket­ing process re­quires fur­ther change man­age­ment.

Aren’t right for all in­dus­tries

Mar­ket mix mod­els are most ap­pro­pri­ate for brands that have large bud­gets, a highly di­verse mar­ket­ing mix, and dis­tri­bu­tion via in­di­rect chan­nels. B2B com­pa­nies whose mar­ket­ing ob­jec­tive is to drive leads and spend lit­tle on cor­po­rate brand­ing are bet­ter served by im­prov­ing their lead man­age­ment sys­tems.6 Sev­eral ven­dors re­port that mod­els usu­ally re­sult in a 10% gain in mar­ket­ing ef­fi­ciency; thus, small­bud­get brands, where these sav­ings won’t off­set the model costs, shouldn’t use the tech­nique. Sim­i­larly, cat­a­logue re­tail­ers with lit­tle spend­ing out­side of di­rect mail won’t de­rive as much ben­e­fit. Prod­ucts like movies, where each re­lease is a new prod­uct launch, won’t have the his­tor­i­cal data to build the model.


Given the ef­fort and ex­pense in­volved, com­pa­nies should care­fully plan and ex­e­cute a new mod­el­ling ini­tia­tive. As one re­tailer noted, “You have to ap­proach a model like a real IT project.” With a decade and a half of his­tory be­hind this tech­nique, mar­keters can fol­low in­dus­try best prac­tices but should be ready to tai­lor the ap­proach to their spe­cific in­dus­try, busi­ness prob­lems, and data avail­abil­ity. These projects evolve in three phases (see Fig­ure 3):

1. Plan: Scop­ing the project and se­lect­ing a mod­el­ling part­ner.

2. Pre­pare: Re­fin­ing the scope and gath­er­ing re­quired data.

3. Build: Cre­at­ing and re­fin­ing the model equa­tion, im­ple­ment­ing model find­ings, com­par­ing them with ac­tual re­sults, and en­hanc­ing the model’s pre­dic­tive abil­ity.

Phase 1: Plan The Project And Pick The Right Ven­dor

While com­pa­nies with econo­met­ric mod­el­ling staff may be tempted to go it alone, nu­ances of mar­ket­ing mod­el­ling re­quire the ex­per­tise of a spe­cial­ist. In ad­di­tion to mar­ket­ing knowl­edge,

mod­el­ling part­ners should have the f lex­i­bil­ity to adapt their ap­proach, the abil­ity to clearly ex­plain the model’s in­ner work­ings, and the will­ing­ness to help the com­pany ac­cess and use the model on a rou­tine ba­sis. Com­pa­nies should eval­u­ate a modeller’s:

Skills be­yond sta­tis­ti­cal ex­per­tise

Banks, auto com­pa­nies, and some re­tail­ers al­ready do sta­tis­ti­cal mod­el­ling of the econ­omy to shape busi­ness plan­ning. While mar­ket­ing mod­el­ling uses many of the same sta­tis­ti­cal tech­niques, the sim­i­lar­ity ends there. The great­est dif­fer­ence is in ad­just­ing mar­ket­ing spend­ing to ac­count for the lin­ger­ing ef­fect that an ad cam­paign has, even af­ter it is over. Trans­form­ing mar­ket­ing spend­ing data into “ad stock” re­quires knowl­edge that an econo­me­tri­cian doesn’t have of the de­cay rate of the im­pact of an ad cam­paign.7 Modellers also need to un­der­stand com­mon me­dia con­cepts like gross rat­ing points (GRP) and cir­cu­la­tion and must be ac­cus­tomed to in­cor­po­rat­ing these dis­parate data types into a model.

Cat­e­gory un­der­stand­ing

Mod­els for fi­nan­cial ser­vices, re­tail, and phar­ma­ceu­ti­cal com­pa­nies will en­counter dif­fer­ent is­sues with dis­tri­bu­tion chan­nels, mar­ket­ing mix, and data avail­abil­ity. But just be­cause modellers don’t have rel­e­vant in­dus­try ex­pe­ri­ence isn’t a rea­son to dis­miss them. To cre­ate its short­list of po­ten­tial mar­ket­ing part­ners, Wa­chovia is­sued a re­quest for in­for­ma­tion (RFI) de­scrib­ing the busi­ness prob­lem and ask­ing ven­dors to dis­cuss the ap­proach they would take. The re­sponses al­lowed Wa­chovia to de­ter­mine which ven­dors had the great­est f lex­i­bil­ity and cre­ativ­ity to adapt their ex­pe­ri­ence to the bank’s needs.

Busi­ness model

Some modellers re­gard the model as their in­tel­lec­tual prop­erty and only give the client re­ports of re­sults. But to get the most out of the mod­el­ling in­vest­ment, mar­keters should make it clear that over time, they will need in­creas­ing ac­cess to the model to up­date it and use it on a rou­tine ba­sis for plan­ning, strat­egy, and on­go­ing eval­u­a­tion of mar­ket­ing ef­fec­tive­ness.

In­te­gra­tion skills

For­rester has heard re­peat­edly that one of the hard­est parts of the model build­ing process is solv­ing the tech­ni­cal and or­ga­ni­za­tional hur­dles to ac­cess­ing nec­es­sary data from in­ter­nal sys­tems and data­bases. As part of the project frame­work, com­pa­nies should plan to build au­to­mated data pulls to ease sub­se­quent model up­dates. Modellers should be able to pro­vide ex­am­ples of sim­i­lar projects for other clients and demon­strate the tech­ni­cal skills to de­fine and di­rect the work.

Phase 2: Pre­pare The Data And Or­ga­ni­za­tion

With a modeller on­board, the next task is to care­fully plan the project, iden­tify data sources, and en­gage other de­part­ments or stake­hold­ers that will be af­fected by the project and the model’s rec­om­men­da­tions. This prepa­ra­tion is crit­i­cal to suc­cess, and mar­keters should guide the project through a se­ries of meet­ings to build a con­sen­sus around the de­sired out­come from the model and gain sup­port for their con­tin­ued par­tic­i­pa­tion.

Re­fine the scope

Us­ing a model to de­ter­mine mar­ket­ing ROI sounds de­cep­tively sim­ple, but be­fore mod­el­ling can be­gin, the com­pany needs to move from the high-level project ob­jec­tives to a more de­tailed def­i­ni­tion of the model: which brands will be in­volved, what the met­ric of suc­cess will be, what type of mod­el­ling tech­niques are best, etc. These de­ci­sions can dra­mat­i­cally af­fect the time, cost, and re­sources re­quired for the model. One re­tailer we spoke with spent the first 60 days work­ing with the ven­dor and in­ter­nal groups to clar­ify such is­sues as whether to use in­creased sales or prof­its as the met­ric and whether to look at short-, medium-, or long-term ef­fects.

Gather the data

The type of model and the met­ric will be­gin to shape the data re­quire­ments. Plan­ning should be­gin to iden­tify which in­ter­nal and ex­ter­nal data sources will need to be tapped. It is not un­usual to be­gin with 20 or more data sources, each of which may re­quire mul­ti­ple rounds to pre­pare the data cor­rectly. Com­pa­nies should staff ad­e­quately for this task. One fi­nan­cial ser­vices com­pany we spoke with had as­signed two em­ploy­ees full-time for two months to seek out, ex­tract, and pre­pare nec­es­sary data from in­ter­nal sys­tems.

Phase 3: Build, Val­i­date, And En­hance The Model

Build­ing a model is a lot of sci­ence, but it is also partly an art, re­quir­ing sev­eral rounds of eval­u­a­tion and re­fine­ment be­fore ac­cept­ing a fin­ished prod­uct. Even so, scep­tics are likely to re­ject the model un­til a cam­paign based on its rec­om­men­da­tions runs in the mar­ket and its pro­jec­tions come true.

Once a model proves ac­cu­rate, mar­keters should work to ex­pand its power and abil­ity to guide busi­ness de­ci­sions. The fol­low­ing steps should be taken:

It­er­ate the model

Ini­tial drafts of the model should be scru­ti­nized for two fac­tors. First, is the “R-squared” (the close­ness of the cor­re­la­tion be­tween a vari­able and sales) high enough?

If not, the model may need added vari­ables, more de­tailed data, or, in ex­treme cases, it may be aban­doned. Sec­ond, de­cide if the cor­re­la­tions iden­ti­fied by the model are co­in­ci­den­tal or causal. Stake­hold­ers from ar­eas such as fi­nance, sales, and op­er­a­tions should continue their in­volve­ment through these stages to help mar­ket­ing iden­tify spu­ri­ous cor­re­la­tions. Be­yond im­prov­ing the bet­ter model, this will min­i­mize the po­ten­tial for groups that don’t like the model, on the grounds that a “black box” for­mula doesn’t ac­cu­rately ref lect the dy­nam­ics of the busi­ness.

Val­i­date the fore­casts

Mod­els typ­i­cally iden­tify sig­nif­i­cant changes in the me­dia used or the bud­get al­lo­cated to dif­fer­ent cam­paign el­e­ments. None of the com­pa­nies we spoke with au­to­mat­i­cally shifted their plans to ex­actly mir­ror the model’s rec­om­men­da­tions. In­stead, they said they would make a less dras­tic, but still mea­sur­able, change in their plans, run a cam­paign, and match up mar­ket­place re­sults to the model’s fore­casts. Only when real re­sults prove the model’s pre­dic­tive ac­cu­racy do com­pa­nies be­gin to use the model to drive strate­gic mar­ket­ing de­ci­sions.

Con­struct tests to ex­pand the model’s power

Be­cause mod­els emerge from his­tor­i­cal data, they high­light what worked best among past cam­paigns. But new me­dia forms, such as prod­uct place­ments, blogs, or, in many cases, on­line ad­ver­tis­ing, may not have ac­cu­mu­lated enough data for the model to ref lect them. To ex­pand the model’s sen­si­tiv­ity, com­pa­nies should con­struct a num­ber of dif­fer­ent mar­ket­ing plans in sets of test mar­kets, track the re­sults, and then use the model to add new vari­ables and ad­just the co­ef­fi­cient of pre­vi­ously known vari­ables.


In the early days, the mod­el­ling in­dus­try was made up of a hand­ful of small, in­de­pen­dent bou­tiques. But ma­jor agency hold­ing com­pa­nies, tech­nol­ogy con­sul­tants, and soft­ware com­pa­nies see busi­ness op­por­tu­ni­ties in mar­keters’ mea­sure­ment pain (see Fig­ure 4, see Fig­ure 5, and see Fig­ure 6). Not only does this broader mar­ket­place give mar­keters more choices, but it also gives them the abil­ity to as­sem­ble their own best-of-breed mea­sure­ment sys­tems by pick­ing the top modeller for their needs, em­bed­ding that model in a tech­nol­ogy plat­form, and then en­gag­ing the agency or a con­sul­tant to im­ple­ment the changes needed to en­hance their busi­ness.

Spe­cial­ist bou­tiques thrive

Since the early ’90s, the power of PCs and soft­ware has en­abled any­one with a strong back­ground in econo­met­rics and sta­tis­tics to hang out a shin­gle as a mar­ket­ing modeller. Early en­trants like Mar­ket­ing An­a­lyt­ics continue to at­tract blue-chip clients like Kraft Foods and Nestlé, while newer en­trants like iKnow­tion continue to spring up. A full-ser­vice ap­proach and deep ex­pe­ri­ence make them good choices for com­pa­nies be­gin­ning the mod­el­ling learn­ing curve. How­ever, com­pa­nies need to make sure that these firms of­fer com­plete ac­cess to the model, in the event that they de­cide to bring it in-house or use it in con­junc­tion with an­other ap­pli­ca­tion.

Mar­ket­ing ser­vices com­pa­nies ex­pand of­fer­ings

Ad agen­cies like DDB World­wide Com­mu­ni­ca­tions and data­base mar­keters like Tar­get base have tra­di­tion­ally served clients by cre­at­ing and im­ple­ment­ing mar­ket­ing cam­paigns. As clients have be­gun to de­mand bet­ter mea­sure­ment of these ef­forts, these firms have added mod­el­ling ex­per­tise to their lists of ser­vices. Be­cause they are mar­ket­ing com­pa­nies at heart, these types of ven­dors are of­ten in the best po­si­tion to en­hance the model with deeper me­dia or con­sumer data. Make sure there is a strong sep­a­ra­tion of re­spon­si­bil­ity be­tween the im­ple­men­ta­tion and the mea­sure­ment sides of the busi­ness to avoid bi­as­ing the anal­y­sis.

Soft­ware com­pa­nies of­fer a mea­sure­ment plat­form

The com­bi­na­tion of data stor­age needs, an­a­lytic tools, and pro­cess­ing ca­pac­ity that drives mod­el­ling makes it in­evitable that a com­pany like Veri­diem would en­ter the field with a plat­form ap­pli­ca­tion. As com­pa­nies em­brace mod­el­ling as a core as­pect of their mar­ket­ing plan­ning, a plat­form like Veri­diem’s gives mar­keters greater ac­cess to the model to con­duct what-if analy­ses of dif­fer­ent mar­ket­ing plan op­tions, while si­mul­ta­ne­ously grab­bing new mar­ket­ing data to en­able more timely model up­dates. But a plat­form alone won’t im­prove mar­ket­ing; com­pa­nies must sub­stan­tially change mar­ket­ing pro­cesses to get value out of this ap­proach.

Hy­brids com­bine mod­el­ing, tech­nol­ogy, and con­sult­ing

As the ecosys­tem of tools and ser­vices be­comes more com­plex, ven­dors have tried to sim­plify the plan­ning and man­age­ment of mod­els with more com­pre­hen­sive so­lu­tions. Mar­ket­ing Man­age­ment An­a­lyt­ics (MMA), one of the orig­i­nal bou­tique modellers, launched its Avista ASP ser­vice in March 2005, i n re­sponse to clients’ re­quests for bet­ter ac­cess, faster up­dates, and bet­ter tools to get value from the mod­els. Ac­cen­ture pro­vides the full range of strate­gic con­sult­ing, tech­nol­ogy build­ing, and change man­age­ment ser­vices. Firms mi­grat­ing to this class of ven­dor should make sure that t hey are not re­stricted to us­ing only this f ir m’s mod­els but can im­port and continue to use their ex­ist­ing mod­els.



For the CEO, mea­sure­ment is as much about pre­dictabil­ity for Wall Street as it is about im­prov­ing the ef­fi­ciency of the mar­ket­ing bud­get. New types of mod­els and new tech­nolo­gies that make them avail­able to ev­ery mar­ket­ing man­ager will en­sure that com­pa­nies rely more and more on their mod­els to shape mar­ket­ing plans that even an in­vestor could love.

Mod­els will evolve from re­sults anal­y­sis to plan­ning tools

Mar­ket­ing mod­els have be­come cen­tral to mar­ket­ing and ad­ver­tis­ing de­ci­sions at lead­ing con­sumer goods com­pa­nies in­clud­ing Kraft, Gen­eral Mills, and Coca-Cola. As these com­pa­nies have gained ex­pe­ri­ence over time, they have be­gun to run what-if sce­nar­ios at dif­fer­ent bud­get lev­els and me­dia al­lo­ca­tions to sup­port an­nual mar­ket­ing plan de­vel­op­ment. To­day’s Mi­crosoft

Ex­cel sim­u­la­tor tools will yield to more so­phis­ti­cated sys­tems to track, col­lect, and model mar­ket­ing in real time.

Data will move from ag­gre­gate sales to in­di­vid­ual con­sumer be­hav­iour

In CPG, the most im­por­tant data comes from ACNielsen or IRI syn­di­cated sales data­bases. As Left Brain Mar­ket­ing expands the num­ber of com­pa­nies with in­di­vid­ual con­sumer data en­riched with be­havioural in­sight like site vis­its and mar­ket­ing pref­er­ences, mar­ket mix mod­els will in­cor­po­rate more gran­u­lar data. Tra­di­tional data­base host­ing firms like Ep­silon and Acx­iom

will fol­low Tar­get base’s lead, adding statis­ti­cians to build con­sumerseg­ment mod­els, draw­ing on in­di­vid­ual trans­ac­tion-level data that the firms hold for their clients.

Mar­ket­ing op­er­a­tions will mi­grate to­ward fi­nance

As mod­els en­able more pre­dictable re­sults from mar­ket­ing, CFOs will see the mod­el­ling process as an ex­ten­sion of the fi­nan­cial con­trols that they wield in other parts of the or­ga­ni­za­tion. To pre­vent a to­tal takeover of mar­ket­ing by the fi­nance depart­ment, CMOs will in­stil greater dis­ci­pline in ex­e­cut­ing and track­ing mar­ket­ing pro­gram re­sults. Cre­ative mar­keters who chafe un­der mod­els’ re­stric­tions will re­de­fine their jobs away from manag­ing ad cam­paigns to new prod­uct de­vel­op­ment, po­si­tion­ing, and com­mu­ni­ca­tion strate­gies.

Mar­ket re­search will be­come less quan­ti­ta­tive

As fi­nance ex­erts greater con­trol in day-to-day ex­e­cu­tion, CMOs will bol­ster mar­ket­ing’s inf lu­ence by re­build­ing mar­ket­ing re­search’s ca­pa­bil­i­ties. Their new mis­sion: Find the unique con­sumer in­sights that can drive in­no­va­tive new prod­ucts. Mar­ket re­search will delve more into “soft” re­search tech­niques like an­thro­po­log­i­cal stud­ies and mon­i­tor­ing on­line con­sumer-gen­er­ated me­dia to divine con­sumers’ la­tent wants, needs, and mo­ti­va­tions.


For­rester in­ter­viewed 31 ven­dor and user com­pa­nies, in­clud­ing: Ac­cen­ture, ACNielsen, The Ad­ver­tis­ing Re­search Foun­da­tion, Drat­field An­a­lyt­ics, iKnow­tion, In­for­ma­tion Re­sources Inc., Knowl­edge Net­works, Se­quent Part­ners, and Veri­diem.Re­lated Re­search Doc­u­ments

“What B2B Mar­keters Need From Tech­nol­ogy” April 8, 2005, Trends “Where Is Mar­ket­ing Mea­sure­ment Headed?” Jan­uary 13, 2005, Trends “Left Brain Mar­ket­ing” April 6, 2004, For­rester Big Idea

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.