Machine Learning for Marketing: User centred design of a decision support system

dc.contributorAalto-yliopistofi
dc.contributorAalto Universityen
dc.contributor.advisorWiberg, Rikard
dc.contributor.advisorNyman, Mattias
dc.contributor.authorDeleuze, Laura
dc.contributor.schoolPerustieteiden korkeakoulufi
dc.contributor.supervisorNieminen, Marko
dc.date.accessioned2018-09-03T12:43:18Z
dc.date.available2018-09-03T12:43:18Z
dc.date.issued2018-08-20
dc.description.abstractOwing to the Machine Learning spread, in particular the Econometric modelling progress, marketers are now able to daily monitor and control the marketing effect of their campaigns, thus optimising their advertising investments. Although several decision support tools and Data analysts consultants currently provide marketers with these data-driven insights, only few manage to understand the algorithms outcomes and act upon the extracted insights on their own. This master thesis thus focuses on understanding marketers behaviours and job to identify where Econometric modelling would be relevant for them to use on a daily basis. To do so, this thesis studied an existing algorithm. The user research consisted in a manifest and a latent content analysis of semi-structured interviews conducted as rigorous Contextual Design Inquiries with 11 marketers familiar with Econo- metric modelling. The extracted patterns were then validated and eventually produced four representative set of marketers work-models and personas. They also grounded corresponding validated design guidelines to help designers build a user-centred tool delivering data-driven insights that marketers can extract alone and autonomously act upon on a regular basis. To deliver understandable and actionable data-driven insights, this user research concludes that Econometric modelling outcomes must be provided through a portable, pertinent and task compliant user-centered tool. First, it should dis- play a centralised overview of how their marketing strategies are currently doing on the market. Second, the tool should contribute to optimise their marketing budget to reach the marketers company business goals. Third, it has to enhance the communication between the marketers, their media agency and their top- management team. By realising these three jobs, the data-driven tool would then constitute a major business asset for the marketers company. Not only would it dramatically increase the efficiency and profitability of marketing activ- ities along with the business managers trust in marketing benefits, but it would also empower marketers to accurately control their budget marketing effect and negotiate their costs down with media channel publishers for instance.en
dc.format.extent150 + 78
dc.identifier.urihttps://aaltodoc.aalto.fi/handle/123456789/33765
dc.identifier.urnURN:NBN:fi:aalto-201809034890
dc.language.isoenen
dc.programmeMaster's Programme in ICT Innovationfi
dc.programme.majorHuman Computer Interaction and Designfi
dc.programme.mcodeSCI3020fi
dc.subject.keywordmarketing strategyen
dc.subject.keywordeconometric modellingen
dc.subject.keyworddecision support systemsen
dc.subject.keyworduser centred designen
dc.subject.keywordPersonasen
dc.subject.keywordwork-modelsen
dc.titleMachine Learning for Marketing: User centred design of a decision support systemen
dc.typeG2 Pro gradu, diplomityöfi
dc.type.ontasotMaster's thesisen
dc.type.ontasotDiplomityöfi
local.aalto.electroniconlyyes
local.aalto.openaccessno
Files