Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

CRM på én dag: Erik Taarnhøj, InsightGroup Nordic

829 views

Published on

  • Be the first to comment

  • Be the first to like this

CRM på én dag: Erik Taarnhøj, InsightGroup Nordic

  1. 1. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved CRM PÅ ÉN DAG Effektmåling Erik Schmidt Taarnhøj, Group Director
  2. 2. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved BrandScience@OmnicomMediaGroup Accelerate / Aug 2013 2 AGENDA Introduktion Metode og analyse setup Cases
  3. 3. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved3 InsightGroup@OmnicomMediaGroup CRM / Januar 2013 BrandScience er en del af InsightGroup • BrandScinece har mere end 15 års erfaring med at bygge kvantitative marketing anbefalinger • BrandScience er en del Af Omnicom Media Group • BrandScience har kontorer I 15 lande • BrandScience I København har mere end 15 ansatte Oversigt over BrandScience kontorer Kommentarer BrandScience Offices BrandScience I København er ansvarlig for Nordics
  4. 4. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 BIO Erik Schmidt Taarnhøj Group Director BrandScience Nordics 9 måneder i BrandScience med fokus på kvantitative analyser i Omnicom M.Sc. in Applied Mathematics fra DTU and MIT Management konsulent hos McKinsey & Co Corporate Finance and Strategy hos DONG Energy 4
  5. 5. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved BrandScience@OmnicomMediaGroup Accelerate / Aug 2013 5 AGENDA Introduktion Metode og analyse setup Cases
  6. 6. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 MÅLING AF CRM AKTIVITETER EFFEKT Direkte måling •DM / eDM •Digitale kanaler •TV Adspurgte effekter •Kundetilfreds. •Survey Isolerede effekter •Modellering Centralt at vi kan måle effekten
  7. 7. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 HVAD ER ØKONOMETRIKS ANALYSE? = Respons Metode kan benyttes til: • Salg • Kundestrøm • Leads • Churn + + MedierKonkurrent adfærd = DistributionSæson Typisk anvendt data Makro faktorer Mediedata Konkurrenters mediedataVejrdata Kalender Markedsdata Below the line aktiviteter Respons- variabel …Baseret på tilgængeligt data Matematisk & statistisk analyse + Tilbudsaviser
  8. 8. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 BIG DATA MULIGGØR HOLISTISK ANALYSE AF HVAD DER PÅVIRKER FORRETNINGEN Word of mouth TV Direct marketing Print Social Media Online display PPCSEO Web tv Mobil Radio OOH PR Offline Online • Øget kompleksitet som følge af stor vækst i antal af kontaktpunkter med forbrugeren • Mediespecifikke evalueringer er gode, men viser kun én brik af det samlede billede
  9. 9. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 FRA DATA TIL VÆRDIFULDE INDSIGTER (OG EFFEKT DOKUMENTATION) 1. Indsamle data 2. Identificere key drivers 3. Kvantificere drivers og udregne ROI på marketing- indsatser 4. Levere værdifulde indsigter
  10. 10. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Data anvendes til at forstå udviklingen i forretningen 4,1% 3,6% 1,6% 0,5% 0,4% 0,4% 0,3% 0,1% 0,0% 6,2% -0,1% -0,4% -0,7% -1,2% -2,5% 0,0% 2,0% 4,0% 6,0% 8,0% 10,0% 12,0% 14,0% Effect% Distribution Shelfpries Marketdev. Highseason Add.FTE Media(ATL) Promotion Media(BTL) Openingdays Comp.media PR Wheather Events Totalsales increase DM ”When sales are down it is either due to the weather or the economy, when sales are up, it is because we are excellent marketers” Anonymous
  11. 11. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Mulighed for at analyser hele kunden livscykel Modelfokus: o Web visits o Svar kort return Customer life cycle Customers leads New customers Existing customers Churn Modelfokus: o Antal nye kunder o Omsætning fra nye kunder Model fokus: o Total el. gennemsnitlig salg fra eksisterende kunder Model fokus: o Antal inaktive kunder o Kunde afgang Setup med både leads og nye kunder giver et billede af kvaliteten af de leads som skabers 11
  12. 12. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Eksemple på projekt setup D & E. In/Out of scope Outbound sales Inbound sales Effects within market place Analysen dækker alt salg som kan tilbageføres til DM 1. Salg I Inbound 2. Salg i webbutikken 3. Salg gennem “outbound” som respons på et svarkort 4. Salg gennem Outbound som følge af en “Ring mig op lead” 5. Det totale salgløft I Outbound måles gennem en salgsmodel Illustration – DM consumer journey Kommentarer Customer calls customer service Web sales Point of sale x 1 2 Send DM to customer Customer has multiple options for response Sales included in analysis Customer sends ‘svarkort’ 3a Customers goes online Customer selects ‘Ring mig op’ “Company” webpage 3b 1 2 3a 3 3 3b “Company”
  13. 13. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved BrandScience@OmnicomMediaGroup Accelerate / Aug 2013 13 AGENDA Introduktion Metode og analyse setup Cases
  14. 14. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Rekruttering af tilmeldinger Rekruttering af medlem Omsætning Betaling af medlemskontingent og tilmeldinger (nye og eksisterende) PROJEKTETS FORMÅL Case At identificere hvilke faktorer, der påvirker udviklingen i følgende KPI’er:
  15. 15. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Case Insight # 1 Uadresserede udsendelser er 4 gange så effektive som TV annoncering i forhold til at rekruttere nye tilmeldinger
  16. 16. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 UADRESSEREDE UDSENDELSER ER 4 GANGE SÅ EFFEKTIVE SOM TV ANNONCERING Case 432 100 TV HUS Indekseret pris per ny tilmelding (uadresserede udsendelser = 100) Implikation  Der bør anvendes uadresserede udsendelser som det centrale element i rekrutteringsstrategien  TV bør ikke være den primære kommunikationskanal i forhold til at genere nye tilmeldinger Uadresserede udsendelser +400%
  17. 17. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Case Insight # 2 DM er lige så effektivt til at skabe meromsætning som TV, men DM står for en større del af medieløftet
  18. 18. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 TV DM 0 50 100 150 ROI(100=DM) Indekseret omsætningsløft (DM = 100) STØRST VOLUMEN FRA DM Case 102 100 TV DM Indekseret ROI (DM = 100) Implikation  TV og Direct Mail er lige gode til at genere meromsætning, men fordi DM aktiviteten er blevet anvendt mere hyppigt har den generet mere omsætning  Begge medier bør anvendes, såfremt målsætningen er at generere meromsætning Størrelsen på boblen repræsenterer størrelsen på investeringen
  19. 19. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Case Insight # 3 2 + 2 = 5 DM virker 17% bedre, når de udsendes samtidig med TV annoncering
  20. 20. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 MEROMSÆTNINGSEFFEKT PÅ 17% SOM FØLGE AF SYNERGI MELLEM TV OG DM Case 57% 26% 17% 0% 20% 40% 60% 80% 100% 120% DM TV Synergi Effekti%(100%=totaleffekt) Indekseret effekt af TV og DM på julelotteriet (omsætning) Implikation  Der opstår en synergieffekt når DM udsendes samtidig med en TV kampagne  Planlæg direct marketing kampagner sammen med masse kommunikationskampagn er
  21. 21. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved Case 2
  22. 22. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 MÅLING AF CRM AKTIVITETER EFFEKT Direkte måling •DM / eDM •Digitale kanaler •TV Adspurgte effekter •Kundetilfreds. •Survey Isolerede effekter •Modellering
  23. 23. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved23 InsightGroup@OmnicomMediaGroup CRM / Januar 2013 CHALLENGE CRM modeling ”In a CRM system with more than 100.000 names who should I send DM”
  24. 24. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved24 InsightGroup@OmnicomMediaGroup CRM / Januar 2013 ”What is the probability that an existing customer will makes a purchase in the next 3-6 months?” CHALLENGE CRM modeling KEY STRATEGIC QUESTION ”In a CRM system with more than 100.000 names who should I send DM”
  25. 25. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 KEY STEPS AND PURPOSE OF A CRM MODEL • Structuring data in order to ensure an efficient modeling of the data – i.e. trans- formation of variables, cleaning data, calculating supplementary variables, etc. • Initial testing of relevant methods • Identify the efficient indicators of consumers buying probability • Identify the buying probability with in a given period (e.g. 6 months) • Define relevant and operational segments based on: • Buying probability • Expected value • Drivers for driver probability • Etc. • Identify initiatives across relevant segments which increases buying probabilities • Extensive enrichment of segments: • Online behavior • Media habits • Attitudes and needs (qualitative and quantitative) • Etc. • Ensure the ability to act proactively PurposeDeliverables • Data structured in required format • Recommended model approach • Clarification of analysis and granularity level • Overview of significant indicators of buying probability • Calculation of probability index for each client ID • Identify top category for near purchase • Overview of relevant buying segments • Efficient targeting (Quick- Wins, tactical and strategic) • Business case • Segments operationalized in CRM database • Insights into efficient activation of segments Structuring data and initial testing Conducting specified analysis Segmenting customers Data enrichment 1 2 3 4
  26. 26. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved26 InsightGroup@OmnicomMediaGroup CRM / Januar 2013 o Example of data from 20 customers: o Number of kids o Age o Education o Calls o Store visits o Did the customer churn o On the basis of the CRM data, relationships between for example age, store visits, education and churn can be identified through the application of a variety of Machine Learning algorithms. o Analysis output is a re-buy probability for each Customer in the CRM database. o The re-buy probability can then be interpreted through a probability index (e.g. low, medium and high) – depending on the scope of the analysis. CASE STUDY: Estimation of re-buy probability from CRM data SamplefromrawdataSamplefromanalyseddata **
  27. 27. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved est@insightgroupnordic.com Spørgsmål?
  28. 28. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved Case Churn prevention
  29. 29. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Churn prevention / Case Aktivering af viden fra kundedatabaser Datasæt holdt op på Kunde id: Forklarende variable • Demografiske • Geografiske • Antal år som kunde • Kundetilfredshed • Typer af køb • DM udsendelser • Nyhedsbreve Afhængig variabel • Churn / krydssalg / life time value / segmentering Identificering af vigtige parametre Beregning af sandsynlighed for churn Segmentering What if scenarios Data: Outcome: Metode: Logistisk regression Cluster Probitregression eller logistisk regression Probitregression
  30. 30. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved InsightGroup@OmnicomMediaGroup CRM / Januar 2013 Churn prevention / Case Aktivering af CRM data Værdi af dialog med eksisterende kunder Selskab analyserede sandsynligheden for churn på baggrund af en række parametre Learning: 14% længere gnsn. levetid hos kunder med regelmæssig kontakt
  31. 31. Copyright © Omnicom Media Group. Confidential and proprietary. All rights reserved est@insightgroupnordic.com Spørgsmål?

×