Attribution Modelling or Customer 360⁰ view engineering: Which comes first & how to automate for intervention strategies
Download as PPTX, PDF2 likes394 views
The document discusses the comparison between attribution modeling and customer 360-degree view engineering in digital marketing, emphasizing their respective roles in analyzing customer journeys and improving data quality. It outlines the value of customer journey analytics in enhancing marketing mix modeling by providing insights from individual-level user data. The author, Sjaun Wong, shares expertise from various roles, highlighting the complexity of tracking customer interactions across multiple channels and the need for effective data management strategies.
Attribution Modelling or Customer 360⁰ view engineering: Which comes first & how to automate for intervention strategies
- 1. Attribution Modelling or Customer 360⁰ view engineering: Which comes first & how to automate for intervention strategies Sjaun Wong – 26 Apr. 2018 Digital & Analytics Dialogue UK
- 2. • Former Global Head of Business Insights for GVC, a FTSE 100 gaming company in the UK. • Brands: Ladbrokes, Coral, Party Poker and many more. • I was responsible for the data architecture, visualisation and data transformation strategies. SJAUN WONG Data Strategist Previous roles for Comic Relief, Havas (advertising), Betfair, CallCredit and Rightster (Brave Bison) Highlights: • Rightster founding team (14) grew to 224 staff and IPOed in 2.5 yrs. • I’ve worked in both analytics, web development team roles, advertising and SEO agency roles – experienced with multiple CRM & marketing tools and Data Management Platforms. @SjaunWong SjaunWong
- 3. Agenda @SjaunWong SjaunWong • Define and compare - Marketing Mix, Attribution Modelling & Customer 360 degree view strategy (C360o) aka Customer Journey Analytics. • Outline the value C360o engineering brings to both models by improving data quality matching off-site web data. • The complexity of tracking customer journeys in C360o. • Explain the engineering solution and a quick example.
- 4. Gartner Forrester Attribution Modelling Individual-level data rather than aggregated data and as a result is usually confined to digital marketing channels. 1 The purpose of attribution is to give fair credit to the tactics – placements, creative ideas, formats – that work. 3 The practice of using advanced statistical approaches to allocate proportional credit to marketing communications and media activity across all channels, which ultimately leads to the desired customer action. 4 Gartner Forrester Marketing mix modelling Aggregate data, not user-level data, and explores the impact of a wider range of channels and factors, including non- digital media such as broadcast television and radio. 1 The process of using statistical analysis to estimate, optimize, and predict the impact of paid, owned, and earned multichannel marketing tactics on future business revenue or any other key metric. 2 Defining the challenge 1 Source: "Clarify marketing impact with attributions and marketing mix modelling", Chris Pemberton, Gartner, Inc., 6 Jan. 2017 2 Source: "Webinar: How Agencies Play A Role In Attribution", Tina Moffett and Ari Osur, Forrester Research, Inc., 8 May 2012 3 Source: Blog article: "Is Accurate Attribution Even Possible?“, Martin Kihn, Gartner, Inc., 29 Sep. 2016 4 Source: Blog article: "Just published: The Forrester Wave: Cross-Channel Attribution Providers“, Tina Moffett, Forrester Research, Inc., 7 Nov. 2014 5 Source: "Market Guide for Customer Journey Analytics", Gartner, Inc., 30 August 2017 Gartner C360⁰ view engineering 4 distinct phases of journey analysis: • Gathering • Connecting • Visualizing • Acting Customer journey analytics (Cust. 360⁰ view) is the process of tracking and analyzing the way customers use a combination of available channels to interact with an organization @SjaunWong SjaunWong
- 5. What value does C360⁰ view engineering bring to Marketing Mix and Attribution Modelling? Data time- range Long-term: Months – years. Short-term: Days-months. Short to Medium-term: Weeks-months. Prediction accuracy Better for evaluating overall media ROI. Lacks insight into customer off-site transactions. Improves unified MMM by delivering ROI per user (ad) impression. Data collection time-lag Months – years. Hourly-daily. Usually longer than Attribution Modelling but less than MMM. Marketing Automation Media macro allocation – no automation. Channel allocation – no automation. Digital User allocation + campaign automation. Marketing mix modelling Attribution Modelling (siloed approach) C360⁰ view Attribution modelling Channel / User Scope Macro-channel: Economic/Market factors, media mix & consumer demand KPIs e.g. sales. Digital Micro-channel: Usually cookie-based / device IDs not customer ID. Unified approach: Multi-device & cookie unification matched to customer ID. @SjaunWong SjaunWong
- 6. WebsiteAffiliates Mobile Retargeting (Search & Display) Display Ads App Store Search Web Personalisation Customer Retention Optimisation Loyalty Programmes Word of Mouth Sample Customer Journey Search EngineSearch Ads Social Ads Reviews Comparison sites News / Blogs / Social Media Social Networks Reviews E-commerce PR Radio, TV, Print Customer Service PURCHASE A typical customer journey MACRO: MARKETING MIX MODELLING CUSTOMER 360 DEGREE ENGINEERING = CUSTOMER JOURNEY ANALYTICS ATTRIBUTION MODELLING CUSTOMER DATA – overlaying the approaches @SjaunWong SjaunWong
- 7. Customer journeys are complex @SjaunWong SjaunWong
- 8. Wouldn’t it be nice to have an ID Graph that joins 3rd party Users IDs to Customer IDs? Customer journeys are complex @SjaunWong SjaunWong
- 9. Digital Marketing Landscape Source: "Gartner Digital Marketing Transit Map", Gartner, Inc., 2 Feb. 2018 @SjaunWong SjaunWong Without an ID graph – this compounds the complexity!!
- 10. Framework for an ID Graph: connecting users’ cookies and device IDs per session Unified User ID cookie / device per session unified deterministically first and then probablistically 2nd Party Data 1st Party Data DWH / ERP data IoT CRM Email SMS Digital Analyics Push notifications Social Media Analytics DSP / Ad Exchanges Publisher Ad Private Marketplace Data brokers Marketing Channel activation targeting per Customer DSPs Email Mobile Push Notifications Social Personalisation Affiliates CRM Test & Learn SMS Attribution Modelling & SegmentationAffiliates Partners 3rd Party Data @SjaunWong SjaunWong
- 11. How does this ID Graph work in practice – a simple example Tag Manager Machine Learning / Deep Learning DoubleClick Campaign Manager • Match GID • Return Ad Bid Price • Creative & Ad position • URL • Targeted keywords Web Analytics • Match GID • Page URL • Conversion clicks • AB/Testing • Site Personalisation Google Ad pixel / tag Google Cookie Matching Service Blog Advertiser Website Customer 360 Data Storage Repository
- 12. Thanks!! Email: sjaun@ukprofessionals.com @SjaunWong SjaunWong Sjaun Wong Data Strategist & Marketing Activator