Marketing Analytics with Mcdonald's Data Scientist
1 like•325 views
Promotable offers educational solutions focused on in-demand skills for career advancement, taught by experienced practitioners from top companies. They provide various upcoming courses and events related to data analytics and marketing measurement. The document emphasizes the importance of understanding data and marketing mix modeling to evaluate business performance and optimize marketing strategies.
Marketing Analytics with Mcdonald's Data Scientist
- 3. Success in today’s economy means understanding Data. At Promotable, we create education solutions that empower career potential and advancement. We teach in-demand and marketable skills that can help anyone advance in his or her career.
- 4. Learn from Real World Practitioners Promotable’s instructors work at top companies and combine real-world experience with people-focused skills.
- 5. Upcoming Events & Courses
- 6. Upcoming Events ● April 30: Cloud Essentials for Managers - What you Need to Know ● May 9: A/B Testing with Facebook’s Marketing Science Partner
- 7. Upcoming Classes Intro to Data Analytics (5 hrs/wk for 6 Weeks) ● April 23: 2 spots left ● May 7: 6 Spots left Data Jump Start (2.5 hr class) ● May 7 (in partnership with the Chicagoland Chamber of Commerce) ● May 22
- 8. Journey Towards Better Marketing MeasurementAlex Gaski Data Scientist- McDonald’s US 4.17.2019
- 9. About me Data scientist: Performance Learning Manager @ McDonald’s US • Drink of choice: Wine, Rieslings current favorite • Pet Peeves: People getting on the L before people get off • Go-to McDonald’s Order: Egg McMuffin • Last Movie Watched: Captain Marvel • What music is on your playlist: Jazz, Big Band Music, along the lines of Nina Simone and Frank Sinatra • Favorite Sport Team: Chicago Blackhawks • Most Admired Leader: Benjamin Franklin
- 10. Today’s topic
- 11. Agenda • How to optimize media to business goals rather than media metrics • Worst to best methodologies for media measurement • Challenges to understanding true return on marketing spend • What is Marketing Mix Modeling • What kind of data should you be collecting • How to get started
- 12. Sales GCs Market Share Marketing results are influencers of but not synonymous with business results Marketing Performance Indicators Awareness Engage Convert Advocate Social media Volume & Sentiment Email open rate Promoted product URWs Search volume Awareness Strength of creative Media plan store visit rate Issue: Leadership wants to know how marketing is contributing to the business but traditional marketing metrics don’t answer this
- 13. Media Metrics In Store Traffic In Store Sales Brand & Creative Metrics Measurement Metric On-Target Reach Ad Verification Frequency Lifts in Awareness, Perception, Association and/or Intent In Store Lift (currently leveraging proxy data) Sales Lift (currently leveraging proxy data) KPIsObjective Increase Consumer Brand Love Drive Reach | Frequency & Meet Benchmarks Drive In Store Traffic Convert Sales Where most marketers sit Where they want to Go Typical way to measure Total Traffic Pre Post Cost Per Inc Visit Multitouch Attribution Surveys Focus Groups Engagement (CTR, Video Completion) Nielsen Ratings Publisher reports Sales Pre Post Media Mix Modeling (MMM) Worst To Best Ways To Measure Media Performance
- 14. We want to know how our marketing performed … but challenge is so many external factors effect performance 7 Outside Company’s control • Weather • Population demographic near stores • Minimum wage law changes • Consumer Buying Power • US Economic Outlook • Gas prices • Unemployment • IEOtrends • Competitive media spend • Competitive pricing strategy • Competitive store location openings/closings Within Company’s Control ► What we are selling ► Customer Satisfaction ► Complaints per Guest Counts ► Pricing Strategy ► Store variables (e.g. # of drive through lanes, # of employees, location) ► Store openings/closings ► Online vs Offline Within Company’s Marketing Control ► Spend per product category & message type ► Numberof campaigns ► Media mix foreach campaign ► National and Coop roles and responsibilities ► Flighting strategies/Campaign overlap ► Instore promotional signage ► PR► Discounts/offers *List not exhaustive of all possible factors
- 15. Econometric modeling -> Marketing Mix Modeling (MMM) is gold standard in evaluating marketing performance measurement • MMM modelingteases apart the effects of media and marketing activities that occur at the same time to determine the impact on sales and ROI measures for individual marketing activities • Accordingto MMA 2018 survey 46% of companiesare using MMM models to assess marketing/media performance $- $100.0 M $200.0 M $300.0 M $400.0 M $500.0 M $600.0 M $700.0 M $800.0 M Baseline Sales Pricing Impact from New product Decrease Competition Marketing Weather launch Service Times Marketing Total sales Example Salesdecompfrom MMM 30% 5% 15%11% 11% 19% 6% 3% Example Contribution to Sales by Channel TV Prin t OO H Radio Digital Display Social Streaming audio Search
- 16. ROMI is used not only to understand past performance but to optimize media spend across channels, products, messages Media Investment Levels/PODR Understand Media Saturation Levels and points of diminishing return Halo and Cannibalization Understand how new changes are affecting the overall ROI on marketing and evaluate the effects of local vs national spending Calendar Planning/Optimal Message Pairings Understand what messages combine when advertised in the same time frame Campaign Optimization Understand factors what drives the best ROI and the most sales across Products, Message types, and Media Vehicles
- 17. Some Examples of how we are using MMM today for Media
- 18. Other MMM outputs – Calendar Simulations • ROMI is an input we use not only to evaluate past marketing performance • We have a simulation & optimization engine built on top of the model that allows us to run possible calendar combinations to determine best ROMI and Sales outcomes • Limited capabilities for Coops at this time as it assumes all Coops go to market the same (e.g. Balanced Archetype used for all Coops)
- 19. What data do I need for this? Economic Variables • Weather • Population demographicnear stores • Minimum wage law changes • Consumer Buying Power • US Economic Outlook • Gas prices • Unemployment • IEO trends • Competitive mediaspend • Competitive pricing strategy • Competitive store location openings/closings Company Variables ► What we are selling► Customer Satisfaction ► Complaints per Guest Counts ► Pricing Strategy ► Store variables (e.g. # of drive through lanes, # of employees, location) ► Store openings/closings ► Online vs Offline Marketing Variables ► Spend per product category & message type ► Numberof campaigns ► Media mix foreach campaign ► National and Coop roles and responsibilities ► Flighting strategies/Campaign overlap ► Instore promotional signage ► PR► Discounts/offers *List not exhaustive of all possible factors
- 20. Step 1 Find your advocate in the organization Step 2 Start collecting the data Step 3 Hire data scientists or analytics firm Not something you necessarily want to do on your own Step 4 Step 5 How can I get started? Create a culture around measurement Benchmark against yourself and cut bottom 20% of what you’re doing and reinvest in top 20%
- 21. Q&A
- 24. Data Jump Start ● Gain confidence navigating datasets ● Learn powerful, advanced functions and features like Pivot Tables, calculated fields and conditional formatting ● Leverage Excel’s most important and powerful features ● Analyze data and derive insights
- 25. Intro to Data Analytics ● Use Excel, SQL and Tableau to collect, clean and analyze large datasets ● Understand some of the basic fundamentals of Data Science ● Present data-driven insights to key stakeholders using data visualization and dashboards ● Tell compelling stories with your data